Wikipedia:Requests for comment/Article feedback

This is a requests for comment about article feedback. It is scheduled to end on February 21. A request for closure of the discussion has been filed.

Background
For over a year, the Wikimedia Foundation has placed a comments box on selected articles that allows readers and other page viewers to add feedback (comments) to articles. This is known as article feedback. A feed of these comments is available at Special:ArticleFeedbackv5.

The Wikimedia Foundation plans to expand this feedback box to all articles on the English Wikipedia by the end of March 2013.

The purpose of this request for comment is to determine the answer to a few questions:


 * 1) What should the scope of the article feedback tool be (cover all articles, opt-in per article, etc.)?
 * 2) Are there sufficient resources to moderate and respond to all of the feedback?
 * 3) How will abuse filters handle articles in which the subject's title contains a disallowed word (e.g., Blue-footed Booby + "boob")?
 * 4) Will the tool continue to only be a(n expanded) box or will it go more minimal? Will it go "above the fold" (e.g., File:Article-link-to-feedback.png)?

Users are encouraged to sign below supporting a particular view or to add their own view.

View by MZMcBride
There are currently insufficient resources to moderate and respond to article feedback for all articles. Consequently, the article feedback tool should be available as an opt-in feature on a per-article basis, allowing individual interested editors to receive feedback on articles that they are willing to monitor and respond to.

The design of the article feedback tool should be as minimally intrusive as possible, recognizing that the content area of articles is sacrosanct. Comments should never appear below an article and the tool itself should be positioned in an inconspicuous place to avoid disruption to readers.


 * Users who endorse this view
 * 1) MZMcBride (talk) 01:33, 21 January 2013 (UTC)
 * 2) Support - I'd add that article feedback as a concept seems to have started as a trick to get people involved in editing Wikipedia, that this experiment has not produced demonstrably positive results, that the comments generated are essentially garbage, that the feedback boxes are disruptively large, and that the whole exercise is a waste of time at this point. "Opt-in" and "Make smaller" are probably the closest principles to this view that have a chance of gaining sufficient support, however, so I will endorse those. Carrite (talk) 01:52, 21 January 2013 (UTC)
 * 3) Weak support - The article feedback tool is almost solely used to submit vandalism and nonsense rather than contructive feedback that engages readers. While the tool itself might be useful for some wikis, on enwiki it obviously does not work. Reaper Eternal (talk) 03:27, 21 January 2013 (UTC)
 * 4) Support. The feedback tool should only be implemented on an article if there is at least one editor who is willing to monitor the feedback. I have other concerns, but will express them in another section.--Srleffler (talk) 04:00, 21 January 2013 (UTC)
 * 5) Agreed I clicked the "Feedback from my watched pages" link once on a whim and not only found not a single useful comment, but the majority of what was there was people wanting the lyrics to a particular song (which would of course have been a copyvio). Not just unhelpful but actively discouraging. Andrew Lenahan -  St ar bli nd  05:33, 21 January 2013 (UTC)
 * 6) This is probably the most we can get away with, but the extension should be obliterated from en.wp altogether. MER-C 09:49, 21 January 2013 (UTC)
 * 7) Support As a direct response to question 1, I support this option of a narrower scope as opposed to site-wide implementation. This would allow a broad/faster sampling of opinion which might be more efficient than (e.g.) peer review, though this doesn't tackle the problems of attracting new users Jebus989 ✰ 14:16, 21 January 2013 (UTC)
 * 8) Semi-support If its decided to keep the tool, this would be the best way of managing it as it would limit the scope for 'comments' to the articles where people are monitoring the comments and following up on them. I imagine that readers would be confused about why they can comment on some articles but not on others though. Nick-D (talk) 07:14, 22 January 2013 (UTC)
 * 9) Sorta. — foxj 19:58, 22 January 2013 (UTC)
 * 10) Support In a way this the same thing as the feedbackdashboard, only concerning an article. Even if we can't respond to them all, we can use the feedback to help improve the article. Also shouldn't this be somehow integrated into the talk page of the article?— Ross coolguyCVU &#124; 00:11, 23 January 2013 (UTC)
 * 11) If the tool must stay, then this is the only option. - filelakeshoe   &#xF0F6;    14:00, 23 January 2013 (UTC)
 * 12) Support However someone should first do an impartial evaluation of whether the tool is worth its cost. The feedback I received through this tool is mostly useless, for its banal contents and for the random order in which it is presented. Ideally feedback should be placed in the talk page, where it can be seen by editors who have not yet edited that article. Editors who care will see such feedback in their watch list. Perhaps the Talk pages can be made a bit more reader-friendly, e.g. by adding newer comments at the top rather that at bottom? Alas, Wikipedia already has way too many features that provide little value to readers but consume inordinate amounts of editors' time (like category tags, sorting tags, stub tags, editorial tags, navboxes, wikiprojects, article evaluation, citation templates, ...). If there is an area where the Foundation should rule without piety is in the weeding out of useless features and complexity. Will it ever have the courage to do so? --Jorge Stolfi (talk) 16:08, 23 January 2013 (UTC)
 * 13) On this basis it might be worthwhile, but there would need to be a process for removing it from pages where whoever had invoked it had lost interest.  Ϣere Spiel  Chequers 16:14, 23 January 2013 (UTC)
 * 14) Support This makes sense and solves most of the problems that exist currently. Vaca  tion  9  16:33, 23 January 2013 (UTC)
 * 15) Michig (talk) 17:13, 23 January 2013 (UTC)
 * 16) support -- TRPoD aka The Red Pen of Doom  22:41, 23 January 2013 (UTC)
 * 17) support While a good feedback tool would be useful, in its current form it has too many problems and mainly attracts spam and frustrates editors, so it should not be rolled out until significantly improved. -- ELEKHHT 00:20, 24 January 2013 (UTC)
 * 18) Support - The spam and abuse filters should be further improved so that they can catch most unconstructive feedback comments. If more unconstructive comments are prevented from being submitted in the first place, then feedback wouldn't be seen as a waste of time to review. Also, how about having the tool act like it submitted the unconstructive comment? This way, it might prevent commenters from submitting actual unconstructive comments if they think that it got past the filter (provided that the commenter doesn't check the most recent comments to see if it was submitted). - M0rphzone (talk) 06:59, 24 January 2013 (UTC)
 * 19) Support I simply don't have the time to check large amounts of article feedback every day. The real worry is that libel or oversightable material will stay on a feedback page longer than it would on a talk page.-- ♦Ian Ma c M♦  (talk to me) 10:33, 24 January 2013 (UTC)
 * 20) Support I agree with this viewpoint and with IanMacM directly above. It's very useful when you want the feedback, but when no one is reading that's just an opportunity for spam. Am I the only cleaning up Minecraft's feedback? Boring by myself. &bull; Jesse V.(talk) 16:24, 24 January 2013 (UTC)
 * 21) Strong support for making it less intrusive. There is no reason to have it jammed between the footer template and the category list. It need only be a link in the toolbox or whatever. As for opt in I have no real opinion. -- Alan Liefting (talk - contribs) 06:01, 25 January 2013 (UTC)
 * 22) Support per Mr  T  (Talk?)  [ (New thread?) ] 07:41, 25 January 2013 (UTC)
 * 23) Support Too much useful feedback is going unread. Shii (tock) 07:58, 25 January 2013 (UTC)
 * 24) Support both opt-in and less intrusion for this rather unhelpful tool. Cheers, LindsayHello 08:54, 25 January 2013 (UTC)
 * 25) Support Too much useless feedback to wade through, too many random letter edits that would never stick in articles suggests that the filters or vandal-bots are not working properly on feedback. The few useful comments are often not noticed because we now have five different places to check for them. Would much prefer a simplified tool that could dump useful comments into the regular talk page (or just a new improved talk page for all). Rmhermen (talk) 18:08, 25 January 2013 (UTC)
 * 26) Support - Though I also would support just removing the tool altogether. (Hey look MZM, I'm supporting one of your proposals, care to look over a couple of mine? grin : ) - jc37 23:19, 25 January 2013 (UTC)
 * 27) There is a lot of noise to relatively little useful signal, but thankfully little of that noise is libel, copyvio, or the like. (In my experience, it's mostly stuff like "Good article" or "You didn't answer my homework question".) Still, opt-in rather than opt-out seems necessary to ensure that at least one editor is looking at the feedback. Also, in the spirit of "minimally intrusive", I'd prefer to see the new more visible link on the talk page, rather than the top of the article. Cnilep (talk) 01:50, 26 January 2013 (UTC)
 * 28) Support per Carrite - if we can't get rid of the whole thing completely and get the devs working on some urgently needed really serious Wkinpedia features. Kudpung กุดผึ้ง (talk) 10:17, 26 January 2013 (UTC)
 * 29) Support because I agree on the lot of noise vs little useful signal argument. I would add that while I don't believe Wikipedia has limited resources (when editors are enthusiastic about something, they seem to grab time from the cosmos), I do believe that the feedback tool distracts editors from editing in the main namespace. It certainly distracted me. Jane (talk) 10:46, 26 January 2013 (UTC)
 * 30) Support - I have read many of the arguments on this page, and it seems like having the feedback option only on articles that an editor is willing to monitor is the best option. I like the feedback option for readers who don't feel comfortable editing, and haven't seen as much vandalism as some others have (maybe we're reading different articles) but I understand the need people feel for some oversight. The feedback tool is already not available for every article, so it wouldn't change anything in that regard. Azaleaa (talk) 22:28, 27 January 2013 (UTC)Azaleaa
 * Support. Opt-in per article seems like an appropriate use of this feature, and I can think of articles I would probably turn it on for. I've signed other ones that oppose the continued use of the feature overall, but I do think this limited use could conceivably be beneficial. But my question below would still apply. -Pete (talk) 19:15, 29 January 2013 (UTC) No longer think this is a good idea, see here. -Pete (talk) 18:43, 17 February 2013 (UTC)
 * 1) Distant second choice to the completely disable views I endorsed below.  This would avoid many problems mentioned in those views, but it seems that it would also avoid many of the supposed benefits as well.  – Philosopher Let us reason together. 23:57, 29 January 2013 (UTC)
 * 2) Support This user has made a good point. Opt-in per article seems to be appropriate here. &mdash; Forgot to put name (talk) 09:30, 30 January 2013 (UTC)
 * 3) Support. Opt-in would enjoy a higher moderation rate. Let active editors propose articles for feedback, just like with the "freestyle-list" on german WP. Shut off AFT5 on high-traffic articles which are already quite good and where the noise in the feedback is quite high (only 25% "useful" according to WMF analysis). Keep AFT on the old "random" articles bucket (~23.000 articles?) as a control group while improving the AFT moderation and feedback quality (see my suggestions for better feedback quality at ). Getting feedback on less popular articles by thoughtful readers that are now deterred by the WikiMarkup on talk pages (just look at the templates... and there will be no visual editor for talk pages...) - that would be great: to improve articles, to motivate editors by feedback, to lead those readers to be bold and become editors. Let's keep AFT, on a smaller scale, and make it better to get there. --Atlasowa (talk) 09:52, 7 February 2013 (UTC)
 * 4) Support. I think it would be great if the AFT were to be linked somehow to the AFC process. Let the people who are trying to keep the door closed to new articles look at the feedback they receive on articles they have stashed in their "cooler". Jane (talk) 14:20, 7 February 2013 (UTC)
 * 5) Support  If not dropped completely.  Too much prime real estate and distraction for too little benefit. North8000 (talk) 19:00, 7 February 2013 (UTC)
 * 6) Support, as a second choice to removing it entirely. If we must keep this, making it an opt-in is important. 1ForTheMoney (talk) 16:48, 9 February 2013 (UTC)
 * 7) Support, as a minimum choice for now, with hopes for wider rollout of improved versions. I've already supported  below, and feel that supporters of Mike's and other more supportive positions should be assumed as supporting this view as a compromise preferable to total abandonment. I only recently figured out that opting in is as simple as adding an article to Category:Article Feedback 5 Additional Articles -- Wbm1058 (talk) 02:09, 21 February 2013 (UTC)


 * Comments
 * What happens if two editors of an article disagree over whether they want the feedback box to be included? Also, it seems to be a question of  rather than . Finally, I find that allowing editors to decide whether they want feedback in "their" article undesirably promotes WP:OWNership.AgnosticAphid  talk 01:12, 18 January 2013 (UTC)
 * If two editors are in disagreement, they do what you do in the case of any other article-related disagreement: take it to the talk page. :-) I understand the distinction you're trying to make between will and resources, I think. You're saying that Wikipedia does have the resources (enough man-power), but the resources just won't be allocated in this way (there's not enough will or determination from editors to actively monitor the feedback)? I can't say I disagree, but I think that still qualifies as having insufficient resources, if the resources are unavailable, as I wrote. I'm more than open to better phrasing, though. What do you have in mind? Regarding article ownership, I think many editors take pride in their work. We see this more clearly in featured articles, but it's still true through the entire project. WP:OWN is about controlling the content of the article in an unfair way, it's not about taking pride in your work or protecting it. --MZMcBride (talk) 01:20, 18 January 2013 (UTC)
 * I guess I feel like saying we have insufficient "resources" implies it's to monitor the feedback whereas I think the issue you're referencing is that people don't seem to  to monitor the feedback.  The former would be an inherent problem with feedback; the latter could be because the tool is new or unfamiliar and might not always remain the case.  To be honest, though, in my limited time spent reviewing feedback I haven't even noticed that there  a lack of will.  My issue with a disagreement over inclusion of the tool is that there seems to be no real basis to resolve a disagreement over whether the tool should be used.  Wouldn't the arguments always be "I don't like it" vs "it seems useful to me!"?  And then it seems like the default would have to be, just don't include it if editors disagree.  W/r/t ownership, it came to my mind because I saw that on the main article feedback talk page that some people were requesting to add "their" articles to the feedback blacklist.  I don't wish to see an undesirable tool shoved down anyone's throat, which would probably cause a loss of editors, but if the project as a whole thinks that it is valuable to allow anonymous editors to more easily comment upon article quality, why should a single editor be able to decide that for aesthetic or other reasons that "their" article doesn't get feedback and keeps the pretty useless "rate this page!" box instead?  It kind of seems like this is allowing a sort of "local consensus" on the feedback tool for each article. I don't think I'm expressing myself very well on this latter point, to be honest.  AgnosticAphid  talk 09:08, 18 January 2013 (UTC).  P.S.  I apologize if I'm not supposed to comment on the views before the RFC actually starts.  I'm not ver experienced and i found myself not quite agreeing with either view presented; yet I don't feel confident writing a view of my own.  But it seems I've spilled a lot of "ink" already; I hope it's not off-putting.  AgnosticAphid  talk 09:15, 18 January 2013 (UTC)
 * Regarding responding to feedback, as I read the stats, only 11% of feedback is moderated within a month. (I'm not a stats guy, so if I'm wrong about the conclusion I'm drawing, please let me know.) I don't think arguments would (or should, I suppose) include "I don't like it" or similar. The feedback from the tool itself should be able to stand on its own, I think. Editors are able to look at the feedback being received and evaluate whether it's serving its purpose to create a better encyclopedia. I doubt there would be (m)any disputes, though perhaps I'm being overly optimistic. :-) I agree that the ratings box (a.k.a. ArticleFeedback, as opposed to ArticleFeedbackv5) is pretty useless. I believe it will be disabled shortly, as it's no longer supported and the data it's collecting is not even being monitored or used, I'm told. I can say that "local consensus" (or an "article-by-article" approach) has been strongly advocated for in the past for certain issues as it's often an effective means to prevent a more top-down, forceful approach to articles. Absolutely no worries regarding "spilled ink." As I said below, I think having comments prior to beginning of endorsing/supporting individual views is a really good thing. I don't find it off-putting at all. Thank you for taking the time to post your thoughts; they've been helpful for me to read. --MZMcBride (talk) 16:48, 18 January 2013 (UTC)
 * If the tool is opt in as you suggest, not opt out, there will not be any feedback for editors to "look at the feedback being received" and resolve the dispute in the manner you suggest. What would happen is editor A says, "I think we should opt in because I like feedback!", and then B says, "Naw, it's obnoxious," and then I guess they just don't enable it since how can such a thing be decided?   also, I'm no WP expert but it seems to me that disputes over WP:LOCALCONSENSUSes have caused problems and hard feelings in the past.  Finally, I could be operating under a misunderstanding if some kind here, but I thought that the ratings box, which I think is AFTv4, was being phased out  of this tool, AFTv5.  If AFTv5 is rejected per GregJackP or some other view, wouldn't it mean going back to AFTv4 rather than just getting rid of AFT completely?  If so, then isn't failing to opt in to AFTv5 for an individual article not that different?  If not, wouldn't that violate the spirit of the RFC that established AFTv5? AgnosticAphid  talk 00:40, 23 January 2013 (UTC)
 * I got kind of lost in your reply, but regarding further deployment of or Wikimedia Foundation engineering efforts toward this project (article feedback), you'll want to have read this comment from the talk page. I can't say for sure how this RFC will end, but given the direction it's headed in, the people who would like to see the article feedback tool's continued deployment here on the English Wikipedia may want to focus on an opt-in strategy as suggested in my view, as a compromise measure. --MZMcBride (talk) 01:16, 23 January 2013 (UTC)
 * Yes, thanks. Given that this has been running for a whole four days and there are about as many editors who support full adoption as support your opt-in view, the two of which combined are about as popular as GregJackP, I decline to endorse an opt-in strategy as a last-ditch compromise measure.  In my opinion, the format of this discussion is not really conducive to saying, "I endorse this view only if this other view prevails". AgnosticAphid  talk 01:27, 23 January 2013 (UTC)
 * I don't see why editors A and B should be able to decide that editor C can't receive feedback on article X simply because A and B don't want to deal with it. Perhaps A and B should be able to enable a user preference that hides the feedback box and links for them only? --Aurochs (Talk | Block) 19:36, 18 January 2013 (UTC)
 * I agree. I don't see why editors A and B would have an issue with receiving feedback if it's useful, helps in building a free encyclopedia, and there's a willing editor (editor C) who has volunteered to respond to the feedback. My concern is not this scenario. My concern is a very plausible scenario in which there are mounting piles of feedback on millions of articles that will receive no attention from anyone. Even during this extended trial period, we've seen thousands of articles with feedback that goes un-responded to. I think this is unfair to the people leaving feedback and I think it creates a lot more potential harm (from bad feedback [vandalism, libel, etc.] that lingers) than potential good. --MZMcBride (talk) 22:08, 18 January 2013 (UTC)
 * It is likely that the current backlog of feedback exists mostly because most of our editors are not familiar with AFT5 or how to use it. The WMF has done practically no publicity work for this tool during the beta period. I expect a lot more editors will be scrutinizing the feedback once it launches to 100% of en.wiki. Further, there will be a couple of software releases soon that address concerns with usability and the moderating tools we have available right now, which might also help engage editors with moderating feedback. Finally, an at-will opt-out just doesn't seem to be a very good way to address this concern. It would probably exacerbate the issue on those articles that continue to carry AFT5. Editors would probably become interested in working with feedback, since it wouldn't consistently be there to work with.
 * Wrt editors A and B, your proposal will almost certainly create the exact situation that I described. Editors steamroll each other all the time, for the stupidest of reasons. And what if A and B are unanimous, and then they decide to leave the project 8 months later? Editor C will land on article X and find that there's been no feedback for the past 8 months that he could use to improve it. --Aurochs (Talk | Block) 01:58, 19 January 2013 (UTC)
 * I'm going to take exception to your comment that the WMF has done nothing to advertise AFT5; it's actually been better advertised than almost all "new" extensions or UI changes I've seen in the past 5 years. Risker (talk) 02:02, 19 January 2013 (UTC)
 * The only reason I ever heard of AFT5 is because I clicked the little link at the top of my watchlist. Otherwise, the only publicity I've seen on this wiki has been on pages dedicated to the tool itself. If I missed something, please let me know, but me missing things doesn't speak volumes about the quality of the publicity. --Aurochs (Talk | Block) 15:08, 19 January 2013 (UTC)
 * In a vacuum, I'd say that putting a prominent link on the watchlist of the English Wikipedia is a pretty high level of publicity. But I understand various mitigating factors here. --MZMcBride (talk) 17:44, 20 January 2013 (UTC)


 * In response to Agnosticaphid (way above), I doubt that arguments over whether or not to turn the feature on would be a big deal. We have more important things to argue about. And before anybody rolls their eyes, answer me this: when was the last time you saw a protracted, disruptive debate about whether an article should be rated C class or B class? Because we had exactly the same kind of concern around the introduction of the C class quality rating. Believe it or not, some arguments really are too trivial for Wikipedians. -Pete (talk) 19:15, 29 January 2013 (UTC)

View by Nyttend
We don't have enough resources to moderate and respond to article feedback for all articles in a rapid manner. However, widening this tool's use to all articles will be in the spirit of "given enough eyes, all bugs are shallow"; better to make a late fix than to make no fix at all. Do what we can to increase editing, there are always going to be people who won't want to edit a page but who will be happy to leave comments. We should encourage them to help, even when we can't address their concerns immediately. The design of the article feedback tool should be as minimally intrusive as possible etc.; no disagreement with MZMcBride's second paragraph.


 * Users who endorse this view
 * 1) Support. Abyssal (talk) 04:45, 23 January 2013 (UTC)
 * 2) Support - feedback does not need to be dealt with in a timely manner, it's just like an easy-to-use version of the talk page. For editors worried about vandalism in feedback, it's worth noting that the feedback tool automatically hides from view by default those comments that seem problematic - just like the Bots do a good work at reverting vandalism. Diego (talk) 17:41, 23 January 2013 (UTC)
 * 3) Support. The presence of unmoderated comments does not cause any serious problems. They are not displayed as part of the article. People can moderate comments for a particular article if and when they become interested in the comments for that article. Dcoetzee 00:44, 28 January 2013 (UTC)
 * 4) support FeydHuxtable (talk) 18:32, 30 January 2013 (UTC)
 * 5) Powers T 14:12, 6 February 2013 (UTC)
 * 6) I think readers should be able to comment on all articles; but if so, why not simply channel their comments to the talk page, where they actually would have a chance to be read? --<sub style="border:1px solid #228B22;padding:1px;">Piotr Konieczny aka Prokonsul Piotrus&#124; reply here  18:11, 9 February 2013 (UTC)
 * Comments

You're not concerned about Wikipedia hosting publicly visible libel or vandalism for months or years? You make a very reasonable point regarding feedback and the lack of a deadline, but in some ways&mdash;and I say this with all due respect&mdash;your view makes me think you're operating with a romanticized view of the feedback that's actually being collected. --MZMcBride (talk) 04:05, 18 January 2013 (UTC)
 * Your not concerned with it in articles? Regards, Sun Creator(talk) 02:51, 23 January 2013 (UTC)
 * Edits to articles are patrolled regularly. Article feedback is not. --MZMcBride (talk) 05:12, 23 January 2013 (UTC)
 * So then WikiMedia must simply commission a dedicated staff (paid or unpaid, or a mix of both) for ensuring feedback is clean and 'good' feedback is highlighted and selected to be answered by either that staff or someone more knowledgeable. --70.50.222.112 (talk) 17:53, 18 February 2013 (UTC)
 * paying people to walk through the tripe found in the feedback is an absolute waste of Wikimedia funds. If I ever found they were flushing money down the toilet for that, I would NEVER EVER give again and would seriously consider whether I would even donate any more of my time. -- TRPoD aka The Red Pen of Doom  19:25, 18 February 2013 (UTC)
 * We wouldn't legally be able to and wouldn't consider it ethically. Can we save caps lock for things legitimately being considered? Okeyes (WMF) (talk) 20:51, 18 February 2013 (UTC)
 * Indeed this wasn't worth caps lock, but it's not like paying staff for tagging/flagging is so unlikely an idea for the WMF: Jimbo Wales said that it's entirely normal for Wikia and should be so for WMF too, «[//en.wikipedia.org/w/index.php?title=User:Jimbo_Wales/Personal_Image_Filter&oldid=502485105 A few hours a week by a staffer to look at new categories or requested changes is no big deal]». --Nemo 21:22, 18 February 2013 (UTC)
 * The image filter is a different situation, insofar as it isn't gauging and checking for the legality of content; Jimbo and Wikia evidently have very different lawyers to us. Okeyes (WMF) (talk) 03:18, 19 February 2013 (UTC)
 * Not particularly romanticised; I just don't see a big difference between this and vandalism to or libel in the article itself. Vandalism to the article itself is visible to everyone easily, unlike feedback, which you can't find as easily if you don't know where to look.  I don't particularly see why we should trust Internet canines to edit our articles but not to leave comments on them.  Nyttend (talk) 04:52, 18 January 2013 (UTC)  Are we really supposed to start commenting before the start of the RFC?  If so, what's the difference between an opened and a not-yet-opened RFC?  Not objecting to having this discussion; I just wonder if we should move it to the talk page.
 * No idea. The only difference is that there is a little "comment period" before endorsing actually begins, which I think is a good thing. Legoktm (talk) 09:29, 18 January 2013 (UTC)
 * The idea behind a drafting period was to prevent the possibility of first-mover advantage. It helps to have a few views for people to compare and digest before endorsing/supporting/voting gets started. I think comments during this time are more than fine: they allow feedback on the views so that the views can be adjusted prior to endorsing. This is much better than attempting to rewrite views after people have signed on to them, I think. --MZMcBride (talk) 15:01, 18 January 2013 (UTC)
 * MZMcBride and I are in compete disagreement on the above. My view is that the RfC drafting period is for drafting: - getting comments on the wording of the RfC. In my considered opinion, MZMcBride's practice of accepting views on the question the RfC asks before the RfC opens gives the authors of those views a considerable first-mover advantage and increases the bias. --Guy Macon (talk) 04:33, 19 February 2013 (UTC)

The feedback tool's current design makes it less likely that feedback will be read and responded to. Feedback is less visible to editors than an edit to the article or its talk page (feedback on an article does not show up in editors' watch pages, and is visible only if they go looking for feedback). It is thus more likely to accumulate inappropriate material. --Srleffler (talk) 04:04, 21 January 2013 (UTC)

View by Aurochs
No, we do not have enough resources to moderate and respond to all the feedback. We don't have enough resources to edit articles either. Wikipedia is a volunteer project, and no initiative we roll out will ever see 100% utilization. However, AFT5 can be seen as an attempt to deal with the resources we have more effectively, because it will help us focus our editing resources where they actually matter for readers.

In addition to the possibility of disputes between individual editors that AgnosticAphid pointed out above, I'm also concerned that allowing at-will opt-out for AFT5 will provide yet another place for WikiProjects to bicker with each other and individual editors, and it will confuse the user experience for the people who leave feedback. ("This article doesn't have a feedback box but all the other articles I've read today do. WTF is going on here?" That seems like a crummy thing to do to our readers based on one editor's personal preference.) It also sounds like an easy way for people to vandalize articles. If the feedback tool is implemented as a patchwork only on certain articles, I fear that editors will become less driven to read and moderate it, because it's not always there for them to moderate, and they may have to get into fights with other editors just to receive it. Finally, we cannot use AFT5 to improve an article that has been blacklisted.

For these reasons, I believe that AFT5 should be rolled out 100% to all mainspace articles. There is already a protection mechanism built into the tool (see Okeyes' view below), so if necessary, it could be disabled on an exceptional basis, but this should be used only when a clear set of objective criteria have been met.

I agree that feedback comments should never appear on the article page, and the feedback box probably doesn't need to be made any more conspicuous than it already is.


 * Users who endorse this view
 * 1) Aurochs (Talk | Block) 03:30, 21 January 2013 (UTC)
 * 2) -- Guerillero  &#124;  My Talk  00:38, 22 January 2013 (UTC)
 * 3) So what if nobody "responds to" feedback (in ways that you can see)?   The same thing happens on talk pages now, and even in articles when newbies add their questions or opinions directly to the mainspace.  I've found the feedback helpful.  If editors at other articles choose to ignore it, then that's their choice. WhatamIdoing (talk) 19:13, 22 January 2013 (UTC)
 * 4) Graeme Bartlett (talk) 20:20, 22 January 2013 (UTC)
 * 5) I find the arguments made against this tool to be unconvincing. <font color="DarkGreen">AgnosticAphid  talk 01:02, 23 January 2013 (UTC)
 * 6) Support. Abyssal (talk) 04:47, 23 January 2013 (UTC)
 * 7) Diego (talk) 17:44, 23 January 2013 (UTC)
 * 8) Support. Also support WhatamIdoing's comment. The feedback helps me become more aware of who reads the articles I work on (a lot of students!), and encourages me to make articles more reader-friendly. I cheerfully confess, however, that I haven't gotten into the habit yet of using all the bells and whistles (flagging and such)—"habit" being the key word. Cynwolfe (talk) 00:12, 24 January 2013 (UTC)
 * 9) --Noleander (talk) 15:40, 25 January 2013 (UTC)
 * 10) Support. Along with Nyttend's view above. The presence of unmoderated comments does not cause any serious problems. Dcoetzee 00:44, 28 January 2013 (UTC)
 * 11) support FeydHuxtable (talk) 18:32, 30 January 2013 (UTC)
 * 12) Support.  bibliomaniac 1  5  02:19, 7 February 2013 (UTC)
 * 13) the wub  "?!"  12:08, 13 February 2013 (UTC)

How will it allow us to focus our editing resources ? In order to do that, you need to process the data, and for that you need resources. So at some point you need to divert human resources, for which benefit ? It would have to be analyzed. From what I see, not one in fifteen comments are potentially helpful with respect to the stated objective of readers to contribute productively to building the encyclopedia (even though it's already filtered). Does the benefit of having this information outweigh the cost of processing it ? Considering that most of the suggestions provided would be evident to any minimally experienced editor ('not enough information'), or unfeasible for policy or practical reasons, I have doubts. Of the rest that can be helpful, I'm afraid there is a way too small probability that the request can get to someone who would be able to fulfill it. Of course some bots could make some basic processing, for example when 'image' or 'picture' is in a feedback (as often), it could somehow notify requested pictures because it's very likely that the reader wants one in the article. So it would allow a little focusing of resources in this instance, but I don't see much other examples. And then it would be easier to just have a 'request picture' function. Cenarium (talk) 06:03, 20 January 2013 (UTC)
 * Comments
 * First, it doesn't take very much time or effort to read the feedback and come to a conclusion about how the readers feel about an article. These aren't PhD dissertations, they're short comments. So no, I don't think the cost of processing it is very high. Second, the suggestions provided by feedbackers are not always "self-evident" - it's not always obvious to us as editors that a particular article needs more images or that it's not detailed enough. I feel like you're really dramatically underestimating the proportion of helpful feedback. I have personally made a number of article edits and even proposed a move based on information posted to feedback. The feedback tool also gives us a better idea of who our readers are, which is something we haven't ever really known (and which is very useful for writing great prose). Finally, editors can browse the list of all feedback and find something they can fix themselves, much the same way people currently monitor recent changes to spot vandalism. --Aurochs (Talk | Block) 16:43, 20 January 2013 (UTC)

The "crummy" user experience you describe is just the status quo: The feedback box only appears on selected articles at present. If feedback were opt-in, implemented only when there is an editor willing to monitor it, the user experience would not be any worse than it is now. The user experience would be improved, since any feedback they leave would be read and responded to if approriate, compared to the present really crummy user experience, which is that in most cases feedback receives no response at all, and may well not be read by anyone. --Srleffler (talk) 04:11, 21 January 2013 (UTC)
 * I will comment on this in your own view below, in order to keep the conversation together. --Aurochs (Talk | Block) 04:41, 21 January 2013 (UTC)
 * Actually, you didn't mention this in your view, so I'll respond to it here. Here's the problem I see with the opt-in proposal: Suppose that a reader leaves feedback on an article. Nobody with that article on their watchlist pays attention to feedback, so it languishes for a few months. Then an outside editor comes across the article, thinks they may be interested in editing it, and checks the feedback listing. S/he finds the feedback left months before, and uses it to improve the article. This is not possible under the opt-in system - that feedback will simply not be left, and the editor will have nothing to go by but the seat of his pants. Further, AFT5 can be an avenue for article discovery, but not if an opt-in system is used. --Aurochs (Talk | Block) 04:49, 21 January 2013 (UTC)

View by Mike Cline
I talked at length last year at Wikimania with the lead of this project and came away feeling there was a tremendous upside potential for this tool. Eventually it will become a tremendous source of data on how the reader community (10,000X larger than the editor community) views wikipedia. So the benefit will accrue long-term to not only article improvement, but to improvements in policies and guidelines based on emphirical data from the reader community. Asthetics and resource issues aside, in no way should there be an opt-out for any article, except under extreme and compelling circumstances. The data this tool generates in the future should be based on the holistic locus of WP articles. --Mike Cline (talk) 16:49, 18 January 2013 (UTC)


 * Users who endorse this view
 * 1) Considering what has occurred with adding revision reviews, I believe that adding this to all article would be an overall benefit to Wikipedia.  --Super Goku V (talk) 01:58, 21 January 2013 (UTC)
 * 2) This is another side of my own view posted above, so I, of course, endorse it. --Aurochs (Talk | Block) 03:09, 21 January 2013 (UTC)
 * 3) Apteva (talk) 22:06, 21 January 2013 (UTC)
 * 4) -- Guerillero  &#124;  My Talk  00:37, 22 January 2013 (UTC)
 * 5) Danger High voltage! 03:07, 22 January 2013 (UTC)
 * 6) I've re-named articles, re-written sections and added a new article based on feedback, mostly from IPs—who are our main reader base. I check once a day maximum, which isn't onerous, and respond where applicable. Is there not a way for individual users to set their preferences so as not to see the tool, rather than having it removed from articles? - SchroCat (talk) 05:42, 22 January 2013 (UTC)
 * 7) * FYI, that preference already exists, in the "Appearance" section of user preferences.--Eloquence* 08:52, 22 January 2013 (UTC)
 * 8) **Many thanks for clarifying. I didn't know that, and I wonder how many people who are so set against it also didn't know they could make it all disappear for them not to worry about. Thanks again. - SchroCat (talk) 09:04, 22 January 2013 (UTC)
 * 9) ***Making it, and the vandalism involved in it, disappear from my view of the pedia doesn’t stop me worrying about the vandalism that others will see.  Ϣere Spiel  Chequers 13:44, 25 January 2013 (UTC)
 * 10) I've made many changes based solely on feedback comments.  WhatamIdoing (talk) 19:14, 22 January 2013 (UTC)
 * 11) Graeme Bartlett (talk) 20:21, 22 January 2013 (UTC)
 * 12) Shiftchange (talk) 01:03, 23 January 2013 (UTC)
 * 13) everyone's voting for multiple views, so I guess I will too.  I don't see how the tool is harmful and it's a vast improvement over the AFTv4 (ratings box) alternative. <font color="DarkGreen">AgnosticAphid  talk 01:06, 23 January 2013 (UTC)
 * 14) Support. Abyssal (talk) 04:48, 23 January 2013 (UTC)
 * 15) Support. The data is valuable in the short term and will be even more valuable for long-term research and analysis. Dcoetzee 00:46, 28 January 2013 (UTC)
 * 16) Support &bull; Jesse V.(talk) 16:26, 24 January 2013 (UTC)
 * 17) --Noleander (talk) 15:41, 25 January 2013 (UTC)
 * 18) Strong support FeydHuxtable (talk) 18:32, 30 January 2013 (UTC)
 * 19)  Wizardman  20:34, 3 February 2013 (UTC)
 * 20) Support, even so-called "useless" comments reveal weaknesses in the articles. I want the Article Feedback Tool to remain in some form.  Abductive  (reasoning) 04:23, 5 February 2013 (UTC)
 * 21) Good point with the data. You can't have too much data, and this is a fair reason for why this should be rolled out for all articles: eventually some scholars will be able to process the data and offer us some conclusions and advice. --<sub style="border:1px solid #228B22;padding:1px;">Piotr Konieczny aka Prokonsul Piotrus&#124; reply here  18:12, 9 February 2013 (UTC)
 * actually: garbage in garbage out - unless you have good, actionable, data, more is WORSE. -- TRPoD aka The Red Pen of Doom  20:15, 9 February 2013 (UTC)
 * The 'GIGO' saying is about 50 years old. There have been advances in computer technology since then. One of the leading strands of the emergent "big data" field is the ability to draw meaningful conclusions from high volume, low quality unstructured data. Since Lehman, a large portion of the ultra bright young developers who used to enter the finance sector are now working on this tech. Theres even off the shelf analytical packages being released, though granted even the simplest gear needs advanced quantitative skills. But as long as the Foundation is headed by someone as inspirational as Sue Gardner, there'd be no problem attracting the necessary talent, if it's not already on hand. Some seem to oppose movements like semantic web (for our purposes, info boxes, Cats, etc) as they think it makes it easier for Corps to make money from the web. In reality they only hurt civic society orgs and low budget accademics, the big players dont need any structure at all. AFT data is relatively good quality in comparison to what these guys can mine. A decade down the line, there will probably be freeware packages for this. As Colonel Cline and Pan Konieczny rightly suggest, squandering AfT data would be a shame. FeydHuxtable (talk) 10:37, 11 February 2013 (UTC)
 * If Wikipedia doesnt have the resources to make an actually user friendly talk page interface, it is far far far from being able to "big data" any actually useful information from the feedback that would be able help improve the encyclopedia. -- TRPoD aka The Red Pen of Doom  21:30, 11 February 2013 (UTC)
 * 1) Support. N2e (talk) 14:27, 18 February 2013 (UTC)
 * 2) Support, per Piotr and Feyd.  As I said in, WMF needs to work on better aggregation (data mining) algorithms for the "general feedback." Reliance on diversion of busy editors to feedback "moderation" tasks isn't going to fly on most articles. Though as Wikipedia's "Hardware" guy, I'd like to see the tool placed on that disambiguation to see if readers are getting sent where they need to go, or if any are still getting stuck, not finding the specific type of hardware they're looking for.  I volunteer to moderate that one.  And, can you put the tool on Brand New? Reader feedback just might help determine consensus at Talk:Brand New (disambiguation). Thanks, Wbm1058 (talk) 12:38, 19 February 2013 (UTC)
 * 3) Support I've found some of the feedback quite helpful and specific. Feedback that was not forthcoming from elsewhere. I see no reason not to continue on with this program and tweak it a bit more to inform those who just want to complain how to constructively provide feedback and contribute. If we are an encyclopedia for all then we should act like it and take in criticism in ways that are manageable and useful. Insomesia (talk) 08:22, 20 February 2013 (UTC)


 * Comments
 * 1) gbawden The flip side is that a lot of the feedback on some of the pages I contribute to is for directory type enquiries - for forms, contact details etc. Everything wikipedia is not. Perhaps its just that the South African user community don't understand what Wikipedia is for. Gbawden (talk) 06:29, 22 January 2013 (UTC)
 * I've found the feedback enlightening. I've seen many a comment that showed the person thought they were talking to the company. People also as questions as if this was Yahoo Answers. But some people have pointed out vandalism, factual errors and errors of omission. Abductive  (reasoning) 09:20, 22 January 2013 (UTC)
 * I personally have considered proposing policy reforms on the basis of such feedback. I see no reason not to include information like permanent street addresses. Dcoetzee 00:48, 28 January 2013 (UTC)
 * FWIW, this is not unique to AFT5 -- OTRS gets a ton of email from people who think they are somehow contacting the company as well. People are ... confused when it comes to directories, which I think will be the case whether or not we have AFT turned on. -- phoebe / (talk to me) 02:08, 6 February 2013 (UTC)
 * 1) Comment - "A tremendous source of data" is not equal to "a tremendous source of useful data" : ) - jc37 23:19, 25 January 2013 (UTC)
 * 2) Just out of curiosity, what kind of information and data are you expecting to extract from all that drivel? What would you do with it as soon as you have it? And why would you need it in the first place? Sorry, but my imagination fails me when I have to think about anything terabytes full of spam and nonsense are useful for. I mean this is an encyclopedia, no? At least in my eyes it's only point of existence is to provide people with a vast amount of information that is correct and up-to-date. But when reading all those comments about data mining it looks to me as if there's a despererate attempt going on to turn it into some sort of Facebook, where the management is constantly trying to come up with new ways to increase their customer base and shareholder value, because otherwise they might dwindle into insignificance and go bankrupt. (Thusz (talk) 20:13, 11 February 2013 (UTC))
 * So, anybody bothering enough to explain to me, what kind of data you want to mine from all those AFT-comments? I mean anything more insightful than the fact that a lot of people who are using Wikipedia appear to be excruciatingly dumb; or that pictures for everything would be a big crowd-pleaser. (Thusz (talk) 18:01, 19 February 2013 (UTC))


 * 1) Comment - The information I have gleaned from about an hour of going through feedback on articles on my watchlist and the conclusions I draw from it are as follows:
 * There was not one item about which I could do anything immediately useful.
 * Most positive comments are not even potentially useful for improving an article.
 * Many of our readers appear to be functionally illiterate.
 * Almost nobody who provides feedback has the slightest idea of how to provide usable feedback, but that is not going to stop them from saying something.
 * Probably the majority of feedback providers appear to have no idea of how Wikipedia works.
 * Going through the feedback is not a productive use of my time.
 * Comments are provided with totally inadequate context.
 * Most of the feedback was quite old. Checking whether the comments were still valid in the absence of sufficient context would take more effort than simply reading the article and fixing what one can at the time.
 * Much of the feedback indicates that the user has either not read the article or does not understand what is already there.
 * Many readers do not seem to realise that the blue words are links,and expect items already linked in the text to be listed in "see also".
 * Requests for images are frequent, which does not help if no suitable images are available.
 * A significant amount of feedback appears to be on pages which have nothing to do with the subject matter in the feedback.
 * The feedback request should provide more information on how it should be used, either directly or through a single step link which is labelled very clearly.
 * YMMV &bull; &bull; &bull; Peter (Southwood) (talk): 10:22, 12 February 2013 (UTC)
 * Sums up my feelings pretty well....get rid of Feedback altogether. Isn't that what the talk page is for anyway?<font color="DarkSeaGreen">Geremy Hebert <font color="MidnightBlue"> (talk &#124; contribs) 19:25, 23 February 2013 (UTC)


 * i dont know that you can extrapolate to "readers" being functionally illiterate, just those who leave feedback. -- TRPoD aka The Red Pen of Doom  12:17, 12 February 2013 (UTC)

View By Beeblebrox
This RFC came up in some comments on the functionaries mailing list. i cannot claim to speak for the entire oversight team in this matter, but I am concerned that there will be a flood of unactionable requests for oversight if this tool is brought into wider service. The reason I have this concern is that we regularly get such requests now with just the limited deployment. The only fix I have seen floated for this is a whole new user right just for suppressing feedback I personally think that is a solution that completely ignores the nature of the problem. The suppression policy applies equally to feedback and actual content, so those entrusted with this new right would still need to be vetted as oversighters. I would like the policy for requesting suppression of feedback to be very, very clear that suppression is warranted only for a very small minority of edits, and that most kinds of vandalism, including BLP violations, do not qualify. Education is the only way to resolve this issue. I have been on a bit of a campaign already to personally speak to users who make invallid requests to suppress feedback and feel it has put a dent in the number of bad requests we get, but I'm only one person and cannot hope to keep up with personally notifying every last person who makes an invalid request if the use of feedback tool is expanded.

TLDR version Feedback reviewers need to have it made very clear to them that suppression is not for "normal" vandalism. Beeblebrox (talk) 21:52, 20 January 2013 (UTC)


 * Users who endorse this view


 * 1) Beeblebrox (talk) 03:27, 21 January 2013 (UTC)
 * 2) Agreed. Any usage of revdel, suppression, etc should be exactly the same as if they had edited a talk page. Legoktm (talk) 03:31, 21 January 2013 (UTC)
 * 3) If applicability of tool usage needs to be clarified, then let's clarify it : ) - jc37 23:19, 25 January 2013 (UTC)
 * 4) support FeydHuxtable (talk) 18:32, 30 January 2013 (UTC)
 * 5) Support. The admins who gave me reviewer and rollbacker rights did so after considering whether I need them and can be trusted with them. Now suddenly I have Oversight permissions, but only for article feedback. Who vetted me for the additional user right? Nobody did. --Guy Macon (talk) 14:15, 2 February 2013 (UTC)

Discussion

 * Why do we need an RfC for this? Whereas this is a clear and valid statement, is not it best done by bot posting on talk pages for instance? Or do you intend to change smth in the policy?--Ymblanter (talk) 22:32, 20 January 2013 (UTC)
 * I intend to see if there is a consensus that education on this issue is an imperative part of expanding use of the tool. As far as I can tell this is a general RFC on the use of AFT, I am not aware of any restriction that says all views must propose a change in policy. Beeblebrox (talk) 03:27, 21 January 2013 (UTC)
 * Given that there has never been any formal policy on AFT, I think this fits in with the scope of the RFC. Legoktm (talk) 03:32, 21 January 2013 (UTC)
 * Commenting briefly; I think Max intends for this to be an RfC covering the scope of the deployment of AFT5. I'm always happy to receive feedback (you'd note that last time the functionaries brought this problem up with us, we invested developer resources in educating editors about what 'oversight' was for) but splitting the RfC off 15 ways may lead to a no-overall-consensus thing, or, a 15-different-consensuses thing :P. Okeyes (WMF) (talk) 03:34, 21 January 2013 (UTC)
 * I do not mind discussing it here, it just seems to me that the issue is already covered by existing policies (suppression policy), and should just be enforced.--Ymblanter (talk) 08:01, 21 January 2013 (UTC)


 * A quick comment here. (I do think this is somewhat relevant to the RFC as it is another manner in which volunteer resources are being used by the AFT5 project.) I took a look at the last 45 days of suppression requests that came in via email to the Oversight OTRS queue; I also did a sample of 45 days prior to when the software was set up for our oversighters to handle suppression requests from AFT5, which automatically sends an email to the OTRS queue. There was no statistically significant increase in requests submitted via OTRS (bearing in mind that there is a +/- 15% variation in requests month over month); in the last 45 days, approximately 28% of requests originated from AFT5. Statistical data including frequency of suppressions is available here with prior months available in the page history, and does not show any statistically significant change in the frequency of suppressions over the past two years. Based on this limited sample, it does not appear that including AFT5 on 10% of the articles has had a significant effect on the number of *emailed* suppression requests, or the frequency of suppressions. It is important to note that an unquantifiable percentage of suppression requests are made outside of the OTRS system, and that the average number of suppressions has not significantly changed. It's very difficult to give a full picture of the workload of oversighters, because a significant portion of emailed requests in particular result in multiple suppressions, while an equally significant percentage of requests (from all sources) result in no suppressions. It is also very difficult to predict whether or not we will see an increase in suppression requests from AFT5, because those requests are entirely dependent on the moderation of individual comments, and it's not obvious that the number of comments that get moderated will change.  Risker (talk) 08:11, 21 January 2013 (UTC)


 * Thanks for that. Hard data is always nice in a situation like this. Of course your observation that we cannot know what the effect will be is also relevant, and that is why I believe education/raising awareness is our best bet at this time. I don't mean to suggest anything drastic or any new rules or anything, just to raise awareness of this issue in the hope that we it will continue to be a small problem rather than a large one. Beeblebrox (talk) 01:12, 22 January 2013 (UTC)

View By GregJackP
The tool is useless. Most of the comments are not even actionable. Additionally, wading through the inane comments to get to one to work on wastes time, and we don't have enough resources to handle it as it, much less by expanding it. We should eliminate the feature.


 * Users who endorse this view
 * 1)  GregJackP   Boomer!   02:16, 21 January 2013 (UTC), as proposer.
 * 2) Strongly support turning off AFT in any and all forms. It does not engage readers; rather, all the vandals seem to just enter random nonsense into the comments field. Reaper Eternal (talk) 03:29, 21 January 2013 (UTC)
 * 3) "Useless" is a little strong. I would've recommended this as a view, but I felt it wasn't tenable (politically, I suppose). That said, I can support this option. --MZMcBride (talk) 03:42, 21 January 2013 (UTC)
 * 4) Strong support - OK, "next to useless" instead of "useless, perhaps, but whatever description is used, it's an unnecessary burden on our resources. Beyond My Ken (talk) 04:49, 21 January 2013 (UTC)
 * 5) Strong Support it should be eliminated or, at the very most, used on rare occasion to invite public comment on contentious articles. It certainly hasn't been shown to be helpful so far, and if it does 'take off' so to speak who's going to volunteer to sift through the litterbox to get rid of all the spam, BLP violations, etc? Andrew Lenahan -  St ar bli nd  05:28, 21 January 2013 (UTC)
 * 6) Strong support. This extension should never have been deployed (or perhaps even created) in the first place. We already have a feedback tool with effective moderation tools and a sufficient barrier to entry to keep the crap out -- it's called the talk page. The developers should have anticipated the massive crapflood coming; it was visible from a mile away. MER-C 09:46, 21 January 2013 (UTC)
 * 7) Support Nick-D (talk) 10:05, 21 January 2013 (UTC)
 * 8) Support I would not call it useless (it was an interesting experiment), however I agree with the general sentiment of the view. Legoktm (talk) 10:36, 21 January 2013 (UTC)
 * 9) Yes. I'm reading the feedback to all articles I watch, and 99% of them either state the obvious, are otherwise unhelpful, or were written by idiots. Sample entries from my watchlist: "Thisted article nejedes a photo" [sic], "what can leeks be put in" (concerning Leak), "stinks need more info", or "it does not tell me how meat is produced or i did not see what i read" (concerning Meat, which does address production in depth). Sorry, we don't need feedback of that sort, and we don't have the capacity to police it for spam, BLP etc.  Sandstein   12:08, 21 January 2013 (UTC)
 * 10) Regretfully support I can see what they were going for, apparently 90% of our traffic is just page views so it would be fantastic to utilise this huge volume of traffic to help improve articles and recruit new editors. Unfortunately, like most of those in this section the feedback I've seen has been unhelpful at best. I had to unwatch Help:Searching because the feedback is flooded with unrelated comments, such as was shown in previous feedback implementations (though of course the talk page there suffers too). We need new editors, but the Q4 report retention stats are underwhelming so maybe it's best not to burden existing editors with these comments and go back to the drawing board. <font color=#000000>Jebus989 ✰ 13:32, 21 January 2013 (UTC)
 * 11) Yes, please kill it with fire.  Salvio  Let's talk about it! 15:55, 21 January 2013 (UTC)
 * 12) Support. "Useless" is probably a bit strong. I'd call it something closer to "Not paying off in proportion to the amount of effort that has to be put into it". It was a really interesting initiative, and I appreciate the sheer amount of developer time and community engagement effort (oh god, if only they did that for all projects!) the Foundation has put into this, but the bottom line is that no matter what we call it, I don't think it's benefiting Wikipedia when all factors are considered. There is far too much chaff and not enough wheat, and at any rate we have almost no threshers to distinguish the two. Even when users leave useful, actionable feedback - which seems to be the exception rather than the rule - it's generally either not noticed in the flood of less useful stuff, or not actioned because most editors don't take requests (maybe we should...but mostly we don't). A fluffernutter is a sandwich! (talk) 16:30, 21 January 2013 (UTC)
 * 13) Regretful support. While in principle this tool sounds like a good idea, in practice it is pretty useless. I've spent hours sifting through and hiding abusive feedback on three hotly contested articles, and am still nowhere near halfway done hiding all the abuse. While doing this, I have come across maybe one or two comments that were actually constructive, a few that were well-intentioned but could or should not be implemented, and the rest were from people who obviously had not read the article in question. Simply put, I don't see the benefit to Wikipedia with this tool. StringTheory11 (t • c) 17:18, 21 January 2013 (UTC)
 * 14) Support I liked this tool in theory, but I've only seen poor to useless results from it personally. -- Khazar2 (talk) 20:31, 21 January 2013 (UTC)
 * 15) Support. I look at feedback occasionally—mainly the "all articles" feed, since most of the articles on my watchlist are very obscure and don't attract much feedback—and out of the hundreds of instances I've seen, I've taken action (usually not what the commenter wanted, but somehow related) on, I think, three of them. The most common nonvandalistic and comprehensible feedback I've seen is (1) "needs more pictures", usually with regard to an article, like one for a video game or film, in which multiple nonfree images could not be used and free images are impossible to obtain; (2) "needs information about X", where that information is already present in the article, or linked to via a "Main article" or similar link; (3) "needs X", where X is clearly beyond the scope of an encyclopedia; and (4) questions, discussion starters, and the like from IPs, which may be pertinent or potentially helpful but which there's no way of answering or discussing with the poster. Gibberish, tests, and vulgarity, however, outweigh all that by a huge margin. I can't imagine who would want (or be able) to wade through the muck intensively enough to pick out the few actionable nuggets, especially when the use of the tool is extended to all articles. I don't know what kind of reviews the WMF has done on feedback, but I fear they may be viewing the results through rose-colored or distorted lenses. Deor (talk) 21:24, 21 January 2013 (UTC)
 * 16) Support. We have the talkpage for comments and feedback. The talkpage is a more open place, though could be made more inviting by perhaps changing the link tab to say "Comments" rather than "talk" and/or greater use of the talkheader template - or a reader/unregistered user coding that presented readers with reader-only information that explained how to use the talkpage.  SilkTork  <sup style="color:#347C2C;">✔Tea time  23:28, 21 January 2013 (UTC)
 * 17) Support, sadly. It would have been great to get substantive feedback but it's true most of the comments are nonsense (asking for info contained in the article, saying they liked/disliked without providing detail, requesting trivia I can only guess was requested by an elementary school teacher as part of a report, i.e. "needs more facts beginning with M", or full-on spam). Editors can't be burdened with this. The space available for lengthy comments on the talk page seems to help increase the likelihood that the comments there are more thoughtful. If the real issue is how to recruit more editors, there's plenty more that can be done, beginning with a giant button on the main page that links to a Wikipedia for Dummies very simple and straightforward cheatsheet for beginning editors. Recruitment internationally through universities, ESL networks, the foreign service and others will help as well. But this tool hasn't proven as useful as the creators had hoped, and honestly I miss the star system - that told me more about where to focus my efforts to improve the article than this feedback box has, and I don't like that one supplants the other. Lemurbaby (talk) 04:30, 22 January 2013 (UTC)
 * 18) Seldom have I seen feedback that was actually 1. coherent 2. able to be done 3. relevant to the article. As an example, we could include the routing by auto from Atlanta to Salt Lake City (two apparently random cities) in Interstate Highway System like someone who commented would like us to do, but why would we? —Scott5114↗ [EXACT CHANGE ONLY] 04:48, 22 January 2013 (UTC)
 * 19) Support - although I disagree that its useless. But useless describes the feedback on a lot of the feedback on some of the pages I contribute to is for directory type enquiries - for forms, contact details etc. Everything wikipedia is not. By including the feedback form we give users the impression that we will follow up and answer their feedback. Gbawden (talk) 06:32, 22 January 2013 (UTC)
 * 20) Support. My experience is that feedback predominantly consists of inappropriate requests to add information that doesn't fall within the scope of the article; also the system is hard to integrate with a watchlist. I don't understand why readers who want to leave feedback can't be directed towards the talk page, perhaps by a button at the bottom of the page that directs them to the talk page in edit mode. Espresso Addict (talk) 16:39, 22 January 2013 (UTC)
 * 21) Strong Support By and large, I find the feedback tool to be more of a nuisance than a help to editors who actually do the grunt work of sourcing, rewriting and maintaining articles. Most of the comments are incomprehensible, offensive and (on a good day) unhelpful. Admittedly, a few of the comments I have come across have been helpful but no more helpful than someone who took the time to go to the article talk page and drop a note. I guess the idea behind the feedback tool was to get people who wouldn't normally edit Wikipedia more involved in the process. It actually does the opposite - it just gives people who normally wouldn't edit or leave a helpful comment on the talk page a more easily accessible platform to complain about stuff we don't include in articles anyway (DVD prices, where to buy stuff, links to copyrighted material, addresses to celebrities' homes (?), etc.) The idea behind the tool is novel but the reality of it is that it mainly attracts people who have no intention of bettering the project - they just wanna complain that an online encyclopedia isn't Amazon, YouTube or Flickr. It's a waste of time and resources in my opinion.  Pinkadelica <sup style="color:black;">♣  21:10, 22 January 2013 (UTC)
 * 22) Strong Support. It's like reading through your spam box after you've got it set up properly. I won't be delving in this stuff again. Aarghdvaark (talk) 07:29, 23 January 2013 (UTC)
 * - Actually I did go back in and trawl through all the stuff: no swearing etc., mostly irrelevant, but picked up one link that was actually very pertinent for John Archibald Wheeler Aarghdvaark (talk) 14:09, 23 January 2013 (UTC)
 * 1) Support, the few useful things to come out of it don't justify the endless garbage it produces. Perhaps a method that people "think" they use the feedback tool, but their feedback actually goes straight to the article talk page instead where it (the comment and if necessary the editor) can be dealt with like we normally do? That way we don't have to checl even more pages to find all comments, and can rely on our normal watchlist instead. If that isn't a realistic option, just delete the whole damn thing. Fram (talk) 08:51, 23 January 2013 (UTC)
 * 2) Support. Sadly true. Editors' time would be better spent editing articles than reading the "feedback". I have stopped wasting my time on the matter.  Axl  ¤  [Talk]  10:54, 23 January 2013 (UTC)
 * 3) Strong Support I have seen only one comment on any of the articles I watch that was helpful. Most of the people using this feature don't understand the purpose of wikipedia or are using it as an additional 'thing to vandalize / draw penises on / write 'faggot' twelve times in'. -- T  K  K  bark  !  12:36, 23 January 2013 (UTC)
 * 4) Strong support, I realise editor retention is a problem, but adding this box which looks like the comment box on youtube and produces roughly the same calibre of content is not the solution. It just serves as a drive by invitation for lazy drive by readers who never even read the whole article (and this shows in their feedback) and would no way think of posting on a talk page. <font face="trebuchet MS">- filelake<font color="#0B0">shoe  <font color="#0B0"> &#xF0F6;    14:03, 23 January 2013 (UTC)
 * 5) Strong support, for all the reasons given above. --Jorge Stolfi (talk) 16:14, 23 January 2013 (UTC)
 * 6) I could live with MZM's compromise, but my preference as Salvio said is "Kill it with Fire". As I and others pointed out from the start this misconceived project was never going to work unless the WMF was prepared to hire a whole load of mods to deal with the feedback. At best it was an unfortunate exacerbation of the trend from wp:SoFixIt to "so tag it for others to fix" that has done so much damage to this project since 2007. It should have been killed long ago - certainly once we knew that a "call to edit" was more effective at getting extra edits.  Ϣere  Spiel  Chequers 16:27, 23 January 2013 (UTC)
 * 7) Largely agree. There probably is some good feedback from the tool, but it's well hidden. --Michig (talk) 17:16, 23 January 2013 (UTC)
 * 8) Support Most feedback is un-actionable. Many requests for contact information for living people, which is against policy, requests for information already contained in the article, and requests for pictures which are difficult to fulfill due to licensing requirements. Also a lot of comments like "Great Article" and "Needs More Information", which do not help improve the article in anyway. Lastly, this tool seems somewhat redundant given that we have the talk page to discuss improvements to an article. --Hirolovesswords (talk) 17:28, 23 January 2013 (UTC)
 * 9) Agree. This isn't a useful feature and creates unreasonable expectations. I gave up reviewing the feedback from my watchlist because it was just flooded with American political commentary and other meaningless nonsense - meaningless in the context of I'm not American and most of it doesn't actually display any thinking. Spartaz Humbug! 17:33, 23 January 2013 (UTC)
 * 10) Support - I wouldn't go as far as to call it useless, but this is the view I'm most agreeable to. I have closely watched the feedback which has come through with articles on my watchlist, and only an absolutely tiny amount is useful - a lot states the obvious, others make inappropriate suggestions, and others say nothing useful at all. On the whole, this tool fails the cost/benefit analysis on the time needed to manage the tool vs. positive outcomes. I also think it is unfair to readers to ask them to leave feedback knowing that a lot of it won't be read, and almost none will be actioned. I would be happy with replacing the tool with some kind of connection with the talk page - problems would remain on a lot of it being inappropriate, unread e.t.c., but readers who are serious about improving the article would benefit from direct interaction with editors, which the feedback tool doesn't provide. CT Cooper · &#32;talk 18:03, 23 January 2013 (UTC)
 * 11) Support — The tool generates horrible feedback and it takes up lots of space at the bottom of the articles it's on. It diverges the editing process by making two "standard" ways of communicating. No sensible way to interpret the data exists. It's been a bad project from the start. Jason Quinn (talk) 19:06, 23 January 2013 (UTC)
 * 12) Support  This attempt to make an alternative place to discuss article issues detracts from an more effective process for discussing article issues more than helping anything.  To the extent that the existing mechanism for discussing article issues (the Talk page) isn't living up to expectations, perhaps improving that, rather than creating an entirely different, second-class alternative would be the right place to begin.  --j⚛e deckertalk 19:24, 23 January 2013 (UTC)
 * 13) Completely useless, and we'd be much better off simply by turning this feature off. It is a timesink for no gain whatsoever. Courcelles 23:27, 23 January 2013 (UTC)
 * 14) Support The way I see it, there are (at least) five types of feedback posts:
 * 15) Patent nonsense (e.g. "jkjkjkjkjkjkjkjk")
 * 16) Errors (asking for incorrect changes to be made)
 * 17) Unnecessary feedback (correct, but already in the article)
 * 18) Pertinent feedback (correct and not already in the article)
 * 19) General comments (e.g. "This article was great")
 * In my experience, there is a very low proportion of type 4 comments (about the only ones which are really useful). Type 5 is also rare. (I am watching about a thousand pages, and I only monitor feedback for those pages, FYI.) Double sharp (talk) 15:35, 23 January 2013 (UTC)
 * 1) Support In its current form the tool is not useful, and should definitely not be rolled out, or simply removed until fixed.-- ELEKHHT 00:25, 24 January 2013 (UTC)
 * 2) Support. For me, feedback posts are nearly always unhelpful, either being vandalism, impossible requests, and even BLP issues within the feedback comments themselves.  Spencer T♦ C 04:42, 24 January 2013 (UTC)
 * 3) Support – GregJackP's viewpoint pretty wells sums up my experience: it's a timesink with insufficient redeeming benefit. Sasata (talk) 04:45, 24 January 2013 (UTC)
 * 4) Support I'm afraid I have to agree. This feature could be removed and I'd neither notice nor care. --BDD (talk) 05:55, 24 January 2013 (UTC)
 * 5) Strong support – I had great faith in this tool's potential, but most comments are unhelpful, and my suggestions for improving it were hardly given serious consideration. If a reader has something to contribute, they will either post on the talk page or edit an article themselves. (That's what I did [and why I joined WP], and it's what many readers do every day.) Comments like "this article needs more pictures" are only overloading WP's servers. Toccata quarta (talk) 12:28, 24 January 2013 (UTC)
 * 6) Support - I think it's best to either get rid of this tool entirely, or at least to go the way MZMcBride has proposed and include it only in articles where somebody has volunteered to sift through the flood of garbage for an occasional nugget, because as it stands pretty much all of the comments left for us to attend to are a waste of time. And in all honesty, it might be an honorable endeavour to find ways to get more people to contribute, but ask yourselves, looking at the tons of shit they are posting, do we really want these people to edit? With all due respect, but 99% of those people using the article feedback tool don't appear to be mentally capable to contribute to Wikipedia in a positive way. I mean, just look at the comments on the Sandy Hook shooting - at least every other person wants to tell you that the Bushmaster was not used, but was left in the trunk of Lanza's car. Do we really want these people here? People who can't even get the facts straight when they are right before their eyes? I think we are better off without them, and anybody who is intelligent enough and willing to contribute, will take his or her time and learn how to edit properly. Yes, editing Wikipedia is nothing you can learn in an afternoon, and you need some perseverance to master it, but instead of constantly nagging that this discourages people from participating, we should also see the positive side - it keeps a lot of the idiots away and therefore limits the amount of vandalism and disruptive behaviour we have to deal with. That said, if AFT5 is destined to stay, at least do us the favour to delete every comment that has not been flagged as helpful after a week or two, just to keep their number managable. There's no need to preserve every trifle for posterity. (Thusz (talk) 18:15, 24 January 2013 (UTC))
 * 7) Support. Finding a useful comment in the flood of inane, rude, nonsense or otherwise useless ones is a waste of time.  -- Ed (Edgar181) 20:29, 24 January 2013 (UTC)
 * 8) support except for the "useless" description. "almost entirely useless and no evidence shown that it is actually providing any of the anticipated benefits". -- TRPoD aka The Red Pen of Doom  20:56, 24 January 2013 (UTC)
 * 9) Support. Entirely pointless. A Kafkaesque waste of time. GRAPPLE   X  00:39, 25 January 2013 (UTC)
 * 10) Support, except, as TRPoD (and others) has already pointed out, useless may be too strong; it's conceivable there are uses for the tool, but all comments in support of it are unconvincing so far. Cheers, LindsayHello 09:12, 25 January 2013 (UTC)
 * 11) Support - I support his more than anything else in this RfC. I especially don't like the bloggish sense that this seems to reinforce. Something we at Wikipedia work very hard to not have go on. Our goal here is to build an encyclopedia, not to blog about an encyclopedia. - jc37 23:19, 25 January 2013 (UTC)
 * 12) Support in reality, whatever issues contributors to Article Feedback are bringing forth, an editor with some experience would pick up on just the same if they really wanted to improve the article. Albacore (talk) 03:01, 26 January 2013 (UTC)
 * 13) Strong support. As a very active editor, I know the quality of the articles I edit. I can see for myself if it is well-written and well-sourced, or not (and if I have the energy and motivation to do something about it). If users would like to comment, they can write it on the talk page. Lova Falk   talk  09:01, 26 January 2013 (UTC)
 * 14) Strong support. I said enough about what a drain it was on development resources while it was in the making, and how potentially useless it would be when implemented, so I'm not going to repeat myself here. Kudpung กุดผึ้ง (talk) 10:29, 26 January 2013 (UTC)
 * 15) Support A distraction for readers and of no benefit in my editing. Not even my articles with hundreds or a thousand readers daily have received anything worthwhile, except my own suggestions. <font style="color:blue;background:yellow;">Kiefer <font style="color:blue;">.Wolfowitz 20:27, 26 January 2013 (UTC)
 * 16) Support well-intentioned idea but has not worked out as intended. Mugginsx (talk) 13:06, 27 January 2013 (UTC)
 * 17) Support -- the current feedback system discourages readers from leaving a comment on the talk page (which is much more likely to lead to the feedback being acted upon) and from making bold edits to improve the encyclopedia. It is likely that Wikipedia has failed to attract many potential new editors because of the feedback feature. Celuici (talk) 15:13, 27 January 2013 (UTC)
 * 18) The end result of this tool needs to interact with the talk page rather than a bottomless pit, but clicking a button and writing a thing, without having to learn the Wiki format rules, seems to have a certain appeal for some people. Shii (tock) 04:01, 28 January 2013 (UTC)
 * 19) Support Of about 10 instances of feedback I've responded to, all but 2 were silly and not really helpful to the process of editing or soliciting contact with Wikipedia.  It would be much better if the feedback would solicit feedback on talk page. -- kosboot (talk) 04:47, 28 January 2013 (UTC)
 * 20) Support. The idea behind the tool is a good one—readers should be encouraged to provide useful feedback—but it's pretty much useless as currently implemented. Far better to encourage readers to make highly specific comments, using the talk page. This could be accomplished through a "feedback wizard" that automatically signed their comments and placed them in a new section at the bottom of the talk page. Rivertorch (talk) 10:57, 28 January 2013 (UTC)
 * 21) Support. Any useful comments should be on the article talk page, not in some obscure location. Very little constructive feedback is left; most feedback seems to be written by schoolchildren, often very short and frequently nonsensical. Simon Burchell (talk) 16:54, 28 January 2013 (UTC)
 * 22) Support. I haven't seen a single helpful feedback in the articles I edit. <font style="color:white;background:#167FF7;font-family:sans-serif;">Correct Knowledge <font style="color:#167FF7;background:white;font-family:sans-serif;">«৳alk»  16:59, 28 January 2013 (UTC)
 * 23) Support Less than 10% of the feedback I've seen might potentially be used to improve an article. Of that less-than-10%, none actually [one] has be so used as far as I am aware. --catslash (talk) 21:39, 28 January 2013 (UTC)
 * 24) Sure there is some usefull feedback - there would be some useful feedback even if people had to send it in by (snail-)mail. Sure feedback is good. Don't stop looking for something that works. But this is not it yet. - Nabla (talk) 21:56, 28 January 2013 (UTC)
 * 25) Support. I agree with MZMcBride, "useless" is a bit strong -- the tool is well intentioned and follows from a reasonable hypothesis. But: I've heard several WMF staff explain why it is needed, I have interacted with it at various points, and have tried to see how it can achieve its purpose of helping new contributors get engaged. I think it actually has the opposite effect - it fractures interaction into a new venue away from the talk page. I believe one-off commenting through AFT comes at the expense of (some) new contributors engaging on the talk page, where they can be exposed to the deliberation that is how we have always addressed editorial issues. I think the feature should be turned off; it may be useful at some point in the future, but careful study of the effects should be done before its use is extended (either in time, or in number of articles). I do feel it would be valuable to consider research on its impact before making a final judgment, so I'm open to seeing a report when it is produced. But generally, I agree with Risker's statements here, that the lack of emphasis on how data is being gathered and studied -- even if it *is* being gathered and studied -- is unfortunate, and doesn't do much to help Wikipedians who are (properly) protective of intrusions into article space evaluate the possible benefit. -Pete (talk) 07:47, 29 January 2013 (UTC)
 * 26) Strong support. I have not found this at all useful, and it seems to be a bit of a nuisance. Article readers should be encouraged to leave comments on the talk pages.  If someone addresses their comments, that would encourage them to register as editors. Wahrmund (talk) 17:32, 29 January 2013 (UTC)
 * 27) Strong support. This feature is redundant and, as others already pointed out, people should instead use talk pages where the feedback can actually be discussed. Plus: Reading up on older feedback is much easier on talk pages. Overall, talk pages are the way to go. Philantrop (talk) 18:34, 29 January 2013 (UTC)
 * 28) Support the concept is possibly a good one, not the implementation, nor do I think that any implementation could succeed at this point. we're an encyclopedia, not a talk site, and we should resist the fashionable trend to expand the roles of sites to include social features; the commercial services do it to increase revenue, but an advantage of our manner of workinging is that such things do not affect us.  The commercial services also have the money to pay for moderation of the comments, however poorly they may do it in practice. We do not have that ability. We can just barely keep up with the problems in actual articles, without adding an additional layer. This diverts effort from where it is needed:working on content.  It's normally pretty clear what needs improvement in an article--the problem is doing the work. Let us thank the foundation for their good intentions, and encourage  them to realize that the way to increase editing of content is to decrease layers of supplementary bureaucracy.  DGG ( talk ) 20:43, 29 January 2013 (UTC).
 * 29) Support Any beneficial content is lost in the chaff of the other comments.  This feature also attracts various kinds of vandalism (including "hate mail") and vandalism-type edits, particularly on controversial articles with a low enough profile that they don't have alot of regular Wiki-patrolling. – Philosopher Let us reason together. 23:42, 29 January 2013 (UTC)
 * 30) Support A great place to push one's trolling silliness, propaganda or inaccurate info with no real oversight. If it's important enough to comment on, they can click edit in talk. The ratings have some minimal use if they are either very good or very bad. CarolMooreDC 14:10, 31 January 2013 (UTC)
 * 31) Support Useless and tiresome. We needed this like a hole in the head. I have an eclectic watchlist too, because I have many interests. None of my feedback prompts have turned up anything I was not already aware of. My favorite: "It's typical of all music-related articles. Written by fans for fans. A biased piece of fluff, which is essentially just promotional material." This was not news nor actionable. This all reminds me of the Facebook "like" feature and a box for lame comments. As I've just said in another discussion on how to link to Simple Wikipedia, I do not believe enough people notice the talk page and even if they do, they do not realize they may speak on it. If participation is found lacking; please link whatever to the article talk pages. Many I edit have never been opened.  <font color="#595454">Fylbecatulous  <font color="#967BB6">talk   15:48, 31 January 2013 (UTC)
 * 32) Support I can only echo what Fylbecatulous and Double sharp wrote above. I've seen quite a lot of feedback comments now but only about two or three of them were actually helpful. I would go as far as saying that this tool discourages the uninvolved IP reader from getting involved and actively edit a page themselves. Why should they bother and try to improve an article the way they see it fit, when there's a handy tool that let's the "staff" know what should be done? De728631 (talk) 17:34, 31 January 2013 (UTC)
 * 33) Support I had considered various more eloquent and creative solutions to this problem and could think of none that actually solves the inherent real world problems of this tool - that the vast majority of the feedback is completely worthless. If the tool is low-overhead then I don't really have a problem with keeping it, but maybe just strip out all the other "features" such as the helpful/not tagging and the ratings-nobody uses those. Just make it a comment box - make it clear that it's for suggestions for the article, not comments about the article's subject, and have them all dump to where they're dumping now. But, again, that won't solve the problem that this is a complete waste of time and doing that would be wasting even more people's time on this boondogle that doesn't actually improve wikipedia in any tangible way, except perhaps giving potential vandals a place to express their inner vandal that's less visible. Maybe we should just have comment boxes that go nowhere and see if it reduces vandalism ;) -Drdisque (talk) 18:28, 31 January 2013 (UTC)
 * 34) Strongly Support I've examined the feedback from many pages I edit and almost all of comments that are actually readable and make some sense seem to ask for information that is irrelevant to, or out of the scope of the article. Some examples from the Bread article: 'bread prObleM in egypt', 'i want a cookie' and 'Zucchini cake recipe'. That's just the negative comments, the positive ones basically just say "good job" which doesn't help at all. This feature is a complete waste of time and should be removed immediently. EvilKeyboardCat (talk)
 * 35) Support Armbrust The Homunculus 10:37, 2 February 2013 (UTC)
 * 36) Support What I found most of the time is either the feedback provided by the reader is abusive or out of context, I mean they ask for view point what is better and whats wrong.I didn't found this tool helpful at all. Gaurav Pruthi talk 12:06, 2 February 2013 (UTC)
 * 37) Support. The talk page of each article can be used to track down its quality or the public reaction to it. - Al Lemos (talk) 13:40, 2 February 2013 (UTC)
 * 38) Support The WMF motto sometimes seems to be "Doing whatever we can to alienate our existing user base to get more involvement out of people who don't even want to contribute anything." Wikimedia Foundation doesn't seem to get that for a large group of its users, reading an encyclopedia is all they want out of this site. For many of those, it's all they are capable of. Going to great lengths to attract the kinds of people who only cause more work for those already here and ignoring everything the users who made the site successful in the first place is a recipe for disaster. DreamGuy (talk) 18:58, 3 February 2013 (UTC)
 * I'm confused by your edit summary here. This is apparently being 'forced down your throat' - by an organisation that has pledged to adhere to the outcome of the RfC? Okeyes (WMF) (talk) 08:36, 4 February 2013 (UTC)
 * 1) Support It's useless clutter and makes us look as if our purpose is to curry favour, not to provide a reference work. The old version with the star rankings asks people to rate coverage of something they almost certainly clicked on because they didn't know much about; and it's a bid for popularity. "Like this article, please!!!" (And yet the Foundation couldn't even be arsed to make it work for someone not using a mouse.) The new version goes to the other extreme on motivation, assuming that a focussed search for specific information was the purpose, so it invites the kind of unactionable responses we've been seeing: "I wanted the words to the song" or "Where do they [snails] come from?" It assumes no one who reads the page is an editor. These are not good ways to lead people to click "edit" except in frustration, because this is not Facebook. Shitcan it. Yngvadottir (talk) 21:58, 4 February 2013 (UTC)
 * 2) Support. I thought that an article talk page was intended for feedback. Why do we need another mechanism? All the feedback comments I have seen seem to be unconstructive or incoherent. How is it that people seem to leave better messages on talk pages? &mdash; RHaworth (talk · contribs) 23:06, 4 February 2013 (UTC)
 * 3) Strong Support. Discussion pages are for feedback. Readers should be guided there instead.--Aschmidt (talk) 00:57, 6 February 2013 (UTC)
 * 4) Support. Because as it turns out, you need at least a low-level understanding of how Wikipedia operates in order to make meaningful suggestions. And the talk page--a small bar to jump--helps screen for that. The unwritten note on the front door "says": you are welcome to participate here, but you must first actually become one of us--an editor--by making an addition to this page. That turns out to be a useful hurdle. Whereas, from what I've seen, the tool's very ease-of-use makes it much less useful. Maybe even useless. Barte (talk) 07:17, 6 February 2013 (UTC)
 * 5) Support. --Andreas  JN 466 09:24, 6 February 2013 (UTC)
 * 6) Support The idea is good, but I prefer people making comments in the talk page, where they can be viewed by anyone, discussed and developed into content edits. --Enric Naval (talk) 11:56, 6 February 2013 (UTC)
 * 7) Stong Support. -- I would have proposed "Get rid of the damn thing." The encyclopedia article content area is sacrosanct and there are much better article improvement tools that the community has religated to the article talk page. The community does not want this useless tool and the sooner we eliminate this feature the better. -- Uzma Gamal (talk) 13:13, 6 February 2013 (UTC)
 * 8) Support. Actual useful feedback is simply so exceedingly rare, we'd do better without feedback and the additional work it creates. Someguy1221 (talk) 01:38, 7 February 2013 (UTC)
 * 9) Strong support Signal to noise ratio? All I see is noise. <span style="font-family:Verdana,Arial,Helvetica;"><b style="color:#333333;">Res</b> Mar  04:07, 7 February 2013 (UTC)
 * 10) Support Too much prime real estate and distraction for too little benefit. North8000 (talk) 19:03, 7 February 2013 (UTC)
 * Support. In theory i support the concept behind the AFT tool, and to a certain extend the tool itself. Conceptually taken more eyes on the various articles is a good thing and equally encourages readers to become involved. In practice, the tool simply doesn't provided substantial added value if one compares it to the time invested. As mentioned multiple times a large share of the comments is vandalism or requests that can never be actioned ("Her personal phne number plz!"). Even decent requests often require looking into the subject matter - or are alterations such as typo's that would have been handled a lot faster if the reader just edited the page. Seeing that we can barely handle vandalism patrol capacity despite have cluebot and some very nice tools to assist with what boils down to an "its vandalism or not" decision i cannot fathom how we can deal with a ton of requests that could literally ask or request anything. I did some feedback patrol in the past but ultimately i end up at the same station: My limited time is better spend on other backlogs as clearing these have more added value for Wikipedia and end up being more fulfilling than sifting trough a garbage belt of user comments. Excirial ( Contact me, Contribs ) 00:04, 8 February 2013 (UTC)
 * 1) Strong support Absolutely worthless. I have received maybe one useful comment out of a hundred on the articles I watch. Users would be better directed towards the article's talk page. That is what it is there for! Make the talk pages more accessible, understandable, this crap is not the solution. Pox on whoever came up with it. TimL &bull; talk 09:23, 9 February 2013 (UTC)
 * 2) Strong support. A grand waste of time. Most feedback commenters seem not to have the first idea what Wikipedia is about, how it works or what it's for. These aspects would be needed before feedback even begins to be useful. Span (talk) 16:03, 9 February 2013 (UTC)
 * 3) Support. First choice. I had the same experience as many other users here - too much heat and too little light to be useful. 1ForTheMoney (talk) 16:44, 9 February 2013 (UTC)
 * 4) Support. At least from my experience, the vast majority of feedback provided falls clearly in the "useless" camp. Moreover, the AFT is useless from a reader's perspective too, since user-provided comments are funneled into an area with very low visibility, and with no ability to check for any feedback. The argument that we need to provide a separate, fenced-off region for users to provide comments is baseless; if anything, it is an argument for switching the resources being used to develop AFTv5 into the development of the Visual Editor (which was supposed to be ready Real Soon Now™ back in 2005...) Tito xd (?!? - cool stuff) 07:26, 10 February 2013 (UTC)
 * 5) Support. I've never spotted a useful piece of feedback from the tool, most of the time it's just pointless spam. Any useful feedback usually comes from responses on the article's talk page. -- The1337gamer (talk) 14:53, 10 February 2013 (UTC)
 * 6) Support per Lemurbaby and others. <font color="#006600">It Is Me Here  <font color="#CC6600">t / <font color="#CC6600">c 20:43, 10 February 2013 (UTC)
 * 7) Support I sometimes look at comment pages for interest and I have hardly seen a single useful or worthwhile comment. Most of it is just pointless drivel. My other beef is that the entry point to this feature does not make it clear that you can leave feedback at all. It simply asks "Did you find what you are looking for?" and initially appears to be a Yes/No choice. I pointed this out almost on day one, and I cannot understand why this obvious deficiency has not been fixed. 86.130.67.118 (talk) 04:08, 11 February 2013 (UTC)
 * 8) Support -- Taku (talk) 05:42, 11 February 2013 (UTC)
 * 9) Support. Personally I think this is redundant as the purpose of the talk page is to provide feedback. The talk page is better anyway because at least there you can actually have a back-and-forth conversation and explain to a reader why their suggestion may not be actionable. I was reading over feedback left for Stacy London and two readers wanted to know her contact information. I'm not kidding. I don't know why anyone would think that Wikipedia would have/post that kind of information. Another reader wanted more pictures which isn't something that can be helped if no free images are available. Another reader wanted to know if she was in a sorority. This isn't useful feedback because this question is already answered in the article. Lastly, almost all of the other readers wanted to know about her love life. Odd to me considering she's not a Rihanna or even an attention-hungry Heidi Montag. She's a stylist and an editor whose level of fame is similar to that of Anderson Cooper, way below the tabloid level. On a side note, ditto to what said: "Why should [readers] bother and try to improve an article the way they see it fit, when there's a handy tool that let's the 'staff' know what should be done?" Agree 100%. It seems to me this tool solicits complaining—only a small percentage of which is actually helpful—instead of editing. //Gbern3 (talk) 18:51, 11 February 2013 (UTC)
 * 10) Support I have yet to see a single useful suggestion. Humongous waste of time and effort... --Randykitty (talk) 09:57, 16 February 2013 (UTC)
 * 11) Support Sadly, the feedback system isn't working out as hoped. Probably best to abandon it as a failed experiment. DocTree (ʞlɐʇ·cont) Join WER 22:08, 16 February 2013 (UTC)
 * 12) Support I have never found an actionable or helpful comment in the feedback from the articles I watch. This feature is supposed to recruit new editors, is it? I wonder if it doesn't do the opposite: if potentially useful editors who might otherwise use the talk page are instead posting their comments in this worthless venue. --MelanieN (talk) 22:53, 17 February 2013 (UTC)
 * 13) Support --Ita140188 (talk) 23:17, 19 February 2013 (UTC)
 * 14) Support - my experience is that it is a complete waste of time. -- Bduke   (Discussion)  22:09, 23 February 2013 (UTC)
 * 15) Support - Feedback is for the talk page. Doug (at Wiki) 01:47, 24 February 2013 (UTC)
 * 16) Strong support I've wasted a fair amount of time looking at those comments and the vast, vast majority either state the obvious, demand something that can't/won't/shouldn't be done, is from someone looking for gossip or trivia or provide no real constructive value ("more info" or "needs more" isn't really helpful). It was a noble effort, but a failed one. Niteshift36 (talk) 18:59, 25 February 2013 (UTC)
 * Comments
 * @ MZM - I think you know I'm not always known for my tact. :D  GregJackP   Boomer!   03:46, 21 January 2013 (UTC)
 * While very few feedback comments provide specific, directly actionable feedback, almost all of them reveal something about the article they're attached to that is helpful for us as editors to know. It is also extremely unlikely that the WMF is going to abandon this after over a year of development. --Aurochs (Talk | Block) 03:55, 21 January 2013 (UTC)
 * I don't disagree that it won't go away, but if they don't want to know our opinion, they (WMF or whomever) shouldn't ask. This is just my 2 cents. Regards,  GregJackP   Boomer!   04:05, 21 January 2013 (UTC)


 * I strongly disagree with this opinion. The feedback on pages on my watchlist has been very helpful and points out what might not be obvious to me (ie. this article is too technical, there aren't any pictures etc). It's not particularly difficult to scroll through the non-helpful comments and I've seen relatively little abuse. (Perhaps the ability to turn off feedback on controversial articles at the same time as semi-protection would be useful.) Danger High voltage! 03:02, 22 January 2013 (UTC)
 * I'm glad that your experience with it has been beneficial. Mine hasn't, it has been a waste of time.  Examples (when they actually write something, many have no comment whatsoever) include: "give me my imformation im looking for", "wikipedia does not have the information i need!", "lyers", "this was a bad site to use", "How do I e-mail your CEO or his assistant" (assuming they meant the CEO of the company the article was about), and "parties who are the primary parties" (in a SCOTUS case, where the article title is United States v. Kagama - i.e., both parties).  By the same token, useful comments are few and far between.  I started out enthused by the tool, and quickly became disillusioned when I realized it hurt more than it helps.  Also, looking at the list of views and the support that each view has received thus far, getting rid of the tool altogether has twice the number of supports as the other views, and a majority of the editors participating in the RfC have indicated a preference to either get rid of or severely limit the application of the tool.  That's not indicative of either support nor usefulness.  Regards,  GregJackP   Boomer!   04:23, 22 January 2013 (UTC)

If article feedback were as eloquent and thoughtful as the posts in this section, we wouldn't have any issue, heh. --MZMcBride (talk) 04:48, 22 January 2013 (UTC)


 * Comment I've re-named articles, re-written sections and added a new article based on feedback, mostly from IPs. I think this is one of those issues like IP editing generally: they may be responsible for 98% of the vandalism or pointless drivel that goes in there, but the vast majority of their comments are designed to help, even if they are sometimes mis-guided in those efforts. "If you don't like it, then ignore it" is a pretty good way to deal with it, but there are some real gems of good ideas in there and it's ridiculous to throw the baby out with the bathwater, unless you want to alienate the readers, rather than the editors. - SchroCat (talk) 05:38, 22 January 2013 (UTC)
 * Feedback ranges from an irrelevance to an annoyance in my editing. I can just ignore it so I'm not over-bothered if it remains, but it doesn't seem to have much use in my personal experience. -LukeSurlt c 14:24, 22 January 2013 (UTC)
 * It is not the tool that is useless, it is the way that it is used. The comments made may be useless to you or me, but to someone else they be of value.  Those they don't want to read the comments don't have to look. Graeme Bartlett (talk) 19:59, 22 January 2013 (UTC)
 * I addressed this fallacy here. --MZMcBride (talk) 20:04, 22 January 2013 (UTC)
 * But you still don't have to read the comments. It's your choice to do so. - SchroCat (talk) 22:19, 22 January 2013 (UTC)

"" It is not the tool that is useless, it is the way that it is used. ""

- Graeme Bartlett

It is used by readers who don't understand the purpose of the tool. Thus it is useless. Axl ¤  [Talk]  11:00, 23 January 2013 (UTC)
 * Actually no: it's a feedback tool and they are leaving us feedback, so most of them are using it entirely appropriately. The issue in a number of cases is that people don't understand what Wikipedia is and is not: that's an entirely separate matter. - SchroCat (talk) 11:11, 23 January 2013 (UTC)
 * Then, to paraphrase Sandstein, they are idiots that are leaving us "feedback". Graeme is correct, it's not the tool itself that is the problem.  It's dealing with inane comments, vandalism, etc.  Last night I looked at it again, and sent a "for a good time, call xxx-xxx-xxxx" to oversight - after it had set in the feedback for over 2 months.  Only those editors that want to use it do so, but we don't have the human resources to deal with the amount of garbage generated.   GregJackP   Boomer!   12:30, 23 January 2013 (UTC)
 * I'm not sure it's terribly helpful labelling readers who do not have accounts as "idiots"—we don't 'do' Wiki just for the sake of other editors, but wider readers too—and there are many more of them than there are editors. Yes, there is vandalism and a fair amount of inane nonsense, but I've been pleasantly surprised by the amount of interaction. Some of this is positive but misguided (requests for images etc) but a decent proportion has been appropriate, balanced and fair. Please see the feedback for the James Bond music article, which is about par for the feedback I've been keeping a watch on. - SchroCat (talk) 13:58, 23 January 2013 (UTC)
 * I didn't label the readers as idiots - I labelled the people that leave inane, stupid, useless feedback as idiots, and I stand by that characterization. A vast majority of out users don't leave feedback.   GregJackP   Boomer!   14:47, 23 January 2013 (UTC)

"" Actually no: it's a feedback tool and they are leaving us feedback, so most of them are using it entirely appropriately. ""

- SchroCat

You seem to be implying that the purpose of the feedback tool is to allow readers to give feedback, irrespective of the nature of that feedback. This page states that article feedback is "A new tool to engage readers to improve articles on Wikipedia." Unfortunately the tool fails to do this. Axl ¤  [Talk]  19:55, 23 January 2013 (UTC)


 * Most of the feedback from this tool is indeed nonsense/vandalism/unhelpful etc., but a small portion is pertinent (an example: a reader wrote on the Charlotte Brontë feedback that they wished to read about letters which Brontë had written to her teacher Constantin Héger; as a consequence, I added a section on the letters, something I hadn't thought of doing before). Therefore I could only support a full roll-out of this tool (as proposed in one section above) if it was modified to make it possible to easily separate the grain from the chaff. The flagging etc. is/was not easy to navigate (I gave up and couldn't see the point of it). If it's not possible to easily ditch the unhelpful feedback, the tool should indeed be abandoned, as not only does it waste editors' time sorting through it all, it gives the illusion to readers that Wikipedia is listening to them when it isn't (because editors have given up), which is a disservice to them. PaleCloudedWhite (talk) 01:50, 24 January 2013 (UTC)


 * How do you know that without the feedback tool the person who merely suggested the comment would not have instead actually joined the talk page discussion and/or added material directly and became a member of the community? The presumption was that giving feedback is a "gateway drug" that would encourage participation, but it could be tossing off one time feedback fulfills an illusion of actual participation and siphons off productive editors from actually contributing directly. -- TRPoD aka The Red Pen of Doom  20:53, 24 January 2013 (UTC)
 * That could be correct, but it might not be; it is all hypothetical. One cannot know either way. My comment is that the feedback tool is not effective as currently employed, so either should be modified or not employed at all. PaleCloudedWhite (talk) 06:10, 25 January 2013 (UTC)

I would like to comment that most of the time I'm looking at article feedback is when I consciously want to do something useful but am too lazy to do something "really" useful like writing an article. Thus, the feedback doesn't waste time I'd otherwise spend contributing, but rather increases the amount of time I contribute by giving me something "easy" to do when I wouldn't be tackling more difficult tasks anyway. Abyssal (talk) 19:38, 24 January 2013 (UTC)

I disagree with this viewpoint. Regardless of the tool's problems:    if this tool did not exist, it would have to be invented. --Noleander (talk) 15:42, 25 January 2013 (UTC)
 * Why's that? --MZMcBride (talk) 16:12, 25 January 2013 (UTC)
 * Without a feedback tool (not to say this exact one) we would miss-out on lots of great input from readers. Readers that cant/wont edit an article will type text into a box.  All serious endeavors need a mechanism for their customers (viz readers) to provide feedback.   If the Feedback tool disappeared tomorrow, attempts would immediately begin to revive it (or a variant of it).  Thus, it is better to fix it than eliminate it.  Personally, I'd be happy if the text simply went to the article's Talk page, so watching editors can act on it.  --Noleander (talk) 16:23, 25 January 2013 (UTC)
 * To the extent that someone cant/wont edit an article to provide feedback, I feel that that's an indictment of the talk page mechanism in general, and suggests replacing that entirely, rather than bolting on an alternative which muddles the two and which undercuts our "anyone is an editor" message" would be a much better direction to head. --j⚛e deckertalk 16:37, 14 February 2013 (UTC)

The tool is not useless. An article for which I am the main author has had feedback from an academic who wrote a leading reference source on the topic, suggesting an upgrade from that reference. I don't think these debates should be conducted in such unmeasured terms. Basically we need an anti-spam post-moderation feature added into the system. Charles Matthews (talk) 08:06, 6 February 2013 (UTC)
 * while not entirely useless, I think it would be hard to have purposefully designed something that would have generated so little feedback of with actual value within the same size mountains of chaff. -- TRPoD aka The Red Pen of Doom  19:08, 25 February 2013 (UTC)

Comment: Since it appears that people are still casting votes on this RfC, I would like to remind everyone that we have made some significant improvements to the Article Feedback tool, which is now ready for testing on our prototype site, as described on this RfC's talk page. These new features were developed to address many of the issues raised in this discussion, and include: We invite you to visit this testing page and try out these new features, then let us know what you think. It appears that only a few of the people commenting on this RfC have tested these features, which seems unfortunate, given that the new Article Feedback software is now significantly different from the old software which most votes are now based on. Together, these new features address many of the concerns we have heard on this RfC, and would seem to deserve to be considered as part of this vote, as I stated in my personal view below. We hope this brief product update will help you all (and the administrator closing this RfC) reach a more informed and nuanced decision about the future of this tool on the English Wikipedia. Sincerely, Fabrice Florin (WMF) (talk) 22:47, 25 February 2013 (UTC)
 * Better feedback filters: surface good feedback, hide useless comments.
 * Simpler moderation tools for editors: moderating feedback is now a lot easier and faster for editors.
 * Separate reader moderation tools: encourages readers to moderate with their own tools.
 * Feedback link on articles: shows up if there is useful feedback for your article.
 * Auto-archive comments: remove comments that are not moderated after a while.
 * Discuss on talk page: share useful feedback with editors on talk page (under development).
 * What is probably more unfortunate is that the date set for the encyclopedia-wide roll out of the tool was less than 6 weeks from any availability for editors to test the new version in the test environment and much less than 30 days to test it in limited circumstances in "the real world" to see if the new features actually do address the concerns and do not create new ones. -- TRPoD aka The Red Pen of Doom  22:57, 25 February 2013 (UTC)

View by Okeyes (WMF)
This RfC raises a lot of interesting questions: I'll do my best to answer them, bit by bit, but I'd like to start with a "how do I see AFT5?" schpiel. Obviously, I'm supportive of it ;p.

We started working on this tool over a year ago, with a couple of objectives: to provide feedback that editors could use to improve articles, and to offer readers a way of contributing to the encyclopaedia and maybe becoming editors. I'd argue that we've succeeded in these goals. The latest research can be found on meta; it shows that on feedback quality, based on evaluation by around 20 long-time editors, between 30 and 60 percent of feedback submitted can be used to improve an article. Of those users who try to sign up and contribute, 66 percent of their edits are helpful (or, at least, not reverted - it's hard to gauge edit helpfulness). We started trying to make a tool that would provide useful feedback and let readers contribute; the tool provides, however you look at it, a lot of useful feedback, and acts as an avenue for readers or first-time editors to make quality contributions.

Along the way we've integrated the spam blacklist, worked in abuse filters so that people can cut out blatantly improper feedback without ever having to bother a feedback patroller with it (in the same way we do edits), and built software so that, on high-volume articles, the feedback tool can be turned off if there are serious problems with contributions people are making. To come, we've got an improved interface that makes getting rid of bad feedback (and highlighting good feedback) easier. We're doing a lot of work to up quality above where it is even now, and make it easier to deal with things that slip through the cracks. I appreciate things aren't perfect right now - as with editing, they never will be.

So, on to the questions.


 * 1) What should the scope of the article feedback tool be (cover all articles, opt-in per article, etc.)?
 * As said, we're aiming at 100 percent deployment (with some caveats: as said above, if you've got a high-volume protected page where people are misbehaving, you can turn the tool off). I'm pretty comfortable with this decision, because the alternative (opt-in, rather than opt-out) creates a lot of problems. As AgnosticAphid says above, you're going to encounter quite a few unnecessary disputes - and even if sticking the feedback tool on an article is clearcut, with over 4 million articles (and that's not to mention help pages, which we've enabled the tool for so that editors can tweak help documentation to meet what new editors need need) it's going to take a heck of a long time to manually go through them.
 * I appreciate that there are going to be submitted pieces of feedback that are unhelpful or even inappropriate; I think this is inevitable in any open system, just like it's inevitable that we get vandalism. The important thing is having ways to deal with it, be it the tools we've built in, third-party scripts (in the same way we have Huggle or Twinkle) or human intervention. And actually, I think the quality of feedback is going to be higher by volume on the pages we haven't applied the tool to, and the volume itself lower: the way we randomly selected the 10 percent that currently have the tool means that high-traffic pages are overrepresented.
 * 1) Are there sufficient resources to moderate and respond to all of the feedback?
 * The honest answer is "probably not", although that's at least in part down to how many people want to use the tool :). But, again, I don't see this as a problem: we're a wiki. Always have been, always will be. Edits will need oversighting or deleting, bad edits will slip through the cracks, and we accept that because it's necessary to produce the good things that an open system gives us. I see no reason not to take the same attitude with feedback.
 * 1) How will abuse filters handle articles in which the subject's title contains a disallowed word (e.g., Blue-footed Booby + "boob")?
 * This one I don't know the answer to - I'm not regex-literate (although I'm pretty good with R these days). But my first question would be "how do we handle edits?" Presumably new people try to edit the Blue-footed Booby page; what's our solution there? Either we have a solution, in which we can probably apply it to feedback (feedback falls under abuse filters in the same manner that edits do), or we don't have a solution but the community considers that acceptable, in which case it's not an article feedback-specific problem or a blocker. If it was a blocker, we wouldn't allow non-autoconfirmed editing.
 * 1) Will the tool continue only be a(n expanded) box or will it go more minimal? Will it go "above the fold" (e.g., File:Article-link-to-feedback.png)?
 * On the first point - at the moment, the plan is to continue using an expanded box. I'd caution against using a more minimal design: when we tested one we found it actually reduced the quality of submitted feedback. I don't think that's something anyone participating in this RfC wants. In regards to the "above the fold" link; that link is something we're working on, but there are some caveats :). The actual intention of the link isn't to solicit feedback comments - it's simply a link to the feedback page for this article. It would only be visible to editors, and is an attempt to target a possible problem mentioned above of not enough people monitoring feedback. Again, I think supporting better monitoring is something we can all agree is A Good Thing.

With all of these things taken into account, I'm in favour of deploying the tool, although as mentioned that's hardly a surprising statement ;p. I hope others are as well.

I hope I've answered the primary questions. To avoid piling-on or badgering, I'm going to be lurking largely on the talkpage from now on. If you have individual questions about our data/upcoming plans/whatever or seek some clarification, just drop a note there and I'll respond as soon as I can. Thanks :). Okeyes (WMF) (talk) 03:14, 21 January 2013 (UTC)

The AbuseFilter answer seems like it needs further thought. The issue is that you've currently implemented fairly heavy-handed abuse filters, which are fine for a limited test, but how will people comment on the article cunt or faggot or bitch? Standard editing is not subject to abuse filters like these. Article feedback, however, is subject to very stringent filters. And article feedback will invariably include currently disallowed words. The Blue-footed Booby example was kind of convoluted... the current feedback filters would hit thousands of legitimate article titles and consequently a lot of legitimate feedback. Is there a solution to this? --MZMcBride (talk) 03:52, 21 January 2013 (UTC)
 * Users who endorse this view
 * 1) -- Guerillero  &#124;  My Talk  00:35, 22 January 2013 (UTC)
 * 2) WhatamIdoing (talk) 19:16, 22 January 2013 (UTC)
 * 3) <font color="DarkGreen">AgnosticAphid  talk 01:08, 23 January 2013 (UTC)
 * 4) Regards, Sun Creator(talk) 01:38, 23 January 2013 (UTC)
 * 5) Yep. --11:09, 23 January 2013 (UTC), Utar (talk)
 * 6) Good analysis. &bull; Jesse V.(talk) 16:30, 24 January 2013 (UTC)
 * 7) Harsh  (talk)  18:17, 25 January 2013 (UTC)
 * 8) Ulflund (talk) 10:30, 26 January 2013 (UTC)
 * 9) Support. I don't really understand abuse filter objections - false positives will lead to rule refinement over time. I also agree that the box should not be made smaller, and this kind of bikeshedding can't stand up to the real A/B testing that went into the design. Dcoetzee 00:58, 28 January 2013 (UTC)
 * How do we distinguish between actual profanity and words that appear to be profanity (to a pre-defined pattern that isn't examining context)? --MZMcBride (talk) 05:55, 29 January 2013 (UTC)
 * By not doing that (that is, by improving the filter to support and use appropriate context - this is not as difficult as it sounds). Dcoetzee 06:29, 29 January 2013 (UTC)
 * Well, actions speak louder than words. Go on, prove me wrong. :-) As it is, the current abuse filters are un-deployable, as I see it. That is, they're simply too broad and too zealous for a full-scale deployment. --MZMcBride (talk) 07:08, 29 January 2013 (UTC)
 * More to the point, show me a single example of any automated anti-abuse filter anywhere on the internet every being effective at stopping human beings who are trying to get past the filter. Filter after filter has been tried, on BBSs, Usenet, Email, blog comment sections, IRC, and even here on Wikipedia. ALL fail without human help. Without the humans fighting the abuse, the humans creating the abuse adapt and learn how to get the abuse past whatever filter you can create. Humans are simply more clever than computer programs. Once you have the humans on patrol, the anti-abuse filter can knock down 95% to 99% of the abuse, lightening the workload considerably, but until someone creates a working A.I., abuse filters by themselves are not the answer. If you disagree, show me a single counterexample. WMF has bot been able to attract enough humans to handle the limited test we have now. Why should we believe that expanding the test will do anything other than making this worse? --Guy Macon (talk) 15:28, 12 February 2013 (UTC)
 * 1) Support (comment below) <font size="4" face="Brush Script MT"><font color="Black">S  Pat   <font color="black" size="1">talk 05:04, 29 January 2013 (UTC)
 * 2) support FeydHuxtable (talk) 18:32, 30 January 2013 (UTC)
 * Comments
 * No, although I think perfection is probably beyond any software we write :/. Thousands of articles, possibly - but when we're talking a wiki with 4 million articles that's, what, a 99 percent success rate :). I'm slightly confused by statements that the current abuse filters do not include abuse filters like these: we've got one that hits personal attacks, one that renders Everyone Poops difficult for newcomers to edit, and one that hits two of the three examples you brought up (forgive me if I've read these wrong; as said, regex is not exactly something I'd stick in my C.V. skills section). Okeyes (WMF) (talk) 03:56, 21 January 2013 (UTC)
 * Additionally, cunt, bitch and faggot are all semi-protected. Apparently, the amount of vandalism they got was greater than the edit filters could handle; perhaps articles like those will also wind up blacklisted from AFT5. --Aurochs (Talk | Block) 04:09, 21 January 2013 (UTC)
 * I deliberately (and explicitly) had a method of blocking AFT5 built, interlocked with protect options - so it'd be trivial for any admin to turn the tool off in those circumstances, yep. Okeyes (WMF) (talk) 04:14, 21 January 2013 (UTC)
 * Sorry, I guess this point isn't as obvious as I thought. I'll be more explicit. I registered a new account here and vandalized the page Constitution of Tonga, adding <tt>The word "cunt" has disputed origins.</tt> to the wikitext of the page and using <tt>The word "cunt" has disputed origins.</tt> as my edit summary (this edit). This was allowed. (Though apparently multiple obscenities for new accounts are disallowed by Special:AbuseFilter/380.)
 * Trying to take that exact same string (<tt>The word "cunt" has disputed origins.</tt>) and inputting it as article feedback was blocked by Special:AbuseFilter/460 (twice, once with and once without the quotation marks).
 * Again: the current abuse filters that are applied to article feedback are much stricter than the abuse filters applied to standard edits. Yes, the example above is contrived. But the point is that I'm wondering how you propose to deal with this issue that will likely affect thousands of articles, from Blue-footed Booby to Cunt (album) to Faggot (food). This is an issue if article feedback is deployed to all articles, so I'd to see some further thought put toward this question (question three of the RFC). --MZMcBride (talk) 03:06, 22 January 2013 (UTC)
 * In regards to point #2, how do you square this with the declining number of active admins? Backlogs in high priority admin tasks are becoming more frequent, and it seems unwise to add extra things which require admin attention unless there's a worthwhile payoff from doing so. What, if any, metric is being used to monitor the cost/benefit relationship here? Nick-D (talk) 10:50, 21 January 2013 (UTC)
 * I square it by noting that we deliberately made the pool of people who could monitor feedback far wider than admins :). Any rollbacker, admin or reviewer can hide feedback. It's very difficult to get a cost/benefit relationship out of this in terms of "how many things had to be hidden for one article to be changed" because, well, how do you measure where article changes are sourced from? As the links above indicate, our research suggests that the actual number of 'improper' pieces of feedback is very low in percentage terms (I'd note that readers and, indeed, any editor who doesn't have the aforementioned rights, can play a part in monitoring and prioritising feedback. So it's very much on the shoulders of all rather than the few.) Okeyes (WMF) (talk) 16:00, 21 January 2013 (UTC)
 * Thank you for your response. I have to say that I'm concerned that the WMF has apparently rolled out this tool without a strategy in place to monitor its impact on how editors' spend their time, and whether it is having adverse effects on this. Your last sentence is particularly worrisome; the WMF shouldn't be dropping any extra workload on the most active editors without a good reason for doing so, as well as a strategy to make sure that it's working out. Nick-D (talk) 07:11, 22 January 2013 (UTC)


 * This whole section (and the "#2" response in particular) illustrates an attitude from the WMF towards normal editors that's cavalier at best and downright contemptuous at worst. (And some people wonder why editor retention is dropping!)  Yes, it's a wiki.  Yes, it's impossible to achieve absolutely zero vandalism, libel, and spam.  But to introduce a whole new vector to introduce vandalism, libel, and spam and then expect the editor community to tirelessly clean it all up, forever and for free, without complaint?  That's madness. Andrew Lenahan -  St ar bli nd  03:40, 22 January 2013 (UTC)
 * So, a couple of things. First, this isn't the WMF's view, this is my view. Hi! I'm Oliver. I've been here since 2006, I'm an admin, and I'm precisely the kind of person who I (apparently) am cavalier and/or contemptuous to (which feels sort of like oratorical self-harm). If it did come off as cavalier or contemptuous, that certainly wasn't my intention - and as a volunteer as well as a staffer it's not like I don't get the problems here. If you think introducing a new vector is inherently a problem, we're going to have to disagree - anything, however good or bad, can be recast as the opposite with the line "vector for [opposing edge case]" - and I'd argue the problem here is in implementation and interpretation. You think that the positive things we get out of AFT5 (implementable feedback, new contributors, a form of micro-contribution we can direct newcomers to) are outweighed by the bad things (vandalism, libel, so on - although I'd note that when we talk about it being a 'vector for libel' the amount of stuff that actually needs to be actively removed on those grounds is tiny). That's fine; we can disagree on that, and I'm happy to have a proper back-and-forth on it. But I am slightly hurt by the tone that people are taking here - and would argue that if we had more civil back-and-forths instead of questioning each others motives or attitudes right off the bat, we'd probably be doing a bit for editor retention on that front, too :P. Okeyes (WMF) (talk) 06:41, 22 January 2013 (UTC)
 * Hi Oliver. I don't doubt that the feedback tool was conceived with good intentions, but as evidenced by the real-world experiences here, it simply hurts more than it helps and, if implemented on every single article would be an enormous source of, yes, vandalism, libel, and spam, which you and the WMF apparently assume the editing community has infinite and tireless patience in cleaning up.  What if you got home one day and found that something had pooped on your floor?  You'd do what any sane person would do, clean it up, spray some Lysol, and get on with your life, right?  Ok, now imagine that instead of one poop, the sewer company has hooked up a main sewage pipe straight into your front window, spewing a constant torrent of filth.  You wouldn't say "Hey, hold on, there might be some good stuff in here!"  Sooner or later you'd find the situation unbearable and cleanup impossible, and you'd have to move.  That's exactly what article feedback is like.  Even if the editors put up with it for awhile (and you can see from this RFC that that patience is wearing thin already), all the extra time spent moderating through useless junk is time the editing community could and should have spent actually editing and improving articles. Andrew Lenahan -  St ar bli nd  18:19, 22 January 2013 (UTC)
 * That's a flawed analogy in many ways, but, lets go for it; I'd go "hey, maybe you guys should build some filters in to keep the turds in the tube! And maybe I should get subsidised lysol". This is not really accurate insofar as it ignores those positive things that come out of the tool. Can I ask; there are obviously some cases when the tool has been helpful. What's your feeling about maybe an opt-in system, or something less-than-full-deployment-but-more-than-turning-it-off? Okeyes (WMF) (talk) 18:26, 22 January 2013 (UTC)
 * Not to state the obvious, but opt-in is my view and completely turning off the tool is GregJackP's view. Andrew has commented on both views on this page, so you may already have the answer to the question you're asking. ;-) --MZMcBride (talk) 18:56, 22 January 2013 (UTC)
 * It's not my real-world experience that the feedback is harmful. It has been my clear experience that the feedback is either helpful or takes two seconds to tag as resolved with an stock explanation.  WhatamIdoing (talk) 19:19, 22 January 2013 (UTC)
 * Yes, this. Tools that help reduce backlogs and that improve the editing experience receive very little criticism, I've found. Tools that help increase backlogs and that degrade the editing experience should be receiving more criticism. We have finite resources. We need to focus on working smarter, I think. --MZMcBride (talk) 03:55, 22 January 2013 (UTC)
 * I think the tool is fantastic. I have personally found it to be incredibly useful in giving a picture of how the content that we put up appears from a readers' point of view. Sure, a large fraction of the comments are usually non-relevant (I've found surprisingly low amount of actual vandal-stuff), the few constructive comments make the whole thing worth it. I think that reducing the barrier from going to the talk page to comment has actually increased number of *constructive* comments, not just comments in general. Contrary to what others have said here, I feel that this tool actually reduces the gap between readers and editors. Sorry if I'm repeating some points - I haven't had a chance to go through all the arguments from all sides :) <font size="4" face="Brush Script MT"><font color="Black">S Pat   <font color="black" size="1">talk 05:04, 29 January 2013 (UTC)
 * I agree strongly with this. Although moderation does require effort, the absolute number of productive comments we receive is much greater with AFT5 deployed, and the people providing them are more diverse. Dcoetzee 21:20, 30 January 2013 (UTC)

View by Srleffler
While I see the potential for a feedback tool to be useful, I believe that the current implementation fails to achieve this, and may do more harm than good. The feedback tool channels readers who want to comment on an article to leave their comment in a place where it is less likely to be read than the article talk page, and where it is impossible for anyone to respond, either to provide information to the reader or to get more information on the changes they would like to see made to the article. Readers get frustrated because they are being asked to give feedback, but they see quickly that their feedback gets no response.

This problem severely limits the utility of the tool. I do not feel that the tool should be widely deployed until these implementation problems are fixed, and have been beta tested on a small number of articles. This tool is simply not ready for wide deployment, and will do more harm than good to the project.--Srleffler (talk) 04:26, 21 January 2013 (UTC)


 * Users who endorse this view
 * 1) Nick-D (talk) 10:52, 21 January 2013 (UTC)
 * 2)  GregJackP   Boomer!   20:04, 21 January 2013 (UTC), as a second choice to getting rid of it altogether, per my suggestion above.
 * 3) Time and resources should be put into examining and improving the feedback tool we already have: the talkpage. Comments on the talkpage are immediately visible to all, including the person who makes the comment. The feedback tool is too obscure, and does not provide the user with sufficient feedback.  SilkTork  <sup style="color:#347C2C;">✔Tea time  23:36, 21 January 2013 (UTC)
 * 4) I think this is a close variant of my view. --MZMcBride (talk) 03:52, 22 January 2013 (UTC)
 * 5) Yes, the current structure of the tool needs fixing. In my mind, it should be just another way to get people to the talk page. Heck, maybe it could just be a link to make a new section on the talk page.  Abductive  (reasoning) 09:27, 22 January 2013 (UTC)
 * 6) Absolutely. If someone comments on a talk page on my watchlist, I'll see it, but if they use the AFT I won't know about it as I've given up reading the feedback due the low SNR. --Michig (talk) 17:20, 23 January 2013 (UTC)
 * 7) Yes, I think this is the closest to my own view. Although feedback has sometimes thrown up occasional good points for correction, in the main, the useful posts would have been formerly have been placed on the talk page where they could be discussed or clarified with anon IP addresses. A significant number of feedback entries seem to be questions, as if they're expecting an answer back, which as they are normally IP addresses mean that replying on an IP talk page is a waste of time. There appears to be no way to leave feedback in reply to other feedback, as it were. Ultimately, it probably makes us look like we're ignoring feedback, whereas if there was a specific problem before it could be discussed on the talk page before implementing. Bob talk 21:51, 27 January 2013 (UTC)
 * 8)  Oreo Priest  talk 20:41, 28 January 2013 (UTC)
 * 9) Exactly - Nabla (talk) 22:00, 28 January 2013 (UTC)
 * 10) Support. – Wdchk (talk) 15:32, 2 February 2013 (UTC)
 * 11) As Bob says, you can't give feedback to the feedback. You can't discuss the feedback, improve on it, point to former discussions, etc. I can't see what other people think of the suggestion, beyond "xx% found it helpful". --Enric Naval (talk) 12:15, 6 February 2013 (UTC)
 * 12) Support – I have yet to be convinced that AFT is an improvement over article talk pages. <font color="#006600">It Is Me Here  <font color="#CC6600">t / <font color="#CC6600">c 20:39, 10 February 2013 (UTC)
 * 13) Support Very well said. Most feedback is drivel, but even worse is the diversion of potentially useful editors away from the true feedback/comment page (the talk page), to a box that gives them the false impression that they are contributing to the encyclopedia. If a reader actually takes the time to post a thoughtful feedback comment, but they get no response and see no change in the article, how does that make them feel? --MelanieN (talk) 23:05, 17 February 2013 (UTC)


 * Comments
 * This argument is undermined by a couple of key facts. First, the strong response to AFT5 relative to talk page comments tells me that readers prefer it greatly over the talk page system. We are not channeling readers away from the talk pages, because chances are pretty damn good that those readers would not have posted on the talk page anyway. Second, the talk page comments that are left by unregistered users are rarely responded to, at least in my experience. I have seen talk page comments willfully ignored by editors who think they're reverting vandalism. --Aurochs (Talk | Block) 05:03, 21 January 2013 (UTC)
 * I agree with your first point, but feel that a different feedback mechanism would better address the clear need for an easy way for readers to leave comments and post questions. For example, the feedback form could automatically create a talk page entry and give the person leaving a comment a link to the correct talk page section. Alternatively, a better feedback system could be devised, which actually meets Wikipedia's needs. I'm not opposed to feedback in general—I'm merely arguing that the current implementation fails to achieve its goals, and may do more harm than good overall.
 * I think your second point is mistaken, or at least disagrees with my experience. On the articles I edit, useful comments are much more likely to get an appropriate response on the talk page than in the feedback system. The inability to reply is a key problem: I have seen many feedback messages where the reader clearly had an idea in mind for how the article could be improved, but didn't express it clearly enough that I could make out what they wanted. If I could ask them what they meant, it's possible I could have done something with the suggestion. In other cases, the reader is looking for information that I know exists elsewhere. I could point them there, if only there were a way to reply to their comment. I've seen comments from readers expressing frustration that their previous comments received no response, but in many cases there is no way to respond, nor even any way to tell the reader why they are not getting a response. All of these readers have been ill-served by being channeled into the feedback mechanism. Some of them would have perhaps found the talk page if the feedback mechanism didn't exist. The ones who did not find the talk page would still be mostly better off for not having left feedback that will not do any good.--Srleffler (talk) 06:08, 21 January 2013 (UTC)
 * To comment briefly; on the first point, the problem with talkpage entries is that it would swiftly overwhelm the talkpage system and the editing format is not suited well to removing vandalism where subsequent contributions have been made. On the second, a reply mechanism is on our to-do list. Okeyes (WMF) (talk) 06:33, 21 January 2013 (UTC)


 * "Readers get frustrated because they are being asked to give feedback, but they see quickly that their feedback gets no response."
 * Can you give an example of this? I haven't seen any readers try to follow up on their suggestions.  I have, however, had several productive conversations with people who have left feedback.  All of these conversations happened on the users' talk pages, which are easy enough to find.  WhatamIdoing (talk) 19:23, 22 January 2013 (UTC)

I understand what this viewpoint is saying. But the remedy is to tweak the mechanics of the tool (for instance, direct the given comments into the article's Talk page, so editors watching the article can respond) rather than discarding the tool altogether. --Noleander (talk) 15:46, 25 January 2013 (UTC)

View by Nick-D
I think that the current feedback tool does Wikipedia's readers a disservice, and likely discourages them from editing articles for the following reasons: Overall, I think that the feedback tool has been a worthwhile experiment, but it's not working and is probably causing harm to the goal of expanding and improving the encyclopedia. I'd suggest that it be removed and be replaced with a prominent suggestion that people either jump in and edit the article or leave a post on the talk page. Nick-D (talk) 10:28, 21 January 2013 (UTC)
 * Any comments left in the feedback tool are hidden away, few editors regularly check to read them and there's no practical way of discussing the comments with the people who leave them. As a result, when a reader does leave a useful comment it won't be noticed for days. When an editor asks a question or leaves a comment which would benefit from being discussed, there's no practical way to do this - I've seen some unspecific comments left by people credibly claiming to be experts in the field, but there's no real way to follow up on these messages with them given the delays in the posts being read and the location they're left in.
 * The tool encourages a division between 'readers' who only comment on articles (at most) and 'editors' who write articles and follow up on these comments. This isn't how any editors I know see themselves, and people only leaving comments in the belief that someone else will act on them are likely to be frustrated. Instead of annoying these people by giving them a feedback tool which isn't an effective way of getting the article changed, we should be encouraging them to jump in and edit the article, and hopefully then go on to edit other articles. Most won't do so, but some will and this will be for the general good.
 * As an example, this is the kind of use of the tool which concerns me in regards to the above two points - the person who left it appears to have specialist knowledge of the topic and some potentially useful but a bit vague ideas on how to improve this neglected article, but instead of encouraging them to jump in the tool has given them a halfway house in which they've asked for someone else to fix the article for them. Nick-D (talk) 10:44, 22 January 2013 (UTC)
 * However, a lot of readers are not aware that they can edit - and the editing interface is hardly friendly. It's also worth noting that as soon as they've submitted feedback they're invited to contribute via editing, and linked to a tutorial. So, AFT5 gets the above example, saves it, invites the person to jump through the hoops and, if they fail (which they would have anyway, either by not knowing how to contribute or not knowing they could), has saved the feedback so that it can be implemented by someone who does know how to edit. Okeyes (WMF) (talk) 11:23, 22 January 2013 (UTC)
 * Yet in this instance the feedback was left on a page which hardly anyone watches or edits (I only watch it as it was the scene of a nasty conflict about 5 years ago) and while the feedback appears to be informed by specialist knowledge and an interest in the topic, it isn't specific enough to be genuinely useful. The person who left this comment is probably wondering why no-one has acted on it, and I'm frustrated that he or she didn't try to edit the article! Nick-D (talk) 10:44, 23 January 2013 (UTC)
 * Judging from the comments I've seen, many of our readers are genuinely confused about the purpose and status of Wikipedia, and are wasting their time by writing comments which can't possibly be acted on or responded to. As an example, I have articles on the military forces of a number of developing countries on my watchlist, and they're attracting a steady stream of questions from people who appear to believe that these articles are the official website of the military and want information on how to enlist.


 * Users who endorse this view
 * 1) Especially agree with the third point.  Sandstein   12:10, 21 January 2013 (UTC)
 * 2) This seems like a reasonable position to me. We tried AfT hoping it would involve readers, but mostly it's inadvertently set up a segregated "reader area" where few editors want to go because of the noise. A fluffernutter is a sandwich! (talk) 16:32, 21 January 2013 (UTC)
 * 3) Concur.   GregJackP   Boomer!   20:06, 21 January 2013 (UTC)
 * 4) I've encountered the third phenomenon with Amtrak-related articles: people who want fare information, or phone numbers, or driving directions from a random location. I think I've seen three genuinely useful pieces of feedback over the last few months, two of which I was able to act on. One of these was left by an established user. Mackensen (talk) 21:54, 21 January 2013 (UTC)
 * 5) I feel that efforts to improve feedback are better directed to improving readers' access to the talkpage than by creating a parallel system.  SilkTork  <sup style="color:#347C2C;">✔Tea time  23:43, 21 January 2013 (UTC)
 * 6) AFT is a failed idea, and it is time to pull the plug on this one, and try something else. Courcelles 00:48, 22 January 2013 (UTC)
 * 7) The "Judging from the comments..." paragraph in particular hits the nail on the head in a much more diplomatic manner than I ever could. --MZMcBride (talk) 03:51, 22 January 2013 (UTC)
 * 8) I strongly agree with all three points, especially the third one. Sometimes, we receive comments (both positive and negative) that aren't helpful in any way or are just simply too vague (e.g. "needs reorganisation", "more details" etc). I mean, the ideal behaviour we expect to see from editors is that they will take the initiative to make improvements to the articles instead of directing others on what to do (when their "instructions" are not even clear). This article feedback system simply opens up more channels for backseat driving.   <font style="color:#000">LDS  <font style="color:#333">contact me 05:08, 22 January 2013 (UTC)
 * 9) I eyeballed Special:ArticleFeedbackv5, and point #3 is a reasonable description of what's going on. MER-C 05:13, 22 January 2013 (UTC)
 * 10) Support. I also very much agree with Fluffernutter's and SilkTork's comments, above.--Srleffler (talk) 06:32, 22 January 2013 (UTC)
 * 11) Strongly agree, especially with point 2. Abductive  (reasoning) 09:30, 22 January 2013 (UTC)
 * 12) I agree with points 1 and 3. Regarding 2, I wouldn't say AfT encourages this division (per Okeyes comment), but to an editor reading the feedback I can see how it highlights that a division does exist <font color=#000000>Jebus989 ✰ 11:45, 22 January 2013 (UTC)
 * 13) I agree wholeheartedly. A look at the "feedback" link on my watchlist provided very few useful commentaries. The idea of seeing what readers think about an article wasn't a ridiculous one, and I agree that it was worth a try. But the experiment has shown that it hasn't worked, and it should therefore be ended. The article talkpage is a superior forum for comments that encourage improvement. Sjakkalle (Check!)  19:01, 22 January 2013 (UTC)
 * 14) <font face="trebuchet MS">- filelake<font color="#0B0">shoe  <font color="#0B0"> &#xF0F6;    14:04, 23 January 2013 (UTC)
 * 15) As well as "likely discourages them from editing articles" there is also the problem that it relies on diverting editors from more useful work to processing the "feedback".  Ϣere Spiel  Chequers 16:40, 23 January 2013 (UTC)
 * 16) Yep. --  TRPoD aka The Red Pen of Doom  22:41, 23 January 2013 (UTC)
 * 17) I agree with 90% of this. There are fundamental problems with this tool that should be addressed, but I think it has the potential to work so I do think it should stay. &bull; Jesse V.(talk) 16:33, 24 January 2013 (UTC)
 * 18) Especially the second point. The solution is to nudge people to edit, not to nudge them towards blogging about it. - jc37 23:19, 25 January 2013 (UTC)
 * 19)  Oreo Priest  talk 20:40, 28 January 2013 (UTC)
 * 20) Agreed - Nabla (talk) 22:03, 28 January 2013 (UTC)
 * 21) Support (I've also supported GregJackP above), the talk-page allows one to enter into a dialogue with the reader, whereas ATF5 is a virtual waste-paper basket --catslash (talk) 22:36, 28 January 2013 (UTC)
 * 22) Support. I supported GregJackP above; the tone of this one is more palatable to me, and I really don't object to further study of how this could be tweaked to work better. But I'd like to see a more comprehensive theory of how to evaluate the success of the tool. One thing that sticks out in the report Oliver linked above is the lack of discussion of why volume and quality of feedback, as raw numbers, can help us gain insight into whether or not the tool is accomplishing its goals. I don't think they are; I think a more sophisticated theory of measurement is needed. -Pete (talk) 08:04, 29 January 2013 (UTC)
 * 23) support This should be seen as a subtler wording than greg's, but it means the same thing. It was a reasonable experiment, but there is no way that full deployment would be practical, so we should try other things entirely  DGG ( talk ) 20:48, 29 January 2013 (UTC)
 * 24) Support Nothing further to say here except "I agree" and point you to the comments to Greg's post. – Philosopher Let us reason together. 23:44, 29 January 2013 (UTC)
 * 25) Support Armbrust The Homunculus 10:34, 2 February 2013 (UTC)
 * 26) Support. Al Lemos (talk) 16:16, 2 February 2013 (UTC)
 * 27) Support. --Andreas  JN 466 09:26, 6 February 2013 (UTC)
 * 28) Enric Naval (talk) 12:17, 6 February 2013 (UTC)
 * 29) --j⚛e deckertalk 16:21, 14 February 2013 (UTC)
 * 30) Support --Ita140188 (talk) 22:47, 19 February 2013 (UTC)
 * 31) Support. — Shmuel (talk) 18:39, 22 February 2013 (UTC)

I concur with SilkTork's comment: "I feel that efforts to improve feedback are better directed to improving readers' access to the talkpage than by creating a parallel system."  Many readers will knot know how to edit, or not want to. On the other hand, many readers will be comfortable with typing a comment into a rectangular blank box. Those are the readers that we need to capture. Where does there comment go? The Talk page? A centralized feedback page? Either (or both) is fine ... but the article should retain the prompt for reader input. --Noleander (talk) 15:51, 25 January 2013 (UTC)
 * Comments

"As a result, when a reader does leave a useful comment it won't be noticed for days." Many talk page comments are not noticed for years. There are a lot of unwatched articles, and a lot of unwatched feedback. This is not surprising or a big deal, as volunteers have limited time. The tool already informs people that they can edit and encourages them to do so. And we certainly don't want to funnel all AFT5 comments onto the talk page, considering that they need to be filtered and curated to identify the useful ones. Dcoetzee 01:02, 28 January 2013 (UTC)

View by Legoktm
For various reasons explained by people above me, the editorial community does not want AFTv5. This does not necessarily mean our readers do not want it either. We need to take in considerations of our editors, but at the end of the day, we are here to serve the readers.

In the past, if someone had feedback, we would just point them to the talk page or give them. Sure their comments might go unanswered, but its arguably better than the current scenario with AFTv5.

A good middle ground would be to make it easier for users to submit feedback on the talk page, where most users will see it. This can be done using a javascript gadget or whatever. Adding an extra tab in the top toolbar, or maybe a small link at the bottom.


 * This has the advantage of not requiring any extra anti-abuse tools (no need for custom AbuseFilters), and bots like ClueBotNG will work just fine.
 * Edits can be tagged as "leaving feedback" for easy review
 * Revdel/supression are done using the standard interface, nothing different
 * Feedback shows up in the watchlist like normal
 * Comments can be left on it just like any other on-wiki discussion/post
 * Probably a bunch more I haven't thought of

tl;dr: Feedback is probably a good idea, but not in its current implementation. Legoktm (talk) 10:33, 21 January 2013 (UTC) The problem with this kind of view is that we can say all we like. Here's a great example of where that failed. Simon Wessely is a psychiatrist who worked on ME/CFS and has taken a fairly controversial view that has earned him a lot of anger from patients. The article we have on him is semi-protected due to vandalism and BLP violation. Similarly, the talk page is similarly protected. Recently, someone left feedback telling us that a link was broken in the article and giving us a replacement link that worked. ? He couldn't. Post on the talk page? He couldn't. How exactly would he have told us if AFT5 wasn't working? OTRS? The simple answer is: he wouldn't have told us, and we'd have a broken link in an article that we wouldn't have found for months or potentially years. —Tom Morris (talk) 13:03, 21 January 2013 (UTC)
 * Users who endorse this view
 * 1) Legoktm (talk) 10:34, 21 January 2013 (UTC)
 * 2) MZMcBride (talk) 03:48, 22 January 2013 (UTC)
 * 3) A simple widget that allowed a reader to type feedback in a text box, then added that text to the talk page as a new section titled "feedback, date/time", would give us the best of both worlds - easy feedback for the reader (no need to to know how to edit a talk page), and all the power of a talk page (reply, discuss, watchlist) for editors to respond with. Mitch Ames (talk) 12:46, 23 January 2013 (UTC)
 * 4) We have feedback, it is called "talk page". Improve, yes, but keep it simple, and do not ask for drive by "like" cliking; unlikely to be made by prospective editors and waaste editor's time - Nabla (talk) 22:06, 28 January 2013 (UTC)
 * 5) Enric Naval (talk) 12:29, 6 February 2013 (UTC)
 * 6) Exactly.Keith Roth (talk) 06:31, 7 February 2013 (UTC)
 * Comments
 * Why the fuck is a talk page protected? Talk pages should never be protected, that defeats the purpose of a talk page. If there's a lot of crap being put there, leave that to the people that maintain the page to deal with. I find the concept of a protected talk page, especially one in the mainspace, to be very disturbing.  S ven M anguard   <font color="F0A804">Wha?  17:43, 21 January 2013 (UTC)
 * A bit off-topic for here, but there are quite a few protected talk pages about - including one fully protected "until further notice" in August (most of the other full-protects are redirects, but there are lots of semis). Nikkimaria (talk) 17:50, 21 January 2013 (UTC)
 * @Tom: Thats definitely one of the success stories of AFTv5, except compared to all the other junk it picks up, it's an outlier.
 * @Sven: Most article talk pages that are semi'd are due to abuse by LTAs/sockpuppets/etc. There really isn't a better way to do it unfortunately. Legoktm (talk) 22:38, 21 January 2013 (UTC)
 * I disagree about the Simon Wessely article being cited as a success for AFTv5 (even as an outlier). If an article AND its talk page are protected that's obviously a problem, but it is a pretty rare problem. Rolling AFTv5 out for all articles just to fix this problem is patently absurd. I would suggest that AFTv5 as it is now be used for those articles where the talk page is protected, but for all other articles AFTv6 is needed. Aarghdvaark (talk) 13:47, 23 January 2013 (UTC)
 * Luckily the silver lining to this is that the vast majority of our 500 million more than a hundred million of unique readers a month don't like it either, or at least not well enough to use it. What we don't know is the proportion of readers who dislike it as a distraction from their reading. My assumption given the numbers involved is that those readers who dislike it are bound to be far more numerous than those who've used it. But I can't see the WMF testing that one.  Ϣere Spiel  Chequers 16:45, 23 January 2013 (UTC)
 * Well, we don't get 500 million unique readers a month, so it's not surprising we don't get 500 million unique uses. If we did, I'd be raising questions as to how it was that not only were we getting more users than the site has users, all of those users were making it down to the bottom of the article - which you seem to be assuming they would. So I think we can say with some degree of certainty that 'the vast majority of 500 million unique readers a month' is not a useful standard to be setting.
 * Your assumption was, in fact, tested by data - data that is available at the meta research portal that has been linked here. Okeyes (WMF) (talk) 16:54, 23 January 2013 (UTC)
 * Sorry my bad, the 500 million figure is all Wikimedia projects. The English language Wikipedia is only a subset of that, but a pretty big subset, big enough that my essential point holds true. Yes an unknown proportion of them will only be reading our minority of long articles and not reading to the point where they see the feedback box. But a large proportion of our articles are so stubby the feedback box is visible to anyone who clicks on the page. Would you care to put a figure on total unique visitors to EN wiki per month, and total unique givers of feedback?  Ϣere Spiel  Chequers 19:13, 23 January 2013 (UTC)
 * Another way of measuring this is to look at an individual article, as has been done at Requests for_comment/Article_feedback. If the total comments are from less than one in a thousand readers, and the worthwhile ones from less than one in ten thousand readers then it is reasonable to assume that the proportion of the other 99.9% of readers who don't like that intrusive feedback box is greater than the proportion who do.  Ϣere Spiel  Chequers 09:28, 24 January 2013 (UTC)
 * There are a number of problems with that argument. First; even with the 'pretty big subset' point of view, you're assuming full deployment. There's not full deployment :). What are the reader numbers of those articles where the tool is actually available? Second; looking at an individual article, any individual article, is pretty unhelpful to get an assessment of the tool as a whole. It varies widely from article to article, and we don't have reliable data on unique users on a per-article basis, only views. Those thousand or ten thousand users could well be a dramatically smaller number of people - who have submitted feedback and moved on. Fourth; you seem to be assuming a false dichotomy between 'uses the tool' and 'hates it'. Maybe users didn't have anything to say and recognised this. Maybe they only wanted to hit yes or no. Maybe they didn't scroll all the way down, maybe they're logged-in and have the tool turned off (the number of people with accounts and 0 edits who have recently used Wikipedia is actually surprisingly high; happy to pull hard data from the db for you if you so wish). These are all confounding variables it's very hard to control for: confounding variables that make the answer plausibly something other than 'they hate it'. WSC, when you end your argument with 'that intrusive feedback box' it feels a lot like you're seeking data to justify your beliefs rather than the other way around. Okeyes (WMF) (talk) 14:33, 24 January 2013 (UTC)
 * "What are the reader numbers where the article is available" - well just have a look at the example I referred to above 54,000 hits, 20 feedbacks that would have to be a pretty extreme outlier to rebut my contention, and if it was unusually low then the risk of the community being swamped by full rollout to AFT increases, if it was unusually high then that boosts the case for this not being relevant to the readers. As for my assuming a false dichotomy, remember I started with "What we don't know is the proportion of readers who dislike it as a distraction from their reading. My assumption given the numbers involved is that those readers who dislike it are bound to be far more numerous than those who've used it". Given the numbers involved I think that is a reasonable assumption. As for who is seeking data to justify their beliefs rather than the other way round, and how open AFT5 proponents are to testing whether AFT5 is really a good idea - well I just look at the history of meta:Research_talk:Article_feedback/Data_and_metrics and rest my case.  Ϣere Spiel  Chequers 15:21, 24 January 2013 (UTC)
 * That's less 'choosing not to test our assumptions' and more 'I don't think any of us saw that message', I'm afraid. I'm happy to discuss such a test with you if you want. I don't know if William Tell Overture is an outlier; my point is that because of limitations in both sampling and the data gathered, we don't know, and decisions shouldn't be made based on it. On the dislike front; again, there are a large number of explanations that retard the likely use of AFT5 that aren't "this interrupts my reading, get it off". Okeyes (WMF) (talk) 15:28, 24 January 2013 (UTC)

Actually, my point in the analysis of the William Tell Overture, was that the number of respondents in six months, 20, is so small compared to the number of page views (and lets face it, out those 54,000 in 6 months there must be at least several thousand unique readers) that we cannot make any assumptions about our readers in general from them or assume that the respondents, are typical of them. Thus, I think we can scratch the notion that they are providing valuable data about our readers. They are simply providing valuable data about the kind of people who click those buttons. They are a highly self-selected group, and it's pretty obvious from the comments, that the vast majority were not responding because they wanted to "help make Wikipedia better". Incidentally, those results are pretty much in line with the feedback on the other articles I watch. If anything, they have less outright vandalism, spamming, and stupidity. Likewise, I don't think we can assume that the unique readers who don't respond fail to do because they "dislike" having the feedback on the page. The tool may have value in other areas, but providing insight into our readership isn't one of them, in my view. Voceditenore (talk) 19:34, 24 January 2013 (UTC)
 * Thanks for the clarification, Voceditenore :). That makes a heck of a lot of sense; I think interpreting things to mean 'here's info about our readers' is probably a waste of time, and not really what we were intending to do. 'What a self-selecting group of readers who speak English and have something useful and coherent to say would like to see on an article' is probably closer to the target. Okeyes (WMF) (talk) 19:54, 24 January 2013 (UTC)
 * Yep, that's an OK target :) My view and analysis were basically a response to Mike Cline's view above which made some pretty extravagant claims about the Feedback tool becoming a "tremendous source of data on how the reader community views wikipedia" and allowing "improvements in policies and guidelines based on empirical data from the reader community". I kind of doubt that Wikipedia readers could be called a "community" in any sense of the word, and positively cringe at the thought of using this skewed data to inform our policies. In any case, although I have some pretty big misgivings about the cost/benefit ratio of the tool, I doubt if I'll come down on either side in this RfC. Voceditenore (talk) 14:54, 25 January 2013 (UTC)
 * Yeah; as much as I like having people support a roll-out - it was never intended to inform policy, guidelines or the sort of general gestalt: it's about informing individual pieces of articles, with individual issues, handled by individual editors. I do not expect it to have any impact on how we write policy, and if it did I'd be sat in GregjackP's column opposing it, because we have a lot of the policies we do for very good reasons. Taking AFT5 to represent widespread change based on input from an identifiable group of 'readers' with uniform desires, needs and opinions is, I think, incorrect. A better analogy would be treating it as an analogue to the OTRS system: it handles individual problems, from individuals. We don't change the BLP policy because we get a lot of BLP tickets :). Okeyes (WMF) (talk) 15:07, 25 January 2013 (UTC)

The main issue with spilling all feedback onto the talk page is that a lot of it is nonconstructive, and talk pages are very primitive. You can't easily present dynamic views such as rearranging things based on usefulness, and comments would get rapidly archived on popular articles due to high comment frequency. It's a lot easier to organise feedback in the more advanced AFT5 moderation system. Dcoetzee 01:07, 28 January 2013 (UTC)
 * I'd argue that the feedback tool is even more primitive, as it doesn't allow replying to the feedback at all, let alone having a cognent discussion of the merits of the suggested edits..... You can't label anything as "solved" or "won't solve", with an explanation. Even if we could do that, we would still need a mechanism to contest the resolution, discuss it, etc. And we need to point to old discussions, cite sources, and indent replies, archive old feedback, etc. We would be reinventing the talk page, but in a clumsy and constricted manner. --Enric Naval (talk) 12:29, 6 February 2013 (UTC)

View by Apteva
I am looking forward to a full roll out of this useful feedback tool and am disappointed that it was delayed until March. I regularly check reader comments where they are available and use them to improve the article.


 * Users who endorse this view
 * 1) Apteva (talk) 22:11, 21 January 2013 (UTC)
 * 2) -- Guerillero &#124;  My Talk  00:29, 22 January 2013 (UTC)
 * 3) WhatamIdoing (talk) 19:27, 22 January 2013 (UTC)
 * 4) — This, that, and the other (talk) 10:25, 23 January 2013 (UTC)
 * 5) --Neo139 (talk) 21:15, 27 January 2013 (UTC)
 * 6) Support. I also regularly use article feedback to improve articles, and also created User:LyricsBot on the basis of reader feedback which has thousands of edits. Dcoetzee 01:08, 28 January 2013 (UTC)
 * 7) support FeydHuxtable (talk) 18:32, 30 January 2013 (UTC)
 * 8) Support, don't throw the baby out with the bathwater. Abductive  (reasoning) 04:20, 5 February 2013 (UTC)
 * Comments
 * You do? I [//en.wikipedia.org/w/index.php?title=Special%3ALog&type=articlefeedbackv5&user=Apteva&page=&year=&month=-1&tagfilter=&hide_patrol_log=1&hide_review_log=1 only see] 2 feedback log entries for your account. Legoktm (talk) 22:27, 21 January 2013 (UTC)
 * Not everyone flags everything -- Guerillero &#124;  My Talk  00:29, 22 January 2013 (UTC)
 * I rarely flag them in any way but I do use them in looking for ways to improve the article. Apteva (talk) 03:03, 22 January 2013 (UTC)
 * You'll find nearly 500 entries in my log so far, and if you compare them to my edits you'll find a significant correlation, but I read almost all the feedback, even if I don't tag it. WhatamIdoing (talk) 19:27, 22 January 2013 (UTC)

Questions from Risker
I think we need to gather some more information, both from those developing this tool, and from those who have been using it, before we're really in the best position to make a good decision here, so this is a "questions" section rather than one that states a view. Please feel free to respond based on your knowledge or experience, as "comments" rather than supports or opposes. Risker (talk) 07:32, 22 January 2013 (UTC)

Technical/statistical questions

 * In my review of multiple pages' worth of feedback, I note that somewhere between 30-35% of the feedback contains no comment; the correspondents have apparently just clicked the "yes" or "no" button. Is there a way for these to be automatically "resolved" (moderated) so that they are removed from the feedback queues?
 * I spoke to Matthias, our developer, about this today - the likely resolution is that (pending slowdowns in the development timetable or other distractions) we'll:
 * (1) massively reduce in size the boxes for that sort of feedback;
 * (2) remove the moderation tools (to make clear that, well, there's nothing to do here;
 * (3) add something along the lines of "this feedback was submitted without a comment" for added clarity. Okeyes (WMF) (talk) 08:03, 22 January 2013 (UTC)
 * If the feedback has no content, why should it appear anywhere but in the percentage tally? 71.212.247.146 (talk) 16:32, 23 January 2013 (UTC)
 * Because the alternative can be a big discrepancy, where you might only see two comments (both positive) but the percentage tally comes up negative. There needs to be some representation of comment-free feedback to avoid confusion on this front. Okeyes (WMF) (talk) 16:34, 23 January 2013 (UTC)
 * Most recent statistical information (as posted on the talk page) indicates that 11.5% of feedback is moderated within one month, and that approximately 90 users/day are regularly moderating messages, and that AFT5 is currently present on approximately 10% of all articles. Based on these facts, which of the following scenarios are being considered in the decision to take AFT5 to 100% of articles: (note—there may be additional applicable scenarios, which can be added and commented upon)
 * If we increase the number of articles tenfold, we will likely need to increase the number of editors moderating messages tenfold (i.e., to 900 users/day) to maintain the current level of moderation
 * If we increase the number of articles tenfold without increasing the number of moderators, the percentage of feedback moderated is likely to be reduced to 1/10 of the current level (i.e. 1.15% moderated within one month).
 * Both of those are subject to debate and discussion, and boosting moderator numbers (and making it a lot easier for a moderator to triage feedback) have been prioritised as goals in our development efforts. So, on the first front, we're talking about a logged-in-only prompt on the article page (out of the way of the article text) when a page has unmoderated feedback. On the second we're removing a lot of the complexity from the feedback page - we'll have a prototype in a few weeks, but in terms of being able to show off what we've planned this is sort of the worst possible time :/. Not the fault of those who started the RfC, obviously, who did not know this.
 * I would say that we're likely to need less than a 10-fold increase to achieve the same level of moderation, because our random lottery system quite probably overrepresented high-traffic articles. But at the same time we're going to need an increase in moderator activity whatever we do
 * How can administrators remove AFT5 from clearly inappropriate pages NOW? (I noted AFT5 on a redirect page and a Wikipedia space page in the last 24 hours.)
 * Category:Article Feedback Blacklist /should/ work - on the Wikipedia space page, which one? We enabled feedback for help pages, for example. Ditto the redirect: those are meant to be blocked off, so I need to find out what's going on there :/. Okeyes (WMF) (talk) 08:03, 22 January 2013 (UTC)
 * The data provided on the talk page indicated that 40% of comments were found to be useful by those carrying out the reviews. This seems remarkably disproportionate to the random viewing of 60 feedback pages I carried out entirely empirically in the last 72 hours, where I would generously estimate "useful" feedback to be less than 10%. Can you provide more information on the raw data that resulted in that figure? What comments were being considered useful, and in what context? Was feedback that did not include a comment included or excluded in the study? (I suggest here that there are widely divergent definitions of "useful" rather than that there is a problem with the data.)
 * Can you explain "entirely empirically"? People were given a set of feedback posts from all articles using a randomised sampling mechanism; this excluded comment-less feeddback. They were presented with four different ways to mark the feedback: "useful" "unuseable" "inappropriate" and "oversight" (where useful was defined as 'this comment is useful and suggests something to be done to the article'. Reading it now we should probably have gone for 'something valid to be done to the article', but it's a bit late now). I'm in total agreement that there are divergent definitions of useful, as is the data: if you read the study I linked in my comment you'll see that two users involved in the test agree on usefulness around half the time. Okeyes (WMF) (talk) 08:03, 22 January 2013 (UTC)
 * I'd be chary of non-expert opinions on whether a particular comment is valid; on specialist topics (eg bovine papillomavirus), it's likely you would need to know the subject to decide. Espresso Addict (talk) 17:14, 22 January 2013 (UTC)
 * Even on such articles you are much likely to get feedback like add more references or need more about the history of discovering than, I don't know, No, prophylactic vaccination of spiromambalus is not going to cause severe strinosomic injuries. And though you may not be sure what-the-heck-are-they-talking-about, just the point of someone saying something is wrong there leads to ok, useful comment - either there really is something bad or the article just needs to be more clear to not cause such misinterpretations. --11:31, 23 January 2013 (UTC), Utar (talk)


 * The call to action statistics (i.e., edit attempts through various means) show that edit attempts from AFT5 occurred far less frequently than attempts from either the page top or section edit tabs, and successful (saved) edits from AFT5 were more than twice as likely to be reverted than successful edits from either type of edit tab. There was a 0.2% conversion to a successful, unreverted edit from those who initiated the feedback process by clicking yes/no as compared to a 3.9% conversion to a successful, unreverted edit through an edit tab. At one point, there was talk of AFT5 being one of the tools intended to engage new editors. Has this been rethought?
 * No. So, yes; the success rate is lower. But the only way that can make an impact (well, except in a cost/benefit of good versus bad edits) is if the people participating in the article feedback tool are being cannibalised away from section edit links and page edit links. m:Research:Article feedback/Stage 3/Conversion and newcomer quality (I appreciate it's written in research-ese) provided pretty strong evidence that this cannibalisation is not happening. The tool is certainly less successful than I (and, I think, others) hoped on this front, but in isolated terms of conversion it is still a valid mechanism for newcomers. Okeyes (WMF) (talk) 08:08, 22 January 2013 (UTC)
 * There is, I think it is fair to say, non-negligible resistance to applying AFT5 to all articles. With this in mind, would you consider alternatives such as application to certain categories of articles and/or the ability of editors or Wikiprojects to have AFT5 applied on an opt-in basis? Would you consider creation of a blacklist of articles for which AFT5 is not appropriate?
 * I'd also agree that's fair to say ;p. This blacklist already, actually, exists: I am perfectly comfortable discussing any and all compromise solutions, although obviously I can't make promises if they involve developer time and will have to refer things up the food chain. Application to certain categories, opt-in...if people want it, I can look into it. Okeyes (WMF) (talk) 08:08, 22 January 2013 (UTC)

Questions for people using feedback

 * Do you read feedback without moderating it? If so, why do you not moderate it?
 * Yes -- I'm not always sure whether it would be useful for another editor, or what counts as useful. Especially extremely brief comments like "Good" or "more info". -- Avocado (talk) 15:40, 22 January 2013 (UTC)
 * Yes, frequently. Sometimes the suggestion is valid but not something I want to do or think is very important, like Special:ArticleFeedbackv5/Bread/839323.  I didn't resolve it, so I can't mark it resolved.  It doesn't seem important enough to feature it.  It's not vandalism or libel, so it doesn't need hidden or oversighted.  Those are the only four buttons.  There is no "I read this, and there's nothing wrong with it" button.  So I leave it alone.  What would you do with feedback like that?  WhatamIdoing (talk) 19:38, 22 January 2013 (UTC)
 * Yes, mostly. Most of the stuff I have no idea how to describe/categorize it or know if it makes sense to others. Some of the stuff like blank pages(a few weeks back) should be auto-filtered. What to do with foreign language text that means nothing to me etc? Regards, Sun Creator(talk) 02:21, 23 January 2013 (UTC)
 * I moderate feedback when I can, but often can't. In many cases, I can't tell whether a given feedback item is "helpful" or not. A large fraction of feedback is neither very helpful nor completely worthless.--Srleffler (talk) 04:12, 23 January 2013 (UTC)
 * No interest in nannysitting the comments, but I avidly use them to improve the article. Apteva (talk) 08:01, 23 January 2013 (UTC)
 * I do read feedback but I have no idea what moderation means in that context. Olli Niemitalo (talk) 00:52, 26 January 2013 (UTC)
 * Are the instructions for moderating feedback easily understood? Do you feel you could look at an article's feedback and moderate a list of 20 comments?
 * Not really. -- Avocado (talk) 15:40, 22 January 2013 (UTC)
 * Sure, I have done that, but the instructions aren't complete. It's a new tool, and we improve them as editors figure out what works.  WhatamIdoing (talk) 19:38, 22 January 2013 (UTC)
 * No, for example I requested oversight too liberally until Beeblebrox asked me to stop. Abductive  (reasoning) 21:24, 22 January 2013 (UTC)
 * I think the feedback response guidelines are less than clear. You can kind of figure it out if you carefully read that page and think about it a decent amount.  For instance, it is not obvious that you should just "resolve" non-actionable feedback.  Youd need to read most of the guidlenes article and think about how it interrelates to discern that.  People shouldnt need to put on their serious thinking caps for such a task.  Those guidelines need a significant improvement, though I support the tool in general. <font color="DarkGreen">AgnosticAphid  talk 00:21, 23 January 2013 (UTC)
 * Understanding is one thing, but the problem is no sense in what to do with much of the feedback. Regards, Sun Creator(talk) 02:21, 23 January 2013 (UTC)
 * It took me a while to figure out how to moderate feedback: when to mark items as "helpful" or not, and marking unactionable entries "resolved". Clearer and more direct instructions would have been helpful.--Srleffler (talk) 04:12, 23 January 2013 (UTC)
 * There's definitely room for improving instructions, some kind of "newbie guide to moderating feedback," but I found it quite straightforward from the get-go. Dcoetzee 01:18, 28 January 2013 (UTC)


 * Have you personally improved articles based on feedback? How frequently? How many articles? Could you give some examples?
 * Yes and fairly frequently. See feedback here for what was called the "James Bond (character)" article. After feedback from a number of IPs asking for information about the films, the article was renamed as "James Bond (literary character)". Subsequent to that, and as we still had feedback asking about the films, a new article, James Bond filmography, was written focussing on the film; this will be going forward for GA shortly. See also the the Bond music feedback: 11 pieces of feedback; three are valid comments and good suggestions (good enough that I will incorporate them when I get round to the re-write). One was a suggested correction (which I undertook), three were praises, one complaint and three understandably misguided requests (more pictures, links to the songs etc). That's a damned site better treatment than the Bond articles get, with vandalism, mindless additions of fancruft and disruptive editing. I'd much rather have the feedback facility than any other bells & whistles that are attached to a page (such as the fairly pointless and ineffectual page ratings facility). - SchroCat (talk) 15:41, 22 January 2013 (UTC)
 * Yes, a couple of times since the tool was launched, which is pretty much all the directly actionable comments I've seen in articles on my watchlist - the great majority are random comments, inappropriate or misguided questions. Nick-D (talk) 10:21, 22 January 2013 (UTC)
 * I'll echo what Nick said. There have been a few instances of useful comments, but most are chaff (I listed some examples in my view, above).   GregJackP   Boomer!   12:46, 22 January 2013 (UTC)
 * No. -- Avocado (talk) 15:40, 22 January 2013 (UTC)
 * Yes, dozens of times. Nearly every image I've added in the last few months is due to article feedback.  I believe that every change I've made to High school diploma during the last six months was directly prompted by feedback.  WhatamIdoing (talk) 19:50, 22 January 2013 (UTC)
 * A lot, but I have to comb through many pointless ones. Abductive  (reasoning) 21:24, 22 January 2013 (UTC)
 * Very rarely. For instance, a user requested an infobox, which I created. However, the vast majority of feedback items are not actionable. GabrielF (talk) 00:28, 23 January 2013 (UTC)
 * Yes (why I love the feedback), some things I would not have thought about such as alternative names used in other cultures. Some of the feedback is indeed awesome. Regards, Sun Creator(talk) 02:21, 23 January 2013 (UTC)
 * Yes, but only small changes.--Srleffler (talk) 04:12, 23 January 2013 (UTC)
 * Yes, occasionally. Eg: Mitch Ames (talk) 13:15, 23 January 2013 (UTC)
 * Very infrequently, last time accidentally today.--Ymblanter (talk) 15:32, 23 January 2013 (UTC)
 * once every couple of weeks a suggestion will be something appropriate and that I can act on. a request for some information about the climate/weather on Brooks Range led me to do some research and add that information. someone wanted pictures in Ape and so I was able to add some images. -- TRPoD aka The Red Pen of Doom  23:04, 23 January 2013 (UTC)
 * Yes, many. I have added images and specific information with references, added links, resolved confusing text, and created at least one bot (User:LyricsBot) with thousands of edits inspired by reader feedback. Dcoetzee 01:18, 28 January 2013 (UTC)


 * How frequently do you see feedback that may seem useful on the surface but is inappropriate to the particular article? (Examples: requests for images in articles that already have 10+ images or on articles where only a fair-use image would be appropriate; information that is already in the article)
 * A fair amount - normally from people asking for extra images, links to clips or downloads of entire books (WP:LINKVIO infringements) - SchroCat (talk) 08:30, 22 January 2013 (UTC)
 * Quite frequently - for instance, the World War II article is attracting a steady stream of requests for photos (it already has lots), very detailed information about individual countries or specific topics (the article is written at a high level as it's the only way of covering the whole war in a non-vast article) or requests to weight it to one point of view or another (typically based on advocating one form of nationalism or another). Other commenters also want chunks removed from the article, generally as they appear to find that topic uninteresting, and lots of school students want us to write their essays for them. Nick-D (talk) 10:21, 22 January 2013 (UTC)
 * Probably 30-50% of feedback -- Avocado (talk) 15:40, 22 January 2013 (UTC)
 * The majority of what I've received has been requests for information to be added that doesn't belong in the article (eg bovine papillomavirus). I suspect specialist articles receive less useful feedback. Espresso Addict (talk) 17:24, 22 January 2013 (UTC)
 * What I see is either obviously appropriate (e.g., an article about a disease that doesn't talk about the prognosis) or obviously inappropriate (e.g., people wanting advice on how to get a copy of their high school diplomas). There is very little that seems plausible initially but later turns put to be worthless.  WhatamIdoing (talk) 19:50, 22 January 2013 (UTC)
 * I see comments meant for completely different articles. Also, why in the f*ck are non-article pages such as Help:Using talk pages included in Article Feedback? Abductive  (reasoning) 21:24, 22 January 2013 (UTC)
 * Yes, but I don't see a problem with it. Readers are expressing what they would like and in some cases I agree with the reader rather then the wiki rules. Regards, Sun Creator(talk) 02:21, 23 January 2013 (UTC)
 * Very often. I mostly edit physics articles.--Srleffler (talk) 04:12, 23 January 2013 (UTC)
 * Sometimes. Eg: Mitch Ames (talk) 13:15, 23 January 2013 (UTC)
 * Once. The articles I watch do not get that much feedback. Olli Niemitalo (talk) 00:52, 26 January 2013 (UTC)
 * These are fairly common and expected. Some readers ask us to include full lyrics of copyrighted songs in the article (I normally resolve these by adding external links to licensed providers). In some cases, I have considered proposing policy changes on the basis of reader feedback (e.g. some readers ask for street addresses of businesses, which we don't include currently but we could). Dcoetzee 01:18, 28 January 2013 (UTC)


 * If you are reading feedback for an article, and you see a comment that looks useful but you are not in a position to pursue it, what do you do?
 * Log it for later and deal with it in an article re-write. See the Bond music feedback. Excellent feedback by and large with people saking pertinant questions (which songs received awards; I would like to have seen a bit more detail on the style of music; Chart positions for all James Bond single releases) When I get round to a re-write of the article these will all be taken into account. The only thing I've done off the list so far is to capitalise a band name, which was requested. - SchroCat (talk) 08:30, 22 January 2013 (UTC)
 * Nothing; beyond copying and pasting it onto the article's talk page (which runs against most editors preference for DIY-type responses), there's no efficient way to log these kind of things or sort the wheat from the chaff - the 'favourite' function is being applied randomly, and most comments marked as such aren't actually useful. Nick-D (talk) 10:21, 22 January 2013 (UTC)
 * Mark it as useful, and if especially insightful, feature it. -- Avocado (talk) 15:40, 22 January 2013 (UTC)
 * Mark it as useful. Abductive  (reasoning) 21:24, 22 January 2013 (UTC)
 * Look for something else I could do something with. Regards, Sun Creator(talk) 02:21, 23 January 2013 (UTC)
 * Mark it "Helpful". If it's really good, I Feature it.--Srleffler (talk) 04:12, 23 January 2013 (UTC)
 * Put an item on the article's talk page, eg to ask someone to fix it - if it's something that I feel is likely to be fixable, but I don't have the info to do it. Example - someone wanted a Russian translation of a fairly short article on English Wikipedia, so I put a request on the (English) talk page in case someone bilingual felt inclined to translate and add it to Russian Wikipedia. Mitch Ames (talk) 13:38, 23 January 2013 (UTC)
 * Nothing. Everything I see is pretty obvious, like requesting to add images when there are no free images available (and if they show up, they will be immediately added to the article) or to add info on the family situation of a BLP for which no reliable sources were found (at some point, reliable source will write about the family situation, and the info gets to the article).--Ymblanter (talk) 15:36, 23 January 2013 (UTC)
 * I only ever found one useful feedback comment, and I posted it on the talk page for consideration. <font face="trebuchet MS">- filelake<font color="#0B0">shoe <font color="#0B0"> &#xF0F6;    11:43, 25 January 2013 (UTC)
 * Nothing.Olli Niemitalo (talk) 00:52, 26 January 2013 (UTC)
 * I mark useful and feature comments like this. This moves them toward the top of the list and makes them easier for others to find and act upon. I sometimes also leave a note on the talk page with a link to the comment and my own thoughts on the matter. Dcoetzee 01:18, 28 January 2013 (UTC)


 * Right now, AFT5 is on a limited and controlled number of articles. Are there articles you watch or otherwise monitor where you think AFT5 would be particularly useful? Are there articles on your watchlist that currently have AFT5 applied, where you find it is significantly unhelpful and would like to remove it?
 * I'd like to see it rolled out to all articles. If I don't agree with it or don't like it then I ignore it. If I think there is a valid point for it, I'll action it. - SchroCat (talk) 08:30, 22 January 2013 (UTC)
 * It's pretty unhelpful on the full range of articles I've seen it deployed on (which range from the 20,000 page view a day World War II article to articles on obscure topics which only specialist readers would ever take an interest in or comment on), so no. Nick-D (talk) 10:21, 22 January 2013 (UTC)
 * It's not useful on the articles where I've seen it.  GregJackP   Boomer!   12:48, 22 January 2013 (UTC)
 * It is more useful on low-view articles. Having it on YouTube and Minecraft is insane. Abductive  (reasoning) 21:24, 22 January 2013 (UTC)
 * Given this comment, I'm a little surprised that you didn't sign on to my compromise view above in which this feedback tool would be opt-in per-article. --MZMcBride (talk) 19:42, 29 January 2013 (UTC)
 * I would prefer it to be excluded from all articles above a certain number of views per day, all GAs and FAs, and all articles above a certain size. Possibly all BLPs too. It works very well for stubs, medical/scientific/geographical/history topics. Abductive  (reasoning) 04:19, 5 February 2013 (UTC)
 * I monitor about 8000 articles but only about 200 are on my watch list - and would like to expand feedback to show related articles. Still a scalability issue for me. Regards, Sun Creator(talk) 02:21, 23 January 2013 (UTC)
 * I'm not convinced that the feedback tool is useful.--Srleffler (talk) 04:12, 23 January 2013 (UTC)
 * I'd like to see it on articles, but it'd be especially handy on math-related articles, where it helps to point out topics that need to made more accessible. Dcoetzee 01:18, 28 January 2013 (UTC)


 * Have you encountered inappropriate feedback comments? What did you do about them? Did you hide them or submit them for oversight?
 * Yes, especially on BLPs (especially when the subject of the article is involved in some kind of scandal - for instance, politicians who mess up and come under attack from their opponents). I've hidden lots of comments, but can't remember asking that any be oversighted, though in retrospect I should have. Lots of other comments are not directly offensive, but are aggressive advocacy of one position of another related to the topic of the article and would be moderated if posted on a general discussion forum (for instance, the Atomic bombings of Hiroshima and Nagasaki article is attracting regular comments declaring the United States a 'terrorist nation' or similar; nothing useful can be done with such ranting, but it's not worth the effort of hiding). Nick-D (talk) 10:21, 22 January 2013 (UTC)
 * Occasional irrelevant cursing, or nonsense/garbage ("fjdskahfjsah"). Mostly marked as unhelpful or hid them. -- Avocado (talk) 15:40, 22 January 2013 (UTC)
 * It is surprisingly common to see feedback where the reader has a very specific personal request and wants someone to call them and help them. Users provide their phone numbers and other personal details (in some cases government ID numbers!) We need to have better instructions for users so that they can understand that this is feedback about the ARTICLE not a place to ask general questions about a topic. GabrielF (talk) 00:31, 23 January 2013 (UTC)
 * I hide some, but very few times have I encountered really inappropriate(as opposed to nonsense) feedback. Regards, Sun Creator(talk) 02:21, 23 January 2013 (UTC)
 * Yes. Hid.--Srleffler (talk) 04:12, 23 January 2013 (UTC)
 * Blatant abuse/insult I flag as such. (eg "[politican's name] is a [derogatory term]"). Obviously useless comments I mark as "not helpful". Mitch Ames (talk) 13:15, 23 January 2013 (UTC)
 * I hide any feedback comment which I would roll back as a talk page comment. <font face="trebuchet MS">- filelake<font color="#0B0">shoe <font color="#0B0"> &#xF0F6;    11:43, 25 January 2013 (UTC)
 * I hide any feedback comment which is nonconstructive or contains no useful information (e.g. commenting on the topic not the article), but I leave compliments as they can be rewarding for page maintainers. I don't normally submit for oversight unless personal details of minors are included. Dcoetzee 01:18, 28 January 2013 (UTC)


 * Did you read/review/moderate AFT5 comments earlier but then stop doing so? Why did you stop?
 * I read them quite a bit when the tool was first introduced, but have largely given up on doing so due to the very low relevance of most comments and my concerns about a reader/editor split developing as a result of this form of interaction (the idea of 'editors' taking instructions from 'readers' is pretty unattractive to me). Nick-D (talk) 10:21, 22 January 2013 (UTC)
 * Signal to noise ratio is extremely poor. Something about how the questions are being asked of the readers is encouraging one-word or two-word and yes/no comments, which aren't helpful to editors and not worth moderating. -- Avocado (talk) 15:40, 22 January 2013 (UTC)
 * Yes, it was to much thought required to explain why I wanted to hide them. Regards, Sun Creator(talk) 02:25, 23 January 2013 (UTC)
 * I do so less, because it just became such a waste of time due to 99% of the comments being completely useless. <font face="trebuchet MS">- filelake<font color="#0B0">shoe <font color="#0B0"> &#xF0F6;    11:43, 25 January 2013 (UTC)
 * I start and stop based on how much free time I have for it. Dcoetzee 01:18, 28 January 2013 (UTC)

View by GabrielF
I can offer a qualitative assessment from reading through many, many feedback posts. The vast majority of feedback is not particularly helpful. There are a lot of people who leave nonsensical or one-word or even blank feedback. The most recent post on an article on my watchlist is "I JUst Wasted afew secconds of ur life". Well, thanks buddy. There are a lot of people who leave feedback saying that they liked the article. This is gratifying but is it helpful? A recent example: "you guys rock". Well, thanks poster. Many posts do include suggestions for improvement: by far the most common are requests for pictures, requests that the article be written in simpler language, and request for the contact information of the article subject. My sense is that none of these are particularly useful. We often lack pictures because freely-available photos are not available. Most likely the editors who are responsible for an article want a photo but they can't find one that meets our licensing needs. A feedback suggestion is not likely to change anything. Similarly, the issue of language complexity is not necessarily something that we can address. Recent feedback on W.E.B. DuBois: "Make more kid-friendly". Well, DuBois lived a long and influential life and he advocated complex ideas. Using simpler language would probably make the article less encyclopedic. And we don't provide contact information by policy. Another common theme is a generic request that the article be expanded: For instance recent feedback on Roger Sherman: "have a bigger selection of information". I appreciate the feedback, but how is this actionable?

So far I've found a relatively small number of cases where users have provided feedback that was helpful and actionable. One user asked for an infobox on an article about a federal judge, and I was happy to create it. Similarly a couple of users have pointed out that after reading Ariel Sharon they are unclear whether he is alive or dead - that suggests that we should change the language in that article to be clearer. Having said that, I'm not sure whether this feedback reflects the quality of the writing or the capabilities of the reader. The first paragraph of the article says: "He has been in a persistent vegetative state since suffering a stroke on 4 January 2006." That seems quite clear to me.

My overall impression is that there is some specific, actionable information coming from these feedback requests, but that it is buried in a mountain of crud. The feedback provided by this tool is more distracting than helpful. Many of the people providing feedback just don't seem to get what we're about.

I'm concerned that we're providing too many means for people to ask questions and provide feedback and that, as a result, they aren't reading carefully and going to the right places. If someone uses a feedback request to challenge a CSD, I would guess that they aren't taking the time to read the CSD notice carefully because they see a button where they can submit feedback and they think this will solve their problem. Similarly, the really useful feedback - specific, actionable items - belongs on the talk page where editors are more likely to see and discuss it. I am concerned that it is going to be buried in feedback tool. We also get questions from people who are looking for information: e.g. recent feedback for Mexican passport "how to renew mexican passport". Those kinds of questions should go to the reference desk.

My specific suggestion for improvement is to modify the feedback request template to better direct people to appropriate forums for their request, for instance by saying something along the lines of:
 * Is there something about this topic that you would like to find out? Try the WP:Reference Desk
 * Is there a specific mistake in this article or a new source that we can use? Try the Talk Page.
 * If you have general feedback, leave it here.

Overall, this is an interesting experiment, but reading through the feedback that our readers have provided leaves me more discouraged with our readers than brimming with new ideas for improving articles. GabrielF (talk) 17:33, 22 January 2013 (UTC)


 * Users who endorse this view
 * 1) I agree with most of what you have written. --Srleffler (talk) 04:22, 23 January 2013 (UTC)
 * 2) &bull; Jesse V.(talk) 16:37, 24 January 2013 (UTC)
 * 3) Agree, This is the best analysis I have seen; it is also one of the few which recognize there is no single solution.  DGG ( talk ) 01:12, 30 January 2013 (UTC)
 * 4) Support. – Wdchk (talk) 15:24, 2 February 2013 (UTC)
 * 5) Yes!. That feedback disambiguation is priceless. Reference Desk must be one of the "better kept secrets" on Wikipedia. WMF needs to work on better aggregation algorithms for the "general feedback" – Wbm1058 (talk) 04:34, 9 February 2013 (UTC)


 * Comments
 * Thanks for posting this. It's been very helpful to read various users' experiences with this tool. Your experience seems to match the experience that many other users have had. --MZMcBride (talk) 18:57, 22 January 2013 (UTC)
 * Second that. &bull; Jesse V.(talk) 16:37, 24 January 2013 (UTC)

I would agree with this post if not for proposed amendment to the template to redirect the user giving specific feedback to the talk page. Aggregating these sort of suggestions is the point of the tool in the first place! Abyssal (talk) 20:10, 24 January 2013 (UTC)
 * What the fuck for? How does aggregation suggestions in some arcane stand alone system instead of getting actionable ones out front (actually in back on the talk page) where editors can actually action them help improve anything? -- TRPoD aka The Red Pen of Doom  23:21, 25 January 2013 (UTC)
 * Send the "I want to see a photo" comments to the AFT "general feedback tool," which later reports, "17 readers requested a photo." Send the reader who says, "I have a picture I took myself, I want to donate it to Wikipedia, but I don't have a clue about how to upload it to the article" to the talk page. Wbm1058 (talk) 04:34, 9 February 2013 (UTC)
 * To assist with aggregation of the obvious questions, per, use specific radio buttons for:
 * "Needs images" – direct high traffic on this button to Category:Wikipedia requested photographs (which is normally populated by Image requested placed on talk pages), after moderation for appropriateness of image requests
 * "I couldn't understand it" – high traffic on this button may mean that a quality Simple English Wikipedia article on the topic, linked to by a hatnote on the more complex main article, is merited—Wbm1058 (talk) 16:11, 19 February 2013 (UTC)

Comment by Guerillero
After reading over the comments here I feel that we are judging our readers too harshly and by the wrong standards. All of us are "power users." We know the policies, rules, and essays that the culture of wikipedia is built upon. Our readers, the people who leave feedback, do not know our mores. For example, a common piece of feedback is that article X needs more pictures. Yes, this would often violate the NFCC to add more images in many articles, but how would John Doe in Iowa know that? This discussion shows that there is an ever widening gap between readers and editors. -- Guerillero &#124;  My Talk  19:28, 22 January 2013 (UTC)
 * I agree with you that there is a large gap between readers and editors. Maybe there are some conclusions that can be drawn from article feedback about Wikipedia's interface in general: maybe we should emphasize the Simple English wikipedia more. Maybe we should try to explain more clearly what an encyclopedia is and how it differs from a dictionary or a business directory. This tool might provide excellent feedback for people at the Foundation and editors who are looking at big-picture ways to improve Wikipedia. However, if the target audience of this tool is editors who are looking for suggestions on how to improve articles, than I'm not sure that hearing what the average person who can click a button says is all that helpful. Most editors know that we need better pictures and most editors want to find better pictures, the limitation is the availability of free content. Hearing it from a reader isn't telling us anything new.GabrielF (talk) 19:39, 22 January 2013 (UTC)
 * I don't intend any judgment of our readers at all when I say that the feedback we're getting is (in my view) mostly not useful. Guerillero is entirely correct that they have no way to know our policies - but the thing is that that doesn't matter. It doesn't matter if they're somehow being deliberately obtuse, or if they genuinely don't understand how Wikipedia works - in both cases, there is no gain from letting them leave that "useless" feedback, and a fair amount of loss because the rest of us have to monitor it. It's not wrong for them to, say, want more photos, but it's not something that helps us either improve articles or draw in new editors. It's not something that's going to get responded to or fixed, and it's not going to teach the feedback-leaver anything. And if that's the case...why are we letting them leave that feedback in the first place? They're shouting (innocently) into a void. A fluffernutter is a sandwich! (talk) 19:53, 22 January 2013 (UTC)
 * No, you don't "have to monitor it". You can ignore it.  If it's not helpful to you, then your volunteer time should be spent elsewhere.  WhatamIdoing (talk) 19:57, 22 January 2013 (UTC)
 * I've found Wikipedians are pretty protective of Wikipedia, surprisingly. We will not allow it to be filled with garbage (libel, vandalism, spam, and other nonsense) in an attempt to fulfill an ill-defined and poorly executed goal of engaging editors by adding a comments box to every article. Sorry. --MZMcBride (talk) 20:01, 22 January 2013 (UTC)
 * But someone has to monitor it, so that the someone can catch all the other feedback that says 'I did ur mum' or whatever. In that sense, every feedback entry adds to the workload, because it's one more entry that needs checking (checking to see if it's actionable, to see if it's abusive...checking for any number of things). And speaking as someone who's tasked with oversighting the worst of these entries...yeah. Someone has to monitor the feedback submissions. There's stuff in there that must be caught and suppressed, by people like me who do those suppressions multiple times a day, because the bad feedback keeps coming. Saying I don't have to be the one monitoring it doesn't change the fact that it has to be monitored. A fluffernutter is a sandwich! (talk) 20:03, 22 January 2013 (UTC)
 * True, it would be better to monitor comments than to let people post garbage. But that work should be done by one of the many volunteers who believe feedback to be a good use of their time, not by a volunteer who believes it to be a waste of time.  If you're in the second camp, then I think that you should stop dealing with feedback and do something that you believe is more valuable.  WhatamIdoing (talk) 21:55, 23 January 2013 (UTC)
 * So far, we have 1 in 500 making a non-reverted edit, and more than that making an effort that gets reverted. That means that we are drawing in a small number of new editors with this. If everyone decided that the feedback was useless, then feedback containing libel (of which I've seen exactly none, by the way) might go unchecked. But "Fluffernutter doesn't have to monitor it" is not exactly the same thing as "everyone refuses to monitor it".  WhatamIdoing (talk) 20:04, 22 January 2013 (UTC)
 * As discussed above, the reason you're not seeing as much libel as you typically would is that Oliver and his team have implemented a number of very heavy-handed abuse filters, which are completely unsuitable for a full-scale deployment of this feedback tool due to their broad scope and high number of false positives. If you're looking for libel in feedback, browse through the abuse log for a feedback filter or two. It's a bit bizarre (and I'd say unfair) to suggest that this tool doesn't collect libel and vandalism and other noise simply because others have made a concerted effort to sweep such garbage under the rug. --MZMcBride (talk) 20:18, 22 January 2013 (UTC)
 * When I see a lot of feedback that says "X needs an image", my thought is that this is completely useless feedback. It's trivial to generate a report of articles without images. I'd rather see us engaging readers by saying "you can upload a photo here" or "help us get Y institution to donate an image". It isn't really about understanding NFCC or similar policies, in my view. It's about finding ways to actually engage readers rather than superficially engaging them with a comments box. --MZMcBride (talk) 19:58, 22 January 2013 (UTC)
 * That's nice in theory, but that hypothetical report isn't why High school diploma now has images. The only reason that article has any images is because of feedback. WhatamIdoing (talk) 20:02, 22 January 2013 (UTC)
 * No, that's simply wrong. High school diploma has images because you added the images in these edits. You were able to add these images to the article because WolfgangMichel and Nevit uploaded these images to Wikimedia Commons. I'd love to see more images in Commons and more images in our articles. This has almost nothing to do with article feedback, though. Research shows that drive-by edits contribute substantially to the quality of our articles; drive-by comments do not. --MZMcBride (talk) 20:11, 22 January 2013 (UTC)
 * And why did I make those two edits on 08 September? I made those two edits solely because on 02 September and 05 September, two different readers left feedback that said the article needed pictures of diplomas.  Until then, I didn't realize that there were no pictures there.  WhatamIdoing (talk) 20:14, 22 January 2013 (UTC)
 * I'm not sure how to politely respond to your suggestion here that you were unable to see that the article didn't have images until it was pointed out to you. This anecdote doesn't seem to speak highly of you. :-/ --MZMcBride (talk) 20:20, 22 January 2013 (UTC)
 * That's not exactly fair, Mz. She's saying that the feedback brought to her attention that readers felt the article was lacking images. Without the feedback, it's possible that it wouldn't have occurred to her or anyone else that from a reader perspective, it needed images. That doesn't mean the people who didn't notice before then were somehow incompetent, it just means they weren't psychic enough to know what readers felt was lacking - and probably that they weren't watching that article in the first place (but instead got the idea to add the photos from scanning all feedback, one item of which happened to mention this article). A fluffernutter is a sandwich! (talk) 20:28, 22 January 2013 (UTC)
 * Fluffernutter is right. That article gets an average of 25 changes per year (despite averaging more than 600 readers per day), and until I started scanning through feedback, I only remembered that it even existed when it was on my watchlist.  I process any changes by diffs and frequently have no reason to look any lower than the diff, so I simply hadn't ever noticed that there were no images in the article.  I've got about two thousand pages on my watchlist, and I've actually read very few of them. The comments about not knowing what interests readers are also accurate.  I don't care much about images, but our readers apparently do.  I fairly often edit medicine-related articles, but the importance of the ==Prognosis== section to our readers hadn't occurred to me (I'd have thought that either ==Signs and symptoms== or ==Treatment== would be most important).  I'm not the target audience.  None of the people who have commented on, or even read, this page are the target audience.  The feedback we get helps us learn about our readers in ways that we otherwise couldn't do.  WhatamIdoing (talk) 21:12, 22 January 2013 (UTC)
 * This hits on my #1 problem with the tool (and the "five star" tool). I should say that I actually feel like this tool works better than the "five star" ratings system, but, from what I've seen, there is a large amount of useless feedback.  I have occasionally seen (and responded to) actionable feedback the tool has provided, but, as is pointed out above, I've found that, on most bio articles, a rating of "50% found what they were looking for" actually means "50% were looking for gossip about who the subject has slept with" or "50% were looking for copyrighted pictures" or some such nonsense.  I'm not advocating to get rid of the tool altogether, but I don't understand why we need to aggregate the number of positives and negatives at the TOP of a feedback page to serve as a badge of shame from people wholly unqualified to be "rating" it.  I don't have a problem with the reader providing "positive" or "negative" feedback, but posting a meaningless statistic at the top of the feedback page creates the exact same problem I had/have with the "five star" system, which is to basically give the average reader unfamiliar with Wikipedia's policies regarding BLP's or copyrighted images, etc, a misleading impression that there is something specious about a the entire article as a whole (although I admit it's now at least better hidden on the "Feedback" page instead of posted on the article itself).  Basically, my view is that we should replace the "five star" system with this tool, but ditch the pointless "X% found what they were looking for" headline at the tops of Feedback pages.  Just one example I can give is the Drew Barrymore article which has a GA rating from a presumably qualified assessor (and did the entire time the tool was being implemented on it), but boasts a whopping 50% rating from readers.  I see the tool has been disabled on the page now, but it was being used on the page for a fair amount of time and gives a perfect example of the caliber of feedback being aggregated in the "five star" boxes and at the tops of feedback pages: Feedback on Drew Barrymore --- Crakkerjakk (talk) 23:16, 22 January 2013 (UTC)

Question about articles you watch
Some of the comments from opponents deal with unencyclopedic feedback that I haven't encountered, like requests for prices or phone numbers. There may be a difference based on subject area. For example, an editor working largely on articles about history might get a higher proportion of useful, or at least not completely worthless, feedback than an editor working largely on articles about recent films (who probably sees a lot more feedback about where to purchase the film on DVD). Would you like to share the subject area(s) that you usually follow, and perhaps your opinion of whether that subject area is better or worse suited for reader feedback? WhatamIdoing (talk) 21:21, 22 January 2013 (UTC)


 * I edit mostly medicine-related or education-related articles. These seem to get fewer unencyclopedic comments in feedback than other users report above.  WhatamIdoing (talk) 21:21, 22 January 2013 (UTC)
 * My watchlist is extremely eclectic (checking the feedback today for my watchlist, I see a children's book, some classic poetry, an internet organization, a feral child, a game show contestant, a few general concepts, some help/project pages, a Disney ride...etc etc), but if I had to generalize I'd say that of the feedback that comes across my desk as far as articles I edit, my reactions range from "...huh?" to "Thanks for saying you're happy/unhappy with us/the article" to "I don't know what language this is, but it's sure not English" to "Ok, thanks for the thought, but this isn't really actionable feedback" to "This is not a Q&A venue and we don't get paid enough to do your homework" - with those types comprising about 90% of it - and the occasional "Huh, I could actually work with this suggestion!" (say, 3%) and "This needs to be hidden/oversighted" (say, 7%). I didn't count them up or anything, but that's my general sense on looking over the latest. So probably 80%-ish of feedback where the submitter meant well but for one reason or another their contribution is useless, 10% where either I or they don't know wtf is going on, a couple percent actionable, and the rest deliberate abuse/libel/someone sharing their personal details. A fluffernutter is a sandwich! (talk) 21:41, 22 January 2013 (UTC)
 * I edit mainly U.S. Supreme Court and Law articles, usually case law. I also tend to edit Native American articles, particularly where they intersect with legal articles (for example, most of the SCOTUS articles I edit are tribal cases).  I also do a little in MilHist (not much).  My watchlist is all over the place though.  In both law and Indian articles, the chaff level is high.   GregJackP   Boomer!   00:26, 23 January 2013 (UTC)
 * I too have an eclectic watchlist. From my watch list there are 10 items of feedback; 2 focus on one BLP one asking for her D.O.B (not given in sources) and another asking for an infobox (not really warranted given how little information we have on her); 6 focus on another BLP 4 asking for more detail than we have sources to provide, 1 praising his latest work, and 1 praising the article; 1 piece of feedback asks for pictures of a historic demolished building (there are none free) the final piece praises our coverage of a historic event. None of this is really useful in helping improve these articles. Stuart.Jamieson (talk) 13:40, 20 February 2013 (UTC)

View from Tom Morris
Wikipedia is at a difficult stage. It is now reportedly the fifth most popular website on the planet. It contains mistakes, omissions and other things which make it suck for readers. And it's openly editable.

Say a reader wants to tell us something is wrong. How do they go about doing that? They can edit the article, but that's scary because the crazy Wikipedians that the press has told them all about will shout at them. Okay, well, then maybe the talk page. You can dump stuff on the talk page... and probably nobody will read it. Or you can post something in article feedback... and nobody will read it. We're complaining because we have a system that's made it easier for non-editors to tell us there is actually stuff that's wrong in Wikipedia (and, yes, we like to pretend there aren't distinctions between readers and editors. That's wishful thinking. In theory, every reader is also an editor unless they are banned, but in practice, there are people who actually routinely press 'edit' and there are others who don't).

We need to make a way for the sort of people who don't know what the WP:GNG is or what goes on at WP:ANI or WP:ARBCOM or whatever to tell us about problems with articles. They want to do so. Talk pages are pretty unfriendly, article feedback isn't. Lots of people send us emails to OTRS. Ideally, I'd like to see a situation where non-sensitive feedback goes into one place. It's sort of like a bug tracker. Someone can dump feedback in, we chuck out the stuff that sucks, then we get the hell on and process the problems. We can moan about too much lousy feedback, and the presence of potential problematic content, but that's obscuring the main problem. There are problems in our articles, and users want to tell us they exist. And we'd rather put our fingers in our ears and say "yeah, yeah, too busy".

What we need to do is actually find a way to triage feedback in a unified way, and to match up the feedback that needs handling with the people best suited to do it. Part of that is some way of efficiently fanning out the queue into manageable chunks, marking feedback as patrolled, and elevating feedback that actually needs action on to the talk page. On top of Wikipedia, we have a really crappy issue tracker. What we need is a really good issue tracker. When someone tells us something sucks, we can work out whether their comment is something we can take any action on: if not, we throw it out. If yes, we make sure we don't lose their feedback, and we get on and do it.

Yes, we'll have big backlogs. I'd rather have big backlogs of things that we know we need to do than four million articles filled with stuff that is also potentially a BLP violation or vandalism or much else besides. We've been pissing around at the edges, but if we care about editor engagement, we should (a) make sure no piece of feedback goes to waste and (b) build whatever technology we need to make it so we can start burning through the backlogs and improve the encyclopedia in large-scale ways.

Shutting down the article feedback tool rather than improving it is a bad strategy. We do need better tools for churning through AFT5 responses and patrolling them. We need something like Huggle or STiki to do basic triage on the feedback we get, to remove libel and the "OMG I LOVE JUSTIN BIEBER" type things. The rest, though, those are telling us about potentially fixable issues with Wikipedia. If a reader, in good faith, wishes to give us feedback about an article, we should listen. We might set the feedback to one side because we aren't the sort of editors who can necessarily do anything about it. But if we stop listening to readers who have information that can improve the article, what's the damn point? —Tom Morris (talk) 23:35, 22 January 2013 (UTC)


 * Users who endorse this view
 * Me, obviously. —Tom Morris (talk) 23:35, 22 January 2013 (UTC)
 * 1) This is one of the best statements I've seen in this discussion. Abyssal (talk) 19:30, 24 January 2013 (UTC)
 * 2) I want to see the comment feature live on. Olli Niemitalo (talk) 00:59, 26 January 2013 (UTC)
 * 3) Avenue (talk) 15:08, 26 January 2013 (UTC)]
 * 4) seems reasonable -- Guerillero  &#124;  My Talk  21:24, 27 January 2013 (UTC)
 * 5) Some feedback will be ignored and some won't. This is the way Wikipedia has always been - talk page comments have sat ignored for years. We should try to collect as much feedback as we can. Dcoetzee 01:24, 28 January 2013 (UTC)
 * 6) Im not sure central triaging is the way to go, otherwise well said. FeydHuxtable (talk) 18:32, 30 January 2013 (UTC)
 * 7) pretty much. -- phoebe / (talk to me) 01:56, 6 February 2013 (UTC)
 * 8) Nicely put. – SJ<font style="color:#f90;"> +  23:28, 9 February 2013 (UTC)
 * 9) the wub  "?!"  12:07, 13 February 2013 (UTC)
 * 10) what! listen to readers !? <font face="Vijaya">Farmbrough's revenge ⇔ †@1₭ 15:17, 13 February 2013 (UTC)

There's some brainstorming at Kvetch about a better system for receiving article complaints. It used to be that even finding an e-mail address was a deliberate maze (Contact us). It seems it's not as bad as it used to be, which is nice.
 * Comments

I sympathize with your view, though I wonder how you or I got started here and why it can't apply to others. Yes, telling someone you found an issue is a good thing to do, but we really need to encourage people to get involved themselves. A sane system might take the "needs an image" feedback and suggest that the user head on over to Commons. A sane system might do a lot of things differently (such as using the talk page or better integrating with user watchlists). But we're discussing what we've got and what the Wikimedia Foundation would like to deploy to all articles at the end of March.

And, bigger picture, I'd say that there was never any shortage of work to do around here. (We can always build lists of articles containing typos or articles missing images or articles that "need more info.") All priorities not being equal, finite developer resources are better spent not developing comments sections and finite volunteer resources are better spent not sifting through sand, in my opinion. I've asked the community to weigh in as well, as they're the ones who will be doing this sifting (or not) if this gets deployed more widely. Look at the comments on a site like YouTube. Look at the comments being filtered at Special:AbuseLog (or the unfiltered comments at Special:ArticleFeedbackv5). Then tell me who you've got lined up to use any tools (and time!) on this garbage. --MZMcBride (talk) 01:40, 23 January 2013 (UTC)

View from Mailer Diablo

 * AFT5 should not become a full substitute for editing the talk page which it seems to have become for the causal user. The talk page has long been the wiki way for editors.
 * AFT5 should be redesigned to integrate with the talk page where editors can watch it for new discussions and address them accordingly. It should be a starting point for visitors to directly interact with the wiki editing interface itself.
 * AFT5 should act as a form of introduction to encourage the visitor to make the first edit (if it's an obvious error, WP:BOLD) and how the wiki works, not as a form of additional workload to existing editors that we are sorely lacking to patrol.
 * AFT5 is most useful for obscure articles where it sorely needs improvement and feedback. Certain keywords can be programmed to give direct assistance to the user (such as translation, Q&A is this way, etc.) Most feedback for major articles do not have a tendency to be useful nor helpful. - Mailer Diablo 01:16, 23 January 2013 (UTC)


 * Users who endorse this view
 * 1) I'm Mailer Diablo and I approve this message. 01:16, 23 January 2013 (UTC)
 * 2) I agree strongly with the part about it being more useful on low-readership articles. For high readership articles such as World War II it is counterproductive.  Abductive  (reasoning) 09:04, 23 January 2013 (UTC)
 * 3) Mailer Diablo's second proposal is especially important. This tool pretty urgently needs to enable more effective communication with the feedback leaver and other community members that may be interested in the points they raise. We should be able to respond directly on the user talk page of the feedback leader or start a relevant talk page discussion directly from the template. Abyssal (talk) 19:28, 24 January 2013 (UTC)
 * 4) I support this, if there is sufficient, clearly communicated scoping, design, and iteration put into this. These are major changes and should not be undertaken unless there is a will to follow through, make necessary course corrections, etc. -Pete (talk) 19:16, 1 February 2013 (UTC)


 * Comments
 * I would tend to disagree - most of the article I edit are on the obscure side, and the feedback isn't any better there. GregJackP   Boomer!   01:35, 23 January 2013 (UTC)
 * I guess you mean "AFTv6", not "AFT5" in your view. ;-) I'd like to see a smarter, saner article feedback tool, too. But I'm not sure it's a very high priority. As I said to Tom above, I don't think there's any shortage of work to do around here without reader comments. And there is a shortage of volunteer and developer resources. --MZMcBride (talk) 01:43, 23 January 2013 (UTC)
 * Feedback can not be left on a talk page using the mobile interface and in a few years(if not already) most traffic will be via that interface. Regards, Sun Creator(talk) 02:46, 23 January 2013 (UTC)
 * Couldn't feedback be transferred to the talk page by a bot? Perhaps when it is marked as helpful but left unresolved? Abductive  (reasoning) 09:04, 23 January 2013 (UTC)
 * Or perhaps if it's marked 'featured'. As it's just a copy and paste operation, this should be relatively simple to do.  WhatamIdoing (talk) 22:08, 23 January 2013 (UTC)


 * You want "visitors to directly interact with the wiki editing interface itself". Are you aware that the entire interface is going to change when the Visual Editor goes online?  WhatamIdoing (talk) 22:08, 23 January 2013 (UTC)
 * Speaking of which, it's been enabled. See https://blog.wikimedia.org/2012/12/12/try-out-the-alpha-version-of-the-visualeditor/ to get information about trying out the VisualEditor.  WhatamIdoing (talk) 17:26, 29 January 2013 (UTC)
 * Speaking of which, the Visual Editor isn't applied on talk pages? So we're making it easier to edit the articles, but if a reader wants to write on the talk page, he still has to deal with strange WikiMarkup? Hm. --Atlasowa (talk) 16:22, 1 February 2013 (UTC)
 * No, the VisualEditor will be enabled for talkpages - we've just not enabled it for that namespace yet. Okeyes (WMF) (talk) 16:31, 1 February 2013 (UTC)
 * Actually, there are no plans for that, since talk pages are super complex next to article pages. We're looking to Flow to solve for the discussion problem.--Jorm (WMF) (talk) 18:57, 1 February 2013 (UTC)
 * Flow? -Pete (talk) 19:16, 1 February 2013 (UTC)
 * Flow. --MZMcBride (talk) 19:24, 1 February 2013 (UTC)
 * I do agree that AFT should recognise certain types of comments and direct readers to further information (e.g. reference desk, info on information that can't be included, etc.), to the extent this is possible to do automatically. Dcoetzee 01:27, 28 January 2013 (UTC)

A comment on requests for pictures
I see a lot of complaints above about the fact that many readers submit requests for pictures, which many editors feel to be unhelpful. It might help if the feedback box did not give "This article needs a picture" as its example of the kind of feedback we would like to receive.--Srleffler (talk) 04:35, 23 January 2013 (UTC)
 * LOL Axl  ¤  [Talk]  11:07, 23 January 2013 (UTC)
 * @Srleffler: See Wikipedia talk:Article Feedback Tool/Version_5. Are you going to be satisfied with this Post without comment button? -11:16, 23 January 2013 (UTC), Utar (talk)
 * I agree with Srleffler. I think the current one might be the worst example you can provide for readers. We don't need a feedback from the readers to know if an article needs more images or not. A more specific suggestion would bring better productivity. Please request ideas from the community, if necessary. ···V ani s che nu「m/Talk」 17:10, 23 January 2013 (UTC)
 * Utar, I don't see how that question relates to my comment above. Anyway, the answer is "No. Why would I be?"--Srleffler (talk) 05:12, 24 January 2013 (UTC)
 * As I said in the discussion linked above, in the german AFT5 test we use a different feedback example than "This article needs a picture". I think the german feedback example (In the article the history since 1989 is missing, compare E. Mustermann (2007).) is more aspiring, maybe even too ambitious. I have seen this same phrase submitted as feedback maybe once or twice. But i read requests for pictures in german wikipedia as well (maybe our feedback is poisened by the en-wiki example ;-P), though often more specific, like "show a whale skeleton". On de-wiki only ~45% of feedback is with comment (with text), on en-wiki ~75% is with comment. But there are tons of differences overall: de-wiki feedback volume is tiny by comparison, almost all readers are native speakers, we have selected the articles for feedback (only 0,1% is random) and we also get a lot of useless comments. I think the new "Post without comment" button will be very beneficial. Another point is the rather small size of the text input field: If you are asked for a tweet-length comment, how specific and detailed will your suggestion for improvement be? How much effort do you put into writing complete sentences in a tweet? When even the example phrase consists of only 5 words? Is there any research we can rely on, about quality of feedback in relation to/depending on text input field size? (I think there must be, somewhere, and i would really like to know!) So, I think these are 3 things we should consider or implement to improve feedback quality: 1) the new "Post without comment" button, 2) a more aspiring feedback example (or 2 examples?) 3) a bigger text input field (if there is reason to believe that this will lead to better quality feedback, not just to longer useless comments ;-). And A/B testing this would be great. --Atlasowa (talk) 19:23, 24 January 2013 (UTC)
 * If we implement a "post without comment" button, the user's feedback should be automatically hidden from view, and should only affect the statistics. Except for its impact on the stats, a feedback without a comment is worthless. As an editor, I don't care whether some particular user's perception of the article is "happy face" or "sad face". That has no value. The statistics do still have some value—an article with an unusually high ratio of "sad face" is clearly in need of further attention, even if none of the readers can articulate why they are unhappy with it.--Srleffler (talk) 06:17, 25 January 2013 (UTC)
 * Lots of our articles lack images, we don't really need our readers to tell us that or even to get lists of articles that lack images - we can use bots to do that. But much more usefully we can produce lists of articles that lack images but which we probably have images available for on Commons. Take Articles with UK Geocodes but without images for example. This is a list produced by computer of UK articles without images, and because they are in the UK we probably have images if anyone looks (we imported nearly 2 million UK images from the Geograph - a UK and Ireland photography competition. Other useful lists that could be created include Articles without images but with Intrawiki links to articles on other Wikipedias that have images from commons. This is a much more productive use of volunteer time than wading through the AFT responses.  Ϣere Spiel  Chequers 23:15, 23 January 2013 (UTC)
 * Actually I'm not convinced that wading through bot-generated lists of articles is more attractive for editors. But i do agree with what you propose. We need a tool that can show/propose suitable commons-images for articles, 1) by showing commons-images that are you used in the interwiki-linked articles on other Wikipedias (for all articles, with or without images), 2) showing geotagged commons-images near the coordinates of the article, 3) showing other commons-images of the same commons-category as the image that is in the article or of the article-linked commons-category. Maybe add 4) FIST etc. This tool, applicable to all articles by a click, would help to actually find suitable images. I think: reader feedback requesting images is a helpful hint and push, if combined with a tool for editors to easily find suitable images for an article, then this can really improve articles. Mere to-do-lists by bots don't.
 * BTW, there's another frequent comment/request, "too many difficult words", "make this text easier to understand", which is a reasonable feedback, but not really helpful for editors. If we had a tool similar to a spell checker, that highlights uncommon words (and proposes wikilinks for them?) and very long sentences (proposes to split them?) etc. in the article text - that would make it easier for editors to actually improve article readability. We currently only have tools that give a readability score (Village_pump_%28technical%29/Archive_95, Help_desk/Archives/2010_December_20, Village_pump_(technical)/Archive_106), mere numbers don't help improving the text, suggestions do. And reader feedback gives a hint and a push to do it. --Atlasowa (talk) 20:55, 24 January 2013 (UTC)

Maybe the tool could be tweaked to treat the example feedback as a variable that can be replaced with randomly selected feedback marked as featured by the community. Abyssal (talk) 19:25, 24 January 2013 (UTC)

This thread needs a picture. Shearonink (talk) 02:59, 26 January 2013 (UTC)


 * I agree with this. I think a much better "default message" would be something like "e.g. the birth date should be 1864, not 1870" or "I found the discussion of gluons confusing" which is so specific that it would be easily recognizable when people just blindly copy it, but also indicates the kind of specific and constructive suggestion we'd like to get. Dcoetzee 01:29, 28 January 2013 (UTC)


 * or the default message "images on wikipedia are complicated: if it's a living person, then david shankbone needs to come round and catch them at a public place. please upload an image where it will be shortly deleted because you didn't use the right license." <font face="Vijaya">Farmbrough's revenge ⇔ †@1₭ 15:24, 13 February 2013 (UTC)


 * I lol'd. --MZMcBride (talk) 07:39, 15 February 2013 (UTC)


 * Maybe the default message should read something like: "If you find any incorrect information in the article, or have any suggestions how to improve it, then please try editing yourself. If you fail to do so, don't have the time, or are just too lazy then leave a message here. Try to phrase your suggestions as specific as possible, and write coherent sentences in proper English, otherwise ýour feedback cannot be processed and will be deleted without further consideration." Perhaps that would get people to actually use whatever language skills they have and yield some better feedback. (Thusz (talk) 17:53, 20 February 2013 (UTC))

Question about financial costs
Not counting the money already spent on developing the tool, will there be additional financial costs involved in rolling it out all over Wikipedia and in maintaining it? If so, how much? The cost/benefit discussion here has focused almost entirely on the potential drain on human resources, i.e. active editors and administrators. It would be good to know the financial side. Voceditenore (talk) 12:52, 23 January 2013 (UTC)


 * I would also like to know, how much is this tool currently costing in terms of bandwidth. How close is it to 1% of the Foundation's bandwidth costs? Jason Quinn (talk) 19:13, 23 January 2013 (UTC)
 * I, on the other hand, would like to ask the foundation to declare how much time and money they have invested in this tool. Money solicited from readers with the annual doomsday campaigns. --Bensin (talk) 18:04, 27 January 2013 (UTC)
 * Warning: speculation follows. Years ago I remember reading some articles in the press (I think on Yahoo, so probably through AP or Reuters) about a large donation to the Wikipedia Foundation (perhaps the largest at the time) and that the money would be used to help make Wikipedia easier to edit for new users, and so on. It may have been this Sloan Foundation donation although the actual one I am vaguely recalling may have even occurred prior to 2009. At some point, I started to interpret some "top-down" projects that seem to come through as WMF fiat as attempts to fulfill the contingents of donations. It really helps to have bullet points on slides showing what is being done with donated money. Donation contingencies are something of a two-way street. The WMF likely commits itself to certain "improvements" during its pitch to score large donations, while conversely, large donators can broadly earmark what they'd like the money to be used for. The old saying "follow the money" would hint that some project spear-headed by the WMF are related to large donations. The press release above says the money will be used for "increasing quality" and "broadening participation". It is quite easy to imagine a WMF board room meeting where it is asked how that could be achieved and something like the AFT results. And on paper, the AFT sounds great to most people and so a team is gathered and put to work on it. Once development starts, it's not easily stopped even with warning signs ahead because that would be viewed as waste of donation money. This is my best guess as to the origin of the AFT. This is an outsider's total speculation but it seems consistent with things so maybe it's true. Jason Quinn (talk) 19:27, 28 January 2013 (UTC)
 * If this is true, then the WMF's biggest mistake was to not trust that the community could come up with something better than AFT regarding both article assessment and recruiting new users. I would really like a comment from WMF on Jason Quinn's post above regarding any stipulations associated with donations. For this and any other donation. --Bensin (talk) 23:30, 28 January 2013 (UTC)
 * Jason, if we had large amounts of money from big donors recently, you'd have heard about it :). The Sloan Foundation donation was used in usability testing and initiatives, not for AFT5. Crucial phrase is 'years ago'; if we'd kept the money idling since ~2009 in a bank account somewhere we'd probably not have been fulfilling the donations. As a general practise we deliberately avoid donations with riders, and have for years. Okeyes (WMF) (talk) 23:48, 28 January 2013 (UTC)
 * Is AFT in any way connected to a specific donation? Are donations with conditions available for review? Are there projects, current or in the past, funded by donations with conditions? If so, which are they? --Bensin (talk) 14:20, 29 January 2013 (UTC)
 * Not to my knowledge, possibly, but I'd imagine it depends on what the donor is comfortable with, and I do not know, in order. Would you like me to try and find someone who can answer them? Okeyes (WMF) (talk) 15:56, 29 January 2013 (UTC)
 * Yes, please. Thank you, Okeyes! --Bensin (talk) 16:35, 29 January 2013 (UTC)
 * Okay, so I've spoken to Lisa (Grunwell; she's our Major Gifts person). Her statement, to paraphrase, is that the WMF deliberately avoids any gifts that involve riders of 'you must spend the money on X' unless we're already interested in X and planning to do it anyway - so, for example, we got a grant for the Visual Editor, which we've been working on for literally years: we accepted the rider of 'it must be spent on the VE' because, well, that's what we wanted to use it for :). Major gifts are announced via the Wikimedia blog; the current projects that are subject to 'restricted' gifts are the Visual Editor, as mentioned, and the Wikipedia Zero program to build infrastructure that lets people access Wikipedia on their phones free of charge. Okeyes (WMF) (talk) 22:54, 29 January 2013 (UTC)
 * I was going to mention the VE during my OP as other potential candidate for my idea. (I had also never heard of the Wikipedia Zero program before. Interesting.) As for the AFT, do you know who originally proposed it and when? And who gave and when did it get the green light? Jason Quinn (talk) 17:24, 30 January 2013 (UTC)
 * Version 5, or AFT generally? Okeyes (WMF) (talk) 17:40, 30 January 2013 (UTC)

I'd actually like to see this sequence: continued further in the past. That way it'd be easy to find such things. If pages already exist somewhere that cover the older years, it'd be good to have all that material organization in single, logical, and clean fashion instead of spread out. Would probably only take somebody a day or two to make from previous annual reports. Jason Quinn (talk) 17:10, 29 January 2013 (UTC)
 * Benefactors (mid-2011 to current)
 * Benefactors/2010-2011
 * Benefactors/2009-2010
 * Benefactors/2008-2009
 * Are you volunteering to work on this? :D
 * If so, please request an account at WMFACCOUNT and I'd be happy to create an account for you. --MZMcBride (talk) 19:36, 29 January 2013 (UTC)


 * I would not expect the bandwidth of the tool to be remotely significant even if rolled out to all articles. It seems like it will require more database resources, since the experience is no longer "read only" and fully cacheable, but I don't know how much this impacts hardware costs. Dcoetzee 01:31, 28 January 2013 (UTC)


 * Hi Voceditenore, thanks for your question about additional financial costs for Article Feedback, going forward. We currently have only one developer working on AFT5, with some additional ops/code review/product support. That developer has been dividing his time between Article Feedback and database tools for both AFT and Flow (our upcoming messaging product). We're prepared to continue to resource it at this level, if it helps make the tool more useful to the English Wikipedia community. We also plan to provide additional support for the release of AFT on other projects, besides the English Wikipedia. Overall, this is a modest investment compared to some of the larger projects under way at the foundation. Dcoetzee, you are correct that the bandwidth costs are likely to be minimal, and the hardware costs are not expected to be significant either. Bensin, the foundation does not break down development expenses for each product we work on, but you can learn more about our editor engagement program as a whole in last year's annual report. Fabrice Florin (WMF) (talk) 22:09, 29 January 2013 (UTC)
 * A qualified estimation will suffice as far as I'm concerned. Time and money. --Bensin (talk) 22:51, 29 January 2013 (UTC)

Technical problems with the tool
I use the tool every few days, but I've had some technical problems with it. I generally run it by clicking "Feedback from my watched pages" from my watchlist. I'm using Firefox 14.0.1 on a Windows 7 PC. Mitch Ames (talk) 13:28, 23 January 2013 (UTC)
 * It doesn't always sort correctly. I raised the matter at the village pump; it was acknowledged, but has never been resolved.
 * For the last month of so, the tool hasn't run reliably at all. Most times when I click the link, nothing happens. The only way I can reliably get it to work is "open link in a new window".
 * Have you reported these to bugzilla yet? If not, developers most likely won't see it on the village pump. Legoktm (talk) 17:38, 23 January 2013 (UTC)


 * For months the feedback (input) tool was not working for me, on WP:en, while working fine eles where, anmely on WP:pt. I reported twice on the village pump, one user replied. I disable it so not to get a bunch of garbled text at the bottom of the articles. Some how it seems to be working again. Good. Two points out of this: one, the more complicated you make the pages, the more errors there will exist, so whatever feedback there will be, please, please, keep it simple (on the interface, add filters, add sorting, add a hyundred things, but do not rely on code that works only on the latestest Firefox browser, or whatever). Second, if the developers can not make a decent feedback for themselves (bugzilla? open an account? watch out, we do NOT take care of privacy! yeah... looks like a friendly feedback tool...) nor they care to take a look at the village pump, what exactly do they understand about "feedback", to create one? - Nabla (talk) 22:15, 28 January 2013 (UTC)
 * Correction: It still does not work. When logged *out* something goes wrong and I only get the text (version 4, or 3, or whatever is the previous one), when logged *in* it seems to be fine. It probably happens that I was mostly logged out until not long ago, and mostly logged in currently. I understand code is never finished (i do a few lines of code myself) but it is unfortunate to have faulty code that much clearly visible. - Nabla (talk) 02:05, 1 February 2013 (UTC)

Question about degraded response time
Whilst most commercial operators seem to focus their projects on the wealthy wired populations of people with good broadband connections, we have an obligation to the whole world, including those who are in areas and with technologies that involve slow Internet access. How much does this degrade the response times for the 5% of our readers with the slowest internet access?  Ϣere Spiel  Chequers 20:06, 25 January 2013 (UTC)
 * How would you propose we work out which 5 percent that is, and what their existing response time is? Okeyes (WMF) (talk) 20:23, 25 January 2013 (UTC)
 * If you're talking about communication bandwidth consumed by additional html for the present tool, it's certainly less than 1% of the total for the page -- and that's excluding images! EEng (talk) 17:55, 27 January 2013 (UTC)
 * If it is less than a 1% overhead then if one were taking a purely commercial view of our readership one might argue that a slowdown of less than one percent is an acceptable price. But we have a remit to get our information out to the world, if cancelling AFT would speed up access even by a smidgen then we should take that opportunity. Of course to answer Okeyes point, there are bound to be maps out there of low and hi speed connection areas, and IPs that are known to run slow copper of fast fibre optic networks. You could address this particular objection to AFT by identifying slow IP ranges and disabling AFT from them.  Ϣere Spiel  Chequers 18:52, 28 January 2013 (UTC)
 * A better question: how many of those 5% are currently savvy enough to give feedback via the talk page, so that their unique knowledge and background regarding topics can be understood and represented? It seems like it would only help if they had a way that they could give us feedback easily. Dcoetzee 01:38, 28 January 2013 (UTC)
 * But is our priority to build the pedia or to make it available?  Ϣere Spiel  Chequers 18:52, 28 January 2013 (UTC)
 * When you're talking about a difference in milliseconds of response time, building is definitely more important. Frankly, misuse of templates is a bigger performance problem than this extension. Dcoetzee 06:32, 29 January 2013 (UTC)
 * If the problem was milliseconds then that would be reasonable, and I don't dispute that template misuse is a bigger problem - but we should deal with both. But for people with slow connection speeds an increase of possibly up to 1% could mean much more than a millisecond. No-one disputes that for Internet users with what would now be considered normal speeds for the developed world this overhead should be miniscule - the concern is for those who have the slowest connections.  Ϣere  Spiel  Chequers 11:32, 29 January 2013 (UTC)
 * Frankly, I think a 1% difference in response time is not realistic. I expect any performance difference to be much smaller than that, particularly for readers who elect not to use the tool (serving up a few extra kilobytes of Javascript is not going to kill even a modem user's connection). Dcoetzee 21:18, 30 January 2013 (UTC)
 * I absolutely agree. I was trying to give an extremely simply metric bounding the degradation at what I thought would be seen as a trivial number, but apparently I guessed wrong about how others would view that. EEng (talk) 01:13, 1 February 2013 (UTC)

View by Voceditenore
This tool may possibly (depending on the topic) generate some useful suggestions for article improvement that we wouldn't have had otherwise, but I seriously question the assumption that this data tells us anything significant about "our readers" or their "perspective". The ones who submit feedback are probably not representative of our readership at all. The respondents so far may largely be a tiny self-selected group of people who enjoy clicking buttons and/or the "like/dislike" feature of Facebook and/or making comments on the internet (for good or ill).

I mostly watch and write opera-related articles. I chose William Tell Overture to analyze because it's not as specialized a topic as some of the others in the area, and many people who don't know about opera know about this piece via commercials and the Lone Ranger. In the last 6 months, the article has had approximately 54,000 page views. It's infrequently edited, and largely by me, so the vast majority of the page views are from "readers". It's had 20 responses at Feedback on William Tell Overture of which 1 provided potentially useful feedback. Of the remainder, several clearly had trouble with English, several had clearly not read the article and probably went straight to the shiny feedback buttons at the bottom of the page, one was truly silly ("this suckd"), and one was charming but irrelevant for improving the article. I would draw no conclusions whatsoever about this article's actual readership or their "perspective" from those 20 responses. The responses tell us only about the kinds of people who fill in these forms. Voceditenore (talk) 18:31, 23 January 2013 (UTC)
 * I also looked to see if the respondents engaged anywhere else on Wikipedia after leaving feedback. None of them did, at least not under their IP address. One of the IPs had previously vandalized an article. Another one (although not necessarily the same person who left the feedback) had previously made some unconstructive edits to the article on their school, and er... that's it. Voceditenore (talk) 09:34, 24 January 2013 (UTC)


 * Users who endorse this view
 * 1) The proportion of readers who like this enough to use it is high enough to cause problems, but so low that it has clearly failed as a readership engagement tool.  Ϣere Spiel  Chequers 09:19, 24 January 2013 (UTC)


 * Comments

View by Zaereth
I saw this RfC, so I decided to drop off my $0.02. Forgive me if I seem rather blunt, because I appreciate the hard work that people put into designing this feature. From what I can tell this feature was designed for two purposes: To improve articles and to get potential new-users involved, so I'm goig to try to examine this from those two perspectives.

On the point of improving articles, the feature can provide useful comments. However, those comments are often unclear as to the specific changes that need to be made, and there is no way to ask for clarification. In this respect, a simple, in-line, citation-needed tag is often much more useful. However, if the comment was to be directed to the talk page, a useful discussion could be implemented. For me, talk page discussions are very useful, not only because they help me to see exactly what others are seeking, but also because responding helps me to organize my own thoughts on the matter. For this reason, I generally ignore the feedback tool, because it doesn't involve me in a useful discussion.

Most of the feedback I've seen are usually very rude, crude, childish responses anyway, and I have no time to weed through them all. In a way, I think it may provide a useful avenue where vandals can go to vent their frustrations, particularly knowing that they won't be getting a profound response of some kind. In this respect, a comparison can be made to the "comments" section of nearly any website. All in all, though, it's another reason for me to ignore it. I have yet to see any useful changes to an article result from the feedback, (but perhaps my watchlist is just too limited).

On the point of user involvement, because there can be no actual discussion, I don't see how we can expect to improve involvement. For me, I have to look at that from two aspects: What kind of user do we want to get involved (who), and what are their motivations for getting involved (why)?

I can only speak for myself. For me, there are really two reasons for joining: I genuinely want to improve the reader's understanding of the subjects I know about, and I also want to engage with people who share the same interests and background knowledge that I wish to share. On the latter point. as a new user, if I had simply left a comment on the feedback tool, I would be very disappointed to see no response --to have no opportunity to converse with someone-- and I would have never looked back. As it happened, I saw a subject which I thought I could help with, and left a comment on the talk page. A highly knowledgeable user gave me some friendly, encouraging advice, and that is why I decided to join. Had it not been for that one user, I never would have bothered. It was the opportunity to have the discussion that gave me the motivation.

On the former point. The satisfaction of adding to the encyclopedia comes from knowing that my contributions actually helped people. With all of the vandalistic, snide remarks that appear on the new feedback tool, it is really difficult to find the good remarks, and impossible to engage with that person. On the other hand, the old tool gave a rather nice way to simply look at ratings. When I provide a definite improvement to an article, I can literally watch the ratings go up, and that helps me to get the satisfaction needed to keep contributing.

To sum it up, I think the tool would work better if: 1.) Feedback comments are directed to the talk page, and 2.) the old rating tool is also incorporated into the new tool.

That is my humble opinion. Thankyou. Zaereth (talk) 20:06, 23 January 2013 (UTC)


 * Users who endorse this view
 * 1) Very nicely put, apart from the summing up. I'd sum it up differently - that a more prominent edit box would be better and I'm loathe to see any aspect of this tool retained other than on those articles where someone wants feedback and opts in to the tool.  Ϣere  Spiel  Chequers 09:42, 24 January 2013 (UTC)
 * 2) i support the idea that any coversations and suggestions of feedback be incorporated into a single place that already exists for commentary / feedback and discussion: the talk page. If we want to bring editors into the project, actually bringing them into the project and not segregating them is the best way. --  TRPoD aka The Red Pen of Doom  19:26, 24 January 2013 (UTC)
 * 3) the feedback tool is basically a one-way channel. Even if you act on the suggestions, you can't leave a note to encourage further edits, and asking them to edit directly the next time. I don't get the feeling that it brings new editors. --Enric Naval (talk) 12:43, 6 February 2013 (UTC)
 * 4) This is the same point I was making; namely that as a feedbacker it seems as if your feedback just goes off into space and has zero effect. That seems very demotivating to me for a feedbacker and not something that would encourage a feedbacker to become a Wikipedia editor. I also agree with the watchlist-feedback issue; I barely spend enough time checking my watchlist, and I really don't want to look at another "watchlist", but at the same time I can't imagine who would ever read feedback on articles I created besides me. Jane (talk) 14:08, 7 February 2013 (UTC)
 * 5) Support. Instead of trying to fix the feedback tool, try to improve the Talk pages. For example with automatic signing, automatic separation of replies, decent sorting (reverse-chronological for topics, chronological/threaded for replies within a topic). Show the first N lines of each topic and collapse/expand bar for the rest. And please delete all those useless WkiProject templates. --Jorge Stolfi (talk) 05:50, 15 February 2013 (UTC)
 * Comments


 * What does "involvement" mean to you?
 * I heard recently that the English Wikipedia has five or six editors each day who reach the 1,000 edits milestone, and that most of those people have had no discussions or contact with other editors. To your way of thinking, do those people count as being "involved" in Wikipedia?  WhatamIdoing (talk) 22:27, 23 January 2013 (UTC)
 * Per the "engage with people who share the same interests and background knowledge that I wish to share" comment. Wikipedia is a very dry atmosphere for discussion, more like a library where no one is allowed to talk to anyone else. None of the talk pages are for discussing the subject, only the article. There is a little whispering that goes on, just like in a library, but the social media aspects of Wikipedia are close to zero. On the other hand, Wikipedia might be a friendlier place if there was more social interaction other than just at the wp:meetups. Apteva (talk) 22:37, 23 January 2013 (UTC)


 * Involvement means to me exactly what it says in the dictionary. I do not go to social media sites because there is very little for me in the way of intellectual stimulation. I do not visit Jimbo's talk page for the very same reason. Coming here to discuss articles, build articles, and improve articles is what attracts me to Wikipedia. That is all that attracts me. Without those aspects, I would not be here. I understand that a hostile, dry, forbidding environment appeals to some, but not to me. I respect your opinion, and now you have mine. Zaereth (talk) 23:53, 23 January 2013 (UTC)


 * My apologies for my rudeness. I think I answered too quickly without giving myself time to think. I going to try to put what I said above into quantitve terms rather than qualitive. This is basically going to rest on the same two points: Energy and entropy.


 * The first is energy. What motivates a person to join? It's interesting that nobody seems to be talking about that. I'm sure we all have our own stories. Why not share them? Perhaps that will help to see the bigger picture. So, in trying to recruit new editors, why not try to pin-down what motivates them?


 * The second is entropy, and this is a bit more complicated. The short version of the question is: How can we make the energy we have produce the same rate of growth? The short answer is: We can't. As wikipedia grows, more and more energy is needed just to support the structure. To construct an article from scratch is easy. To improve an existing article is harder, and to improve a long-standing article is very difficult. This, for the most part, requires a concentrated effort on the part of not only the editors, but also the writers. In other words, it may become necessary to discuss the subject as well as the prose to some extent, if it gives someone a better idea of how to explain it. Providing a somewhat cordial atmosphere cannot decrease the entropy, but it goes a long way toward slowing its increase. Zaereth (talk) 03:07, 25 January 2013 (UTC)


 * You will probably not be surprised to learn that the WMF has studied this question extensively. There are several stages in turning a reader into an editor.  The first stage is to get the person to do something other than read.  If they click and then type something, so much the better.  The feedback tool does that.
 * The next stage is to get the person to make a few edits. The problems here are largely technical:  Can the person figure out how to make a (valid) change and save it?  The interface is hurting us because it's so hard to figure out what things mean.  For example, as a newbie, I thought that Cancer and Category:Cancer were the same thing.  There's a problem with editors yelling at the newbie for making an honest mistake (nobody yelled at me, or I probably wouldn't be here today), but the bigger problem is simply that they can't figure it out, so they don't even save the edit in the first place, or they save the edit and never come back because they have learned that Wikipedia is incredibly complicated.  The VisualEditor is intended to reduce the barriers here.
 * The third stage is when the editor manages more than a few dozen edits. After that, the problem is largely social.  There are, unfortunately, no reliable solutions for problem.  I suspect that it's not a coincidence that the clear majority of people who reach 1000 edits have had no contact with other Wikipedians.  If we were always as kind as the people whom you and I first encountered were, then we might have more experienced editors.  But it seems that we're mostly not, so much of our contact with developing editors is to yell at them for mistakes they made.  WhatamIdoing (talk) 19:23, 4 February 2013 (UTC)

Maybe we should post some quantitative anecdata on the feedback we receive
In light of discussions regarding the quality of the feedback we've been receiving, maybe we should go through our feedback and try to evaluate their quality. Make a subheading here and post your results. It might help to note how many articles you checked and what sorts of articles you have on your watchlist. I've divided my results into a few categories, but you don't necessarily have to use those. My first category was useful feedback that might actually stand a good chance of being directly applied to improve and article. My second category were arguably useful feedback that at least in principle could be used to improve an article, but might not end up being used at the article in question, specifically (for instance request for a specific image that would be nice to have in an article but might not fit in practice for space reasons). My third category was for users who made a good shot at leaving useful feedback, but didn't quite succeed. My fourth category are the utterly useless like oneworders and vandalism. Lastly I noted whenever the feedback filter rejected a submission. Abyssal (talk) 20:06, 24 January 2013 (UTC)
 * It's not that clear to me what these results below are adding to the discussion. If users are just doing it for their own quick check that's understandable, but to refer to these as quantitative results is unhelpful (anecdata I accept). Picking a handful of responses and judging them into arbitrary subjective categories isn't quantitative and gives almost no information about the actual population of feedback received. If people really want to gather data on this, it needs to be on a large stratified random sample of feedback with a good number of reviewing editors—like the research done by the foundation in their Q4 report, but ideally with more than two reviewers. I'm not trying to be nitpicky but these results percentages could be quoted later in the discussion out of context to support any position; i.e. "Abyssal found over 70% of feedback to be useful, AfT is great!" or "Voceditenore found almost 70% of the feedback to be useless, AfT is awful!" <font color=#000000>Jebus989 ✰ 20:28, 25 January 2013 (UTC)
 * Agreed, anecdata is misleading as it is prone to misuse. The Foundation has the resources and skills to do a real quantitative study of the data paying attention to factors like inter-rater correlation, and in fact has already done several quantitative studies of comment quality during development. Dcoetzee 21:16, 30 January 2013 (UTC)


 * If you look at the "View from Guy Macon" section you will see that we are working on evaluating 1000 feedback entries selected with a random number generator. This should give us an adequate statistical sample. I will look into having multiple eyes looking at whether an entry is useless or useful. --Guy Macon (talk) 23:37, 25 January 2013 (UTC)

Abyssal's results
I've gone through the last 30 pieces of feedback I've received and tried to evaluate the utility of each. My articles are mostly on paleontology and modern animal groups. Here are my results:

Useful: 10 (33%)

Arguably useful: 12 (40%)

Good faith but not useful: 3 (10%)

Useless: 4 (13%)

Rejected: 1 (3%)

Overall my results have reinforced my generally positive view of the tool, but I did grade a bit leniently, and I'm a bit more open minded than some users to feedback like image requests, so keep that in mind when you evaluate my report. Still, nearly three fourths of the feedback I've received (my watchlist contains over ten thousand articles) lately could at least in principle be used to implement improvements. Abyssal (talk) 20:06, 24 January 2013 (UTC)

Fluffernutter's results
I looked at the 53 most recent items in the feedback for my watchlist, most of which occurred on project-space pages or 20th-century books and stories. I called "useful" anything that could be used to directly improve the article, "arguably useful" things that might be useful if the person responding to the feedback went off and read the article and did research (things like "talk about theme more" went in this basket - if someone researched the story, they could probably add some info about themes, but the feedback itself didn't help), "good faith, but not useful" the i-like-it comments and such, "useless" anything that not only couldn't improve the article but had nothing to do with improving the article, and "rejected" the ones that needed to be hidden or oversighted. The breakdown is as follows:


 * Useful: 4 (7.5%)
 * Arguably useful: 11 (20.7%)
 * Good faith, but not useful: 13 (24.5%)
 * Useless: 21 (39.6%)
 * Rejected: 4 (7.5%)

This was slightly fewer outright gibberish feedbacks than I expected, and slightly more "if you squint real hard it could sort of be useful" feedbacks. Nevertheless, 71% of the feedback I evaluated was of no use, whether on purpose or not. A fluffernutter is a sandwich! (talk) 17:00, 25 January 2013 (UTC)
 * Right. The abuse filters are quite strong, as I've been trying to point out. In addition to blocking profanity (or anything resembling it), they block out really short posts, posts in all caps, etc. It's completely unclear whether the current abuse filters are sustainable on a full-scale deployment. I personally don't believe so due to their heavy-handed implementation. In some ways, I think even just logging an AbuseFilter hit each time might quickly turn expensive, to say nothing of the feedback that actually manages to get through. --MZMcBride (talk) 06:02, 27 January 2013 (UTC)

Voceditenore's results
I mostly edit opera-related articles. I looked at all the feedback for 3 on my watchlist: Libretto, La traviata, and William Tell Overture. There were 51 responses in all. My criteria were somewhat similar to Fluffernutter's. I called "useful" anything that could be used to directly improve the article; "arguably useful" things that might be useful but are too vague to action without further research or a compliment on an aspect of the article that could be used to improve a similar article; "good faith, but not useful" suggestions that were inappropriate, e.g. adding copyright media, entire lyrics to an opera, how-to guides, etc.; "useless" anything that had nothing to do with improving the article, any response (pro or con) with no commentary (if someone clicks "Did/Didn't find what I was looking for" but doesn't say what it was, it's utterly useless), any comment which shows they clearly didn't or couldn't read the article; "rejected" for serious abuse, nonsense, spam, BLP violations. The breakdown (percentages rounded):


 * Useful: 6%
 * Arguably useful: 12%
 * Good faith, but not useful: 12%
 * Useless: 69%
 * Rejected: 0%

Overall, 18% of the responses were useful or arguably useful. In some of the "Didn't find what I was looking for" responses, I suspect the readers had put some key words into Google, got sent here, but it wasn't what they were really searching for, e.g. "I was looking for some quiet music of Rossini" (the William Tell Overture is not quiet music). Voceditenore (talk) 19:07, 25 January 2013 (UTC)

Georgejdorner's results
My niche is biographies of World War I aviators and articles corollary to them. Late add: I have 1,555 items on my watchlist.

I have received 15 notices of feedback, all from anonymous accounts on ISPs. Not one of them has been useful.

On the two occasions a reader has contacted me via email, I have received useful feedback from identified parties that led to improvement of the articles in question.

Conclusion: The feedback system is worthless.Georgejdorner (talk) 00:28, 27 January 2013 (UTC)

Ynhockey's results
Contrary to most people here (not just the empirical data, but also those posting in the discussion), I have found most feedback to be useful in some way, even though I have absolutely no time to act upon it. I watch over 6,000 articles, most of which are within the scope of WikiProject Israel. Here are the results for the 30 most recent entries:


 * Useful: 7 (23%)
 * Arguably useful: 9 (30%)
 * Good faith, but useless: 12 (40%)
 * Useless: 2 (7%)
 * Rejected: 0 (0%)

—Ynhockey (Talk) 00:22, 27 January 2013 (UTC)

Unionhawk's results
Unfortunately, as the bulk of the feedback I see is from Minecraft, I don't see a whole lot of variety. However, my results shed some light on the usefulness of feedback for articles within the scope of WP:VG:


 * Useful: 2 (5%)
 * Arguably useful: 8 (20%)
 * Good faith, but useless: 23 (56%)
 * Useless: 3 (7%)
 * Rejected: 5 (12%)

Most of the feedback I read was certainly in good faith, however, the majority of suggestions involved the mechanics of the game, which I can't include due to WP:GAMEGUIDE. The redirect from Herobrine doesn't help either...--Unionhawk Talk E-mail 20:52, 28 January 2013 (UTC)

FeydHuxtable's results

 * No of feedbacks evaluated: 89
 * Found useful: 85 (96%)
 * Not useful: 4 (4%)

I evaluated the whole set of 89 feedbacks from Capital Market, which is admitedly the article from where I've found AFT most helpful. Elsewhere I'd agree that maybe only about 60% or less is useful, as per Oliver's feedback quality assessment.

The specific feedbacks I didnt find helpful were: 'please give the required information', 'maybach', 'wiki capital market' & 'whatever we ask put in full'. The first 3 I didnt know what they meant, the last one seemed rather cheeky. I found the rest useful for one of 4 reasons. 1) Specific actionable requests – admittedly there were only 6 or 7 of these. 2) Thankyous or complements on the articles quality - these are my faves as I find it motivational to see folk appreciate our work, especially as lots of the IPs seem to be from non English speaking countries. 3) Non specific requests for clarity. Individually these arent all that helpful, as an article's clarity is partly a function of the quality of attention paid to it by the reader. Also some subjects have irreducible complexity, where an increase in clarity might not be possible without over simplifying.  But this sort of feedback is useful in aggregate, as if I see a lot of feedback requesting clarity, I know to spend time looking to see if key sentences could be reduced in size,  if simpler words could be used, if ambiguity could be reduced with better grammar, etc etc.  4) General requests for more info. Similar to type 3 these are more useful in aggregate rather than individually. Im expecting this sort of feedback to be even more useful once AFT is rolled out globally, as then I'll know where there's greatest demand for expansion for all the articles on my watchlist.

PS. I never rate feedback as I find almost all of it useful, but I have been responding to it for months, e.g. per this edit summary where I made an approx | 2000 word expansion 'per feedback', or when I've left messages after seeing the same IP make feedback multiple times. PPS. Obviously Im a big fan of feedback, but I dont think it should be retained indefinitely if most of the editing community continue to oppose it. FeydHuxtable (talk) 18:32, 30 January 2013 (UTC)

Phoebe's results

 * I've gone through lots of feedback, but I just went through the feedback for one specific article, pre-eclampsia, which is #678 on WP:5000, receiving ~70K hits last week. There were 42 feedbacks. A number were blank, two were in spanish, only one was nonsense. I found at least 9 comments that were genuinely helpful, including a citation to a specific study, and a few more well-intentioned but less helpful comments. (By genuinely helpful, I mean things like "I was curious about when they first began using calcium to treat eclampsia." -- as a researcher, I can act on that). For this article -- highly trafficked but likely not one that attracts a lot of drive-by vandalism, and where a substantial portion of the readers have a strong interest in getting immediate good information (as well as a number of people apparently researching the subject for homework etc) -- it seems worth it. -- phoebe / (talk to me) 18:15, 7 February 2013 (UTC)

View by Fabrice Florin (WMF)
Hello, MZMcBride -- and thanks to everyone for your thoughtful comments about Article Feedback v5. You've given us much to think about, as we plan our next steps for this reader engagement tool.

As product manager for editor engagement features at WMF, I would like to offer some clarifications about this project, to help us reach a more informed decision about its future on the English Wikipedia.

I would first like to point out that most of the comments below [above] are based on the old version of the software which we released last September, rather than the new version, which will be available for testing in a couple weeks -- based on recommendations from the many community members who have guided this project since its inception.

This new version is expected to greatly reduce the editor workload by providing simpler moderation tools, as well as better filters to surface useful feedback and make the best comments more visible to editors. We also plan to show unreviewed comments in a separate tab, and auto-archive comments which have not been moderated after a while, to limit exposure to potentially inappropriate feedback. To learn more, check out this project update, or read the full details on new features under development.

As a result, this discussion seems premature, even if it is well intentioned, because many of the issues raised here will be addressed by our new version. With that in mind, I would like to respectfully suggest that we hold off on passing a final judgment until everyone has had a chance to test that new software. Would you be open to extending the term of this RfC by a few weeks, to allow for a fair and thorough evaluation of the new version?

And while we fully empathize with everyone's concerns about increased workload and feedback quality, please consider that last year's comments are not an accurate indicator of the feedback we expect to collect with the new version, for a couple reasons. First, we only turned on a few automated filters at the end of the year, which allowed a significant amount of inappropriate feedback to be posted; this problem can be effectively prevented by our new abuse filters, without requiring any further moderation. Second, the amount of unmoderated feedback is largely due to the fact that we limited the tool's visibility while it was still in development: many editors don't even know that comments exist for their articles, as there is no visible link to comments on article pages -- until the new version comes out, when we aim to provide a link to useful feedback for editors who have this feature enabled. To sum up, we expect the new filtering and moderation tools to significantly reduce the amount of noise and overall workload, compared to last year's implementation.

Lastly, I would like to clarify the original purpose of this tool: we started this project to engage the millions of readers who would otherwise not contribute at all, by giving them an easy on-ramp towards participation. The current talk pages do not appear effective for collecting reader feedback, because they are very intimidating to new users (who are not as tech-savvy as you are), requiring them to learn wiki markup just to post a quick comment; these talk pages are not visited often (with ), and anonymous users rarely post on these pages. Nowadays, most large information and journalism sites provide simple feedback tools so that readers can easily communicate with editors; as a top 10 web site, Wikipedia also has a responsibility to provide a practical solution to the public we serve. It appears that Article Feedback addresses that need effectively, and about 70% of users who posted feedback expressed satisfaction with the tool in our last surveys.

From an editor engagement standpoint, about after posting feedback, and we project that hundreds of thousands of new members are likely to register each year as a result; these new members can then be invited to make productive improvements to the encyclopedia through our other engagement programs. The net effect is that this tool would give readers a voice and lead them to contribute more actively -- which can help curb the editor decline and ultimately reduce the overall workload for current editors over time.

I hope these observations will be helpful in your continued discussion of this tool. As my WMF colleagues Eloquence [//en.wikipedia.org/w/index.php?title=Wikipedia_talk:Requests_for_comment/Article_feedback&diff=534460308&oldid=534458421] and Okeyes have pointed out, the foundation will respect the community's decision regarding a full deployment of this tool on the English Wikipedia. But we recommend that this decision be based on the latest version of the tool, and that it take into account the broader question of how to best engage our readers to contribute to the growth of our movement.

Thanks for listening. I'm very grateful for all your good insights, which are invaluable to us!

Respectfully, Fabrice Florin (WMF) (talk) 20:54, 24 January 2013 (UTC)


 * Users who endorse this view


 * Comments
 * Hi Fabrice. Thanks for adding your view here. When I read "most of the comments... are based on the old version of the software," I find myself very frustrated and annoyed, as this is unambiguously untrue. Similarly, your comment that "we recommend that this decision be based on the latest version of the tool" is similarly frustrating, annoying, and arguably deceptive and unfair. The comments here on this page are based on the current version of the software; the decision being made here is being based on the latest version of the tool. You've discussed abuse filters, both current filters and proposed future filters. As I discussed above, these abuse filters are incredibly heavy-handed and will result in thousands of false positives if this tool is deployed to all articles. How do you plan to address this? --MZMcBride (talk) 22:33, 24 January 2013 (UTC)
 * Hi MZMcBride, I'm sorry that you found my comments frustrating and unfair. I can assure you that I did not intend to annoy or deceive anyone. I believe that I wrote a polite, honest and accurate report on the Article Feedback project as it now stands, with a goal to provide useful information that might help this discussion. For example, it is indeed the case that most comments on this page are about the old version of the software that was deployed last summer (with last major feature tweaks made in September) -- not the new version which we have been developing in recent months, but which has not yet been released. Our developer was busy working on Wikivoyage, code refactoring and infrastructure development through the end of 2012, so no significant new features were deployed since September. The new version, which will be available for testing in mid-February, includes some important new features that were designed specifically to reduce the editor workload. For that reason, we recommend that this Request for Comments be extended through mid-March, so that editors have enough time to try out this new version, then vote on whether or not the improved tool belongs on the English Wikipedia. This seems like a fair and reasonable proposal, which we hope you can agree with. In response to your question about abuse filters, Oliver has already provided an accurate answer above, so I will only add a couple points: 1) we started with the exact same vandalism regex code for feedback filters as for edit filters, so the same swear words were being disallowed for our tool as for any edit on the English Wikipedia, using a methodology which the community appears to find acceptable already; we since removed a few words that seemed ambiguous, and are open to removing more words, if the occurrence of false positives for these words proves to be statistically significant; 2) when a comment is disallowed by the abuse filter, a text prompt lets the user know why and invites them to revise their comment and try again, as shown in this screenshot; so it is possible for users to post revised comments by removing any words that are disallowed by the filters. These abuse filter changes are relatively easy to tweak, so I expect that we could make improvements together to better calibrate this part of the tool if it is widely deployed. So this particular issue doesn't strike me as a show-stopper for this discussion, since we have a number of remedies at our disposal. Lastly, I encourage us all to collaborate more closely on this project, building on common ground we can agree on, rather than focusing on what separates us. In online discussions like these, it's easy to get polarized into opposing camps, which can create mistrust and be frustrating for all. But I believe we all share the same overall goals (e.g.: improve people's lives, spread free knowledge), regardless of our backgrounds or methodologies. I know this sounds naive, but do you think can we put our differences aside and find ways to work together as partners? This may require some compromise on everyone's part, but the outcome could surprise us, if we all give it a try. Fingers crossed ... :) Fabrice Florin (WMF) (talk) 09:51, 28 January 2013 (UTC)
 * Sigh. You seem to have a fundamental misunderstanding of how abuse filters work. If I write "The article feedback tool is a cunt" in an article or in an edit summary, nothing will stop me (autoconfirmed or not). If I write this same sentence in a feedback comment, I will be stopped (if not autoconfirmed). Why is this a problem? Because many "profane" words come up in thousands of contexts. Fag, dyke, bitch, cunt, fuck, etc. all have their own articles, but they're also related to other subjects (book titles [Dude, You're a Fag], album titles [ Cunt (album) ], band names [ Fuck (band) ], etc.). Given that nearly all of this feedback is from non-autoconfirmed users, there's a serious problem here. You still haven't addressed how you plan to deal with this. Instead, you provided an answer that shows you fundamentally don't understand the current architecture (i.e., that feedback filters are far more potent than edit filters), which I find really worrying and upsetting, given your role in the project. Regarding building common ground, compromise, etc., I posted (at length) here about my views toward a compromise position. I encourage you to read that comment and respond, as you see fit. --MZMcBride (talk) 05:49, 29 January 2013 (UTC) P.S. And it wouldn't hurt to observe basic wiki-etiquette, such as referring to a user by his or her username (or a shortened form of his or her username), unless their user page (or user talk page, I suppose) indicates that he or she can be called otherwise. As it is, you've now created confusion in this section (and you're probably fast approaching a violation of WP:OUTING, at least in the letter of the law). :-/
 * Hi MZMcBride, thanks for inviting me to comment on your proposed compromise, which I will do shortly. I was encouraged to hold off from participating actively in this discussion, so as not to interfere with the community's deliberation. So I have limited my comments accordingly, but am happy to provide more information, if you think it can be helpful. Regarding your concerns about the abuse filters, I believe this issue is not as drastic as you suggest, because users have the option to remove words that are triggering the filters, and filter editors also have the option to remove some of these words, if they believe they are causing too many false positives. Thanks as well for the reminder to use exact usernames rather than first names, which was an oversight on my part. I have since corrected this in my comments on this section. Fabrice Florin (WMF) (talk) 23:14, 29 January 2013 (UTC)
 * I participated in some usability testing last month for what I believe were prototypes for the new version of the tool. While they did indeed try to address some concerns people have expressed here, I can't say they appeared to be all that different from what we have now - certainly not so different that it would change my mind about whether the tool is a net positive. Now, obviously I was shown only a handful of test cases, and you probably have far more changes than I've seen, but I honestly don't think that some tweaks - even some well-directed ones - can save AFT from its own weaknesses, which are many. As far as us !voting on the non-current version of the tool...that's pretty much always the case. Software development seems to be continuous for most WMF features, which means discussing any deployed feature means we're always discussing an "outdated" version that's not what the bleeding-edge devs are hacking at at the moment. That doesn't change the fact that we're discussing what we have here and now. If, two months from now, we were discussing the new version, you could make the same argument again. "Yes, but we've made some changes to the software that haven't been deployed yet! Please wait for v6.1 before making a decision!" And then maybe v6.5. Or 7. You deployed AFT fairly widely in this incarnation because you wanted to test it with the community. If the community doesn't like it, it's unfair to return to us with "Ok, ok, this one doesn't count. The next one will be great!" It's not that we can't believe v6 might be an improvement, it's just that this RfC is asking us to discuss based on what we know and have used, and part of that is weighing the possibility that the Flying Spaghetti Monster might touch the next version AFT with his noodly appendage and render it perfect against the fact that, well, the version we've been using patently isn't. As I've said in my comments in other sections, I have tons of respect for what you guys are and have been trying to do with this tool, and I don't want you to think people are making these comments out of some sort of ignorance about how great AFT could be. We know it might be improved, but we're balancing that against the fact that "how much" is unknown and the fact that current reality is the only thing we can know for sure. A fluffernutter is a sandwich! (talk) 22:57, 24 January 2013 (UTC)
 * Hi Fluffernutter, thanks for testing our early prototype of one of the new features now under development. Despite your first look at that feature, I recommend that you wait until all these features have been deployed to pass a judgment on their effectiveness, as all they all fit together into an improved user experience where the whole is worth more than the sum of its parts. We are only asking for a one-time deferral of the community's vote about this improved version -- not for a series of votes, as you seem to suggest. If a majority of editors who have tested the new tool decide that they don't like this improved version, and if we don't reach a compromise such as Max's proposed opt-in version, we are prepared to remove all versions of Article Feedback from the English Wikipedia after that vote has been finalized. Thanks as well for letting us know that you respect the hard work that has gone into this tool. This means a lot to me and other team members, who have poured our hearts and souls to develop this editor engagement tool. We are deeply aware that it may increase the workload for some editors, but sincerely believe that the new tool will significantly reduce that workload to the point that the tool's editor engagement benefits can outweigh its downside. We will not stand in the way if we are proved wrong, but are asking for one more chance to test these improvements. Would that be all right with you? Fabrice Florin (WMF) (talk) 09:51, 28 January 2013 (UTC)
 * With all due respect, it sounds like y'all have already decided what you want to do, which unfortunately appears that y'all intend to disregard the input of the WP community. Currently the clear consensus is to do away with the tool, which has more !votes than the next two options combined.  When you consider that the next two largest views supported are an "opt-in option" and a "replace with a suggestion to edit", it is overwhelming evidence that the community does not desire the tool.  I hope that I'm wrong about y'alls intentions, and I invite your response to this concern.  Regards,  GregJackP   Boomer!   23:19, 24 January 2013 (UTC)
 * I'm confused; Fabrice explicitly said " As my WMF colleagues Eloquence and Okeyes have pointed out, the foundation will respect the community's decision regarding a full deployment of this tool on the English Wikipedia." immediately above. Okeyes (WMF) (talk) 23:42, 24 January 2013 (UTC)
 * I'm confused as well. Greg's view is about completely disabling the tool. Fabrice's comment is about not going forward with a full deployment of the tool. Can you clarify or ask Fabrice to clarify? --MZMcBride (talk) 23:57, 24 January 2013 (UTC)
 * Sure; the WMF viewpoint (I think I can now safely say) is that if the community doesn't want the tool, the community doesn't get the tool, full stop. We respect the consensus the community comes out with on this feature, whether that's 'put it everywhere' or 'put it nowhere' or something in-between. Okeyes (WMF) (talk) 23:59, 24 January 2013 (UTC)
 * Sorry, I wasn't clear. I saw Fabrice's comment, but full deployment doesn't mean remove, it can be partial deployment, leave as is, etc.  Based on the his overall comment, it appeared to me that while the current tool might be pulled, the new tool might be deployed.  I just wanted clarification.   GregJackP   Boomer!   00:45, 25 January 2013 (UTC)
 * "do away with the tool, which has more !votes than the next two options combined" ... what I just posted to the talk page may help to explain why some of those !votes are there. --Demiurge1000 (talk) 00:00, 25 January 2013 (UTC)
 * Updated here --Demiurge1000 (talk) 08:12, 25 January 2013 (UTC)
 * I don't have a clue about any canvasing, as I don't get on IRC or any other media. All I did is look at the count on this page.   GregJackP   Boomer!   00:45, 25 January 2013 (UTC)


 * Thanks for your insight. Despite my reluctant opposition above, I am glad that efforts are being made to stem the tide of editor decline. My question is: has there been any more quantitative projections of how this will impact editor signup and retention? How did you reach the figure of hundreds of thousands? From what I've read (correct me if I'm wrong), 205 editors signed up over a 12 day period on a trial of 10% of articles. So naïvely we'd expect 170 per day on 100% of articles, or just over 62,000 new editors per year (of course, that sample is insufficient to make that claim with any accuracy). Are these new accounts still being monitored? What number are still "active" in some sense? Because a sustainable body of editors is presumably the desired outcome, rather than the surrogate measure of signups. If rough projections were made available before the next trial phase the community might be able to more objectively assess how successful AfT is proving to be relative to the Foundation's expectations <font color=#000000>Jebus989 ✰ 23:56, 24 January 2013 (UTC)
 * Hi Jebus989, thanks for your question about quantitative projections on editor signup and retention. Our signup forecast is based on these two estimates for a 100% deployment: 900,000 feedback posts per month x 2.7% signup rate = 24,000 new signups per month, or 290,000 new signups per year. We do not yet have any conclusive data on the productivity of these new editors, but plan to collect that data as part of a wider deployment. Keep in mind that research studies like these take time to set up and analyze, so it's not likely that we would have these results until at least a month after our next deployment. We are, however, prepared to set definitive targets for a wider release, so we can track our progress. These targets can be finalized once we reach an understanding with the community on how widely the tool can be deployed and under what terms. Fabrice Florin (WMF) (talk) 09:51, 28 January 2013 (UTC)
 * Sounds great, thanks for the response <font color=#000000>Jebus989 ✰ 10:23, 28 January 2013 (UTC)
 * You are very welcome, Jebus989. Glad this response was helpful.


 * Thank you for this update. Is the development team looking at modifying the instructions given to readers? For instance, we could remind readers more clearly that this feedback is about the article, not the subject of the article and perhaps refer questions to the reference desk. Could we perhaps provide one or two bullet points suggesting answers to common feedback requests - for instance, explicitly referring users to an article on simple english wikipedia where available? Can we ask for specific feedback? GabrielF (talk) 03:08, 25 January 2013 (UTC)
 * If you (the WMF) want us to evaluate the tool based on the new, improved version, I suggest you amend the deployment schedule so that the improved version of the tool is deployed to the current test pages well before deploying the tool to all Wikipedia pages. The current test on a limited selection of Wikipedia pages has been a failure. Continuing to work on improving the tool is a good idea, but given the experience with the current versions of the tool the Wikipedia community are going to need several months to evaluate the improved version of the tool before full deployment occurs. --Srleffler (talk) 06:33, 25 January 2013 (UTC)
 * Hi Sreiffer. You make some very good points, thank you. Of course, if the community needs more time to evaluate the improved tool, it would make sense to delay the wider deployment, as you propose. Rest assured that we will not deploy widely without community approval. Though our hope is that we could get an evaluation completed within a few weeks from the deployment of the improved version, if possible. But we're all reasonable folks, and are happy to adjust our plans as needed. Fabrice Florin (WMF) (talk) 09:51, 28 January 2013 (UTC)

It is false that one needs to learn wiki markup to write in a talk page - and it is likely an editor would assist whenever some is needed. Note that the only wikimarkup I'll use here is my signature, but that is not a must have for a casual reader (plus it is quite likely a bot will add it). Though your reasoning makes sense overall (don't use old data in evaluating, and given there is a new version wait for it) adding obviously wrong reasons quite powerfully makes you look silly and downgrades your opinion. (I have no idea who the heck is "Max", why use names other the the WP-usernames?) - Nabla (talk) 21:42, 28 January 2013 (UTC)
 * I've been asking this question myself for years. Who's Max? --MZMcBride (talk) 05:50, 29 January 2013 (UTC)
 * :-) From the context2 I'd guess who (at least one of) Max is. But I might get it wrong... - Nabla (talk) 21:39, 29 January 2013 (UTC)
 * Hi Nabla, thanks for correctly pointing out that wikimarkup is not strictly required to post comments on a talkpage. The fact remains that posting on talkpages appears difficult for most new users, who are used to simpler feedback tools that are less intimidating than our current editing tools. We are in the process of starting usability studies to assess the severity of this issue, so we don't yet have definitive evidence to support this hypothesis. But based on prior experience with these types of issues, I expect that our findings will demonstrate that posting on talk pages is a serious obstacle for most readers, unlike the article feedback tool. If AFT is not going to be deployed widely on the English Wikipedia, then a better solution appears needed to engage readers to participate more actively. Given that our team is getting ready to move on to other projects, we may not be able to revisit this issue for another year or more. Which is why I think that MZ's opt-in compromise is a more practical proposal than removing the AFT tool entirely.  Fabrice Florin (WMF) (talk) 23:14, 29 January 2013 (UTC)
 * P.S.: I'm sorry that I incorrectly used 'Max' instead of 'MZMcBride', my bad. I have since corrected this in my comments on this section and will refer to actual usernames going forward. Thanks for the reminder! Fabrice Florin (WMF) (talk) 23:14, 29 January 2013 (UTC)

View by Guy Macon
(I have read https://www.mediawiki.org/wiki/Article_feedback/Version_5/Feature_Requirements and https://www.mediawiki.org/wiki/Article_feedback/Version_5/Technical_Design, and unless I really missed something, the following applies to the proposed future version as well.)

Like many of you, when I first heard about "an encyclopedia that anyone can edit" my first reaction was that it could not possibly work. I saw USENET get ruined by trolls and spammers. I have seen web site after web site prove the dual laws of online nature: if you don't have heavy restrictions, moderators wielding hair-trigger banhammers, intricate registration requirements and software that doesn't let anyone edit anyone else's work, the site gets flooded with abuse and all the good users leave. And if you do have all of the above, it doesn't scale -- there is a limit to how much traffic the moderation system can handle.

Wikipedia, and pretty much no other online entity (with Reddit and Slashdot being possible exceptions), somehow put together a system that avoided the failure and which grew and grew without size alone destroying the system. I think we need to carefully consider why this happened and to avoid any addition that does not have whatever attributes make Wikipedia so abuse resistant.

One thing in particular seems to be the core of Wikipedia's success. Yes, someone can replace a page with "PIGGERS ARE AWSOME!! GO PIGGERZ! WO0T!!!", but they have to type 40 characters to do that, only to have someone revert it with one or two clicks 2 seconds later (or maybe zero clicks and zero seconds if a bot catches it). So when I saw the new feedback system I looked for that. I saw that there was just a couple of buttons to click and no way to enter text, so I had no problem with it. But later, there was a new system that lets anyone type in anything, but which does not allow me to revert vandalism or spam.

Yes, I am a rollbacker and thus can hide abuse, but looking at the help pages I see that this is really changing the visibility, which screws up another basic feature that makes Wikipedia work: anyone can view and undo my revert. So we end up with WP:BRD without the R. And really, without the D as well; has anyone actually had a discussion with someone about abusive article feedback they posted?

I don't think article feedback is useless. I am unconvinced that our resources cannot handle the load. I think that something a lot like what we have now could turn out to be a huge success. But a post that says "PIGGERS ARE AWSOME!! GO PIGGERZ! WO0T!!!" or "asdffghjkl" or "0n1in3 \/|ag.ra!! www.areyoustupidenoughtogivemeyourcreditcardnumber.ru" needs to be revertable by any autoconfirmed user with a click or two, and that revert itself needs to be revertable by any autoconfirmed user. --Guy Macon (talk) 03:43, 25 January 2013 (UTC)

ADDITION: All of my concerns are addressed by the "Compromise proposal by Noleander". --Guy Macon (talk) 01:45, 9 February 2013 (UTC)


 * Users who endorse this view


 * 1) No endorsements? Well I certainly endorse this view! In fact I agree 100% with Guy Macon on everything! What are the odds? :) --Guy Macon (talk) 11:24, 26 January 2013 (UTC)
 * 2) I only agree with 99.9% of what Nabla does and write... And I agree with just about that much of what you wrote here - Nabla (talk) 22:21, 28 January 2013 (UTC)
 * 3) Yep. Pete (talk) 02:08, 30 January 2013 (UTC)
 * 4) Additionally, talk pages have better responses to useless feedback? "Article has no images" can be answered with "We need images with a free licenses, you can upload one like this and blah blah". Other people will see the reply and won't ask the same question again, or they might upload a picture. If someone repeats the question, you can point them to the first discussion. Feedback numbers may be inflated because people keep asking the same question again and again, and there is no mechanism to address this. --Enric Naval (talk) 12:51, 6 February 2013 (UTC)


 * Comments

I'm not sure what "revert" means in the context of AFT5, but I do think we need to give more powers to more moderators, and I would be happy to see all rollbackers having the "Hide" ability in AFT5. Dcoetzee 21:15, 30 January 2013 (UTC)

Quantifying abuse
The article feedback tool currently doesn't have the features you mention in your last paragraph. In the short-term, what are your thoughts on the deployment of the tool? Should it go from 10% of articles to all articles, as the Wikimedia Foundation planned? Should the tool be completely disabled until it can be improved? Something else? --MZMcBride (talk) 04:50, 25 January 2013 (UTC)


 * That's a tough one. (...goes away and ponders it for an hour or so while doing chores...) Here is what I would do if it were my decision. I would randomly pick 1000 article feedbacks (I have a really good random number generator) and count how many are blank, spam, nonsense, etc. if it was less than 5% I would say go ahead and deploy 100%. If it was 5% to 10% I would say leave it at 10% of articles and work on the abuse problem. If it was 10% or more I would say remove it from all articles without prejudice - meaning that they can come back and do a 1% trial then a 10% trial after they believe that they have the abuse problem solved. --Guy Macon (talk) 06:26, 25 January 2013 (UTC)


 * All right, let's do it. I believe I have the equivalent of Special:ArticleFeedbackv5 --> "All visible" at my fingertips. I ran this query on the Toolserver's enwiki_p: "select count(*) from aft_article_answer where aa_field_id = 17;". The database says there are approximately 531,717 rows. The text associated with these rows seems to be exactly what appears at Special:ArticleFeedbackv5 --> "All visible". That is, when I pull the most recent text responses directly from the database and compare to the text responses at that Special page (Special:ArticleFeedbackv5 --> "All visible"), I'm seeing identical results. I guess we're on the 17th iteration of this tool's design (aa_field_id = 17)? Someone else will surely correct me if I'm wrong about my logic.
 * I pulled the associated feedback IDs (using "select aa_feedback_id from aft_article_answer where aa_field_id = 17;") and put them in a text file here (warning, it's a 531,724-line text file, you may want to "save as" rather than opening it in your Web browser). If you can pull 1,000 random (or as close to random as possible) IDs from that list, I'll pull the associated comment text, stick the sampling on a wiki page, and we can code the comments together.
 * I don't think anyone (myself included) will go for your percentage system as a means of deciding whether to deploy this tool site-wide, but I'm interested enough in the results of surveying such a sampling that I'm willing to donate some time and energy to conducting this survey. Let me know when you have the 1000 random IDs selected.
 * Let me save you some work. If all you want to do is decide, yes or no with good confidence, whether the true proportion for all IDs is above or below 5%, then you don't need anything like a sample of n=1000; n=600 will give you a give-or-take figure of 1% or better (assuming the % you see in your sample is about 6% or smaller -- if it's above 6%, then you've resolved your question anyway). Contact me if you want to know more. EEng (talk) 11:59, 27 January 2013 (UTC)
 * I have spent many happy hours punching numbers into my HP35 calculating sample sizes, until I realized that there is a book with the answers in a big multipage table. Later I wrote a MS-DOS program, but now I simply plug the numbers into [ http://www.nss.gov.au/nss/home.NSF/pages/Sample+size+calculator ].
 * I did not want to lock in my 5% number -- what if someone likes the general idea but with 10% or 25%? Plugging in confidence level=95%, population size=61887, proportion=0.05 and sample size=600 does indeed give me a confidence interval of 0.017, but with proportion=0.1 the confidence interval is 0.024. Bumping up the sample up to 1000 gives me a slightly better confidence interval of 0.018 for a proportion of 0.1 and 0.013 for a proportion of 0.05. --Guy Macon (talk) 13:44, 27 January 2013 (UTC)
 * I wrote this reply and then realized that the list of feedback IDs I pulled will exclude any comments that were filtered, of course (which includes comments that are blank, short, or that contain profanity). The AbuseFilter catches feedback prior to database insertion. So we'd need to come up with a better way of sampling the IDs, I guess. Hmmm. --MZMcBride (talk) 09:30, 25 January 2013 (UTC)
 * I ran a quick  and got 61887. Maybe you can somehow factor in that a certain percentage of supposed feedback will be rejected? Legoktm (talk) 20:55, 25 January 2013 (UTC)
 * I think that the filtered list is what we want for this purpose. Knowing what percentage of the raw total are blank, etc. would be interesting, but for our purposes it is the percentage that make it past the filters and the editors and can be seen by the reader that tells us whether the system has too much unfiltered and unhidden abuse to deploy. I just grabbed the file and will select 1000 at random. --Guy Macon (talk) 21:02, 25 January 2013 (UTC)
 * List is at User:Guy Macon/Workpage --Guy Macon (talk) 21:33, 25 January 2013 (UTC)
 * I really interested in what the results will be. I would like to have the 1000 feedbacks I randomly selected visible somewhere so that anyone who doubts the useful/useless criteria can do their own count. --Guy Macon (talk) 09:33, 26 January 2013 (UTC)
 * Okay, posted. Though I'm not sure about this output format. Maybe links are needed to the specific comment? Or perhaps an article column is needed? Let me know what you think. --MZMcBride (talk) 10:27, 26 January 2013 (UTC)
 * Interesting! Far different than the ones on my watchlist. Some of those make me want to click on a link and check the article (or do I want to see the version as it existed when the comment was posted?). It might be nice to have a place to add notes like "Poster missed the section on dwarfs and the link to "list of dwarfs in universe X" or "date of birth is in the infobox". --Guy Macon (talk) 11:19, 26 January 2013 (UTC)
 * (edit conflict) The sample is nice for people to have fun, but research already proved that 12 % of posts are abusive to the point of being moderated and 60 % are unhelpful, unless I'm misreading reports. --Nemo 11:23, 26 January 2013 (UTC)
 * See http://blog.wikimedia.org/2012/12/20/article-feedback-new-research-and-next-steps/ --Guy Macon (talk) 18:29, 26 January 2013 (UTC)
 * I asked to see the raw data from these reports. Do you know if it's available? I'm curious to know whether comments such as "needs image" were considered helpful when coding the responses. The subjectivity of rating responses seems very problematic to me. --MZMcBride (talk) 15:14, 26 January 2013 (UTC)
 * I am sure that they meant well, but when I see a bunch of slides, (one of which actually has "(Raw Data)" on the slide) but which don't actually say where I can find the raw data, I suspect whatever unknown methodology was used to create those slides. I have 1000 samples of current (not months old) article feedback, chosen at random from feedback that is visible to the users, and I am going to make sure that the process of deciding which ones are useless is completely transparent. --Guy Macon (talk) 18:53, 26 January 2013 (UTC)
 * They used double blind; "completely transparent" doesn't mean better, you should ask the criteria I guess. They're probably on Meta somewhere. --Nemo 21:13, 26 January 2013 (UTC)
 * Transparency isn't just "better", it is the difference between science and non-science. If I cannot replicate the experiment, then it is not science. You should not have to guess that the methodology is "on Meta somewhere". the methodology and the raw data should be referenced in the result. As for whether "they used double-blind", you know this ... how? Can you explain how the control group was selected? --Guy Macon (talk) 21:45, 26 January 2013 (UTC)
 * You seem to be a bit confused about what's science. Methodology and selection were in the slide you removed from this section. The links to more details are also there but don't work in the PDF, you can surely ask DarTar. I've not spent more than 1 min looking for more info because I suspect it may also be on this wiki rather than Meta and this makes me angry, but you can surely find more with a few minutes' work. --Nemo 22:40, 28 January 2013 (UTC)
 * Please be WP:CIVIL. "You seem to be a bit confused about what's science." is an insult. I would argue that your making claims like "double blind" after admitting that you have no clue as to what the methodology was is evidence of being "confused about what's science". --Guy Macon (talk) 19:07, 19 February 2013 (UTC)
 * Pete at found and analysed it: m:Research:Article feedback/Final quality assessment. --Nemo 12:08, 30 January 2013 (UTC)
 * Was providing a citation backing up your claims really so difficult that you had to resort to personal insults? --Guy Macon (talk) 15:13, 2 February 2013 (UTC)
 * I'm sorry you felt insulted, it's surprising to me. The citation for my claims was provided since the beginning (it's all written in that slide). It's you claiming it wasn't true; thanks for elaborating below. --Nemo 21:46, 2 February 2013 (UTC)
 * To answer above; nemo, a standard requirement of scientific studies (for them to be scientific) is falsifiability - that the methodology and evidence can be tested, and that the subject can be further tested. Guy's comment on this seems reasonable. In terms of original data, we're currently getting permission from the hand-coders to make the dataset public. And, no, the research information is all on meta: I hope this makes you not-angry :). Okeyes (WMF) (talk) 00:22, 3 February 2013 (UTC)
 * Nemo: before I say the following, I want it to be crystal clear that WMF did nothing wrong and has been completely above-board in all of this. The differences I have discussed are mostly because I was looking for the answer to a different question. That being said, [A] There exists no slide which contains a link to the raw data or to a detailed description of the methodology, [B] "You seem to be a bit confused about what's science" is indeed insulting, [C] neither effort is a true double-blind study (nor is one needed in this situation -- we aren't exactly testing anticancer drugs here), and [D] questioning the methodology is a Good Thing which the WMF clearly welcomes, not some sort of attack which is cause for anger. We are all on the same team here, and this conversation will end up with us having a better answer to the question of what percentage of the article feedback is useful. We already learned one thing; checking just the feedback from your watchlist gives you wildly inaccurate results. --Guy Macon (talk) 00:50, 3 February 2013 (UTC)
 * Oliver, I don't think that "raw data" is a requirement at all here, I guess it would be nice; and yes I'm very happy that the page is on Meta. Guy Macon, [A] already answered above, [B] sorry about that, [C] ok, [D] you are the one who didn't read the slide about methodology or looked for it on Meta. ;-) --Nemo 16:37, 3 February 2013 (UTC)
 * [Citation Needed]. It would seem to be an easy thing to post a link to this alleged slide that contains a detailed methodology. Is there any reason why you refuse to do so? I did search for it, and found several candidates that summarize parts of the methodology, but until you say which one you think contains a detailed methodology, I cannot examine that claim. You do realize that a detailed methodology cannot possibly fit on one of those slides using those fonts, right? --Guy Macon (talk) 11:38, 4 February 2013 (UTC)
 * Of course there's a reason, my dear: you undid me last time I did. Also, who spoke of detailed methodologies on a slide? --Nemo 18:12, 4 February 2013 (UTC)
 * "They used double blind" --Nemo 21:13, 26 January 2013 (UTC)
 * "Methodology and selection were in the slide you removed" --Nemo 22:40, 28 January 2013 (UTC)


 * "The citation for my claims was provided since the beginning (it's all written in that slide)." --Nemo 21:46, 2 February 2013 (UTC)


 * I am done with you. You won't provide a link because you can't provide a link. Nowhere was the methodology published, I never removed anything, there is no such slide, I am not your dear, and your claims are completely without merit. You can reply if you wish, but I won't read it. I have better things to do that read reply after reply from someone who cannot back up his assertions with actual evidence. --Guy Macon (talk) 18:59, 4 February 2013 (UTC)

Quantifying abuse two: The Wikimedia research
After much difficulty, I finally received a link to the methodology used for arriving at the figures that we keep seeing on slideshows. It is at [ https://meta.wikimedia.org/wiki/Research:Article_feedback/Final_quality_assessment ].

I was most interested in whether that study would, as several here have suggested, make analyzing my sample of 1000 article feedbacks moot.

It turns out that that study did not measure the same thing that we did. This is not to imply that they did anything wrong. It just means that they were trying to answer a different question.

Let's look at the differences:

First difference:

We took a random sample of 1000 from a snapshot of all visible article feedback at one moment of time. They took a stratified sample of the feedback posted during a one month period, then re-sampled it, giving them 808 samples.

Our sample is better for answering the question "after we have did our best to get rid of abuse, what percentage of the feedback that users can still see (old and new) is useful?" Theirs is better for answering the question "what percentage of the incoming feedback is useful?" Both are legitimate questions, but they are not the same.

Second difference:

We have made our raw data visible for anyone to see, If they have the raw data someplace, I cannot find it. Showing the raw data will allow the user to find out how the numbers change if, for example, "needs a picture" or "Good article" is counted as useful or useless.

Third difference:

They used "at least two volunteer Wikipedians." The page for selecting volunteers and the evaluation instructions are at [ https://en.wikipedia.org/wiki/Wikipedia:Article_Feedback_Tool/Version_5/Feedback_evaluation ]. Significantly, at Article Feedback Tool/Version 5/Feedback evaluation the instructions s ay "It is only the most entirely useless feedback that should be categorized as 'no' (not useful)."

We have not finished the evaluation portion of our study. I think it would be safe to say that none of us has come up with a way to avoid selection bias of our evaluators.

Given the above, I think we should continue with our evaluation of our sample of 1000, have several editors evaluate which ones are useful, and publish a summary along with the raw data --Guy Macon (talk) 15:13, 2 February 2013 (UTC)

Heavy-handed abuse filters I have noticed some comments about heavy-handed abuse filters. These are another symptom of an attempt to allow anyone to post article feedback without allowing anyone to revert the abusive edits. No online forum has ever succeeded at a robot-only spam filter. It just starts an arms race in which the dedicated human beats the robot every time. Robots are great at getting rid of 90% or even 99% of the spam, but if it stops there the 1% multiplies as the spammers figure out what works. --Guy Macon (talk) 09:33, 26 January 2013 (UTC)

View by Jane023
I quickly scanned this page and realized most of the comments here are regarding the usefulness of the tool from the editors' point of view, but no one has looked at it from the point of view of the user. When I post feedback anywhere, I expect two things; 1) I can view my feedback post and monitor it for any other responses, and 2) I expect to see timely responses. So if I post feedback to someone's Facebook post or blog post or even newspaper articles, I expect to be able to find my contribution again. I would also expect to be able to edit this feedback if I wrote it in an inebriated state in the evening and changed my mind after coffee in the morning. Often is is dependent on the feedback interaction whether I like the website or not and whether I will be drawn eventually to "joining the project" and contributing my own pages that can receive feedback from the community at large. The current feedback tool does not take into account these motivations of the user. I believe current users might take 10 minutes to enter thoughtful feedback once, and when they find out it has disappeared into never never land, they don't come back. I also believe that the interface is so different from talk pages that there is no "growth path" for reader/feedbackers to turn into editor/feedbackers. It is a small step from editor/feedbacker (i.e. poster on a talk page) to editting articles. Jane (talk) 14:13, 25 January 2013 (UTC)


 * Users who endorse this view
 * 1) In my personal experience with feedback forms on other Web sites, nearly all of them do not allow post-submission editing. Most of them are "black boxes" (forms that get [hopefully] routed to someone's e-mail inbox), but even more traditional comments sections don't allow for changes to be made in most cases, in my experience. That said, I can endorse your view, as it I agree that the ability to reply to or edit remarks is important (and is a key feature of the talk page system we already have!). --MZMcBride (talk) 16:19, 25 January 2013 (UTC)
 * 2) Pretty much to the point.--Ymblanter (talk) 17:58, 25 January 2013 (UTC)
 * 3) The lack of a growth path is an important flaw. Of course the low response rate is another indication of this being bad for readers, whilst setting it to a much stricter edit filter whilst clearly necessary has also increased the proportion of people for whom it will be a counterproductve bad experience.   Ϣere  Spiel  Chequers 20:01, 25 January 2013 (UTC)
 * 4) I also agree. This pretty much goes in step with what I said above. Nobody's talking about the motivations of a new user to join, nor the motivations of an old user to stay. I think both sides are equally important. Zaereth (talk) 23:09, 25 January 2013 (UTC)
 * Sorry, I missed that and I do agree with your view; especially about the fact that the feedback tool basically creates yet another watchlist for editors to look at, whilst many editors barely have time for their watchlist. Jane (talk) 14:11, 7 February 2013 (UTC)
 * 1) Agreed. If one of the intents of the AFT is to engage users/readers in the hopes that they might become contributors or just to make WP a better encyclopedia, then having them leave AF with no action/low-visibility would seem to be counterproductive.  I think the idea in the next section about possibly having an AF talk page/sub-page might be one possible way to address this issue.  Shearonink (talk) 16:49, 26 January 2013 (UTC)
 * 2) Nick-D (talk) 07:20, 27 January 2013 (UTC)
 * 3) While I think the current system is a step in the right direction, I think sending people a link to their AFT comment that they can follow and get notifications about if they wish is a really critical future feature to implement. I imagine doing this by having a final screen that simply says "enter your e-mail to receive notifications about your comment". Dcoetzee 01:45, 28 January 2013 (UTC)
 * 4) The talk page already has these features. Feedback tool is trying to reinvent the talk page, but in a clumsy way that misses the point of how wikipedia became successful... --Enric Naval (talk) 12:55, 6 February 2013 (UTC)
 * 5) Pete (talk)
 * 6) --j⚛e deckertalk 16:43, 14 February 2013 (UTC)


 * Comments

I think it's unusual to get a response to website feedback. In fact, most websites have a note about "all feedback is read, but we can't reply to most of it". Perhaps a similar disclaimer would help manage expectations here. WhatamIdoing (talk) 19:39, 4 February 2013 (UTC)

View by EEng
I don't think my experience with feedback quality and usefulness is typical so I won't share it. I have just one specific suggestion:

Whatever the nature of the invitation to readers, if the reader clicks to take the bait it should simply open the article's talk as "add new section" (or maybe a ===-level subsection to a special Feedback section). All this accessory apparatus storing feedback in some mysterious place seems unnecessary. Automatic edit summaries, special links automatically added during Save, and similar machinery can provide the functions of the "Resolved" and other buttons in place now. The obvious advantage is one-stop shopping (the talk page) for discussion and feedback, but more subtly it introduces the potential new editor to our normal editing apparatus.


 * Users who endorse this view


 * 1) Strong Support I think the idea of a special feedback section on the talk page (or on a sub-page transcluded to the talk page) is a brilliant idea. --Guy Macon (talk) 22:40, 25 January 2013 (UTC)
 * Not that I regret making the contribution, but the fact is that making it has left me way overdrawn at the brilliance bank. They want immediate payment, so if any actually brilliant people could send in whatever brilliance they can spare, I might be able to avoid intellectual bankruptcy. Thanks in advance. EEng (talk) 23:00, 25 January 2013 (UTC)
 * I like the idea of a subpage transcluded to the article talkpage. That way any editor visiting Talk gets at least a glancing exposure to the feedback (maybe transclude it at the top). The subpage would be a separate item on watchlists. EEng (talk) 18:17, 26 January 2013 (UTC)
 * Perhaps a transclusion that is collapsed at the top of the talk page? This could very well address my problems with the present system. Keep whatever fancy form / edit box that the WMF decides on, but instead of storing the data in some new way, just have the form / edit box store the information on the subpage. That way we can revert spam, check the history, etc., just as we do with any other page. --Guy Macon (talk) 19:53, 26 January 2013 (UTC)
 * 1) support - if the effort is to engage new editors, then we should be engaging them in the current systems and discussions and not having some type of dis-engaged posting at a standalone, outside-of-the-normal editing.  --  TRPoD aka The Red Pen of Doom  23:08, 25 January 2013 (UTC)
 * 2) I agree. I have no objection to having a second, flashy button that directs comments to the talk page, and think that may be helpful. I still want to reiterate that I think the ratings are very useful as well, and would like to see them kept. Zaereth (talk) 23:13, 25 January 2013 (UTC)
 * 3) Support. I like the idea that readers wanting to leave Feedback would be directed into the workings of the present talk-page system, instead of AF-posts being segregated off, the talk page/sub-page idea seems especially intriguing.  If the AF is integrated within the present Wiki-processes, I think that would increase the odds that experienced editors will actually deal with AF-posts, instead of the Feedback being given and then withering away without anything being done).  By having visible action, it would also increase the odds of possible reader-engagement (which is one of the AFT's goals). Shearonink (talk) 16:39, 26 January 2013 (UTC)
 * 4) Weak support. Better than the tool, but I'm concerned we would just move crap to the talkpage, especially crap that needs oversighting.   GregJackP   Boomer!   17:02, 26 January 2013 (UTC)
 * I'm agnostic on whether or not there should be a feedback system at all -- my very narrow proposal is that if there is one, it should funnel to the article's talkpage. EEng (talk) 18:17, 26 January 2013 (UTC)
 * Comment to GregJackP... Maybe if AF was set up as being a sub-page of the article talk page with a clear sightline back to the associated article and to the article talk page, crap wouldn't be sitting on the talk page itself and actionable/useful AF might not wither away unseen. If a major component of AF is to engage readers and possibly help them to become useful editors, having AF disappear into some type of netherworld does not help readers understand that they can actually  become  editors.  And if material that should be oversighted instead disappears into the standalone AF area and remains on view, if editors don't see it and therefore don't notify Oversighters, as I understand it the material just remains there (forever?).  The idea mentioned elsewhere on this page, that many/most readers apparently are unable to find the AF after it is given, should not be seen as a safety/protective feature. Shearonink (talk) 19:27, 26 January 2013 (UTC)
 * 1) I'm concerned that we are developing multiple overlapping systems for user feedback (feedback requests, talk pages, etc) and I agree that specific, actionable feedback about the article should be redirected to the talkpage where it can be discussed by editors. However, we should not direct ALL feedback to the talk page - we should make some effort to try to keep questions about the subject of the article and other inappropriate things off the talk pages. GabrielF (talk) 17:18, 26 January 2013 (UTC)
 * My proposal is that sending feedback to talkpage would replace the current specialized machinery. I'm unclear about your point about questions about the subject -- are you saying they should be routed to refdesk? That may or may not be a good idea, but let me point out that if someone has a question about the subject that he didn't see in the article, then maybe it should be added to the article (or the article modified to make the answer more findable). EEng (talk) 18:17, 26 January 2013 (UTC)
 * Comment to GabrielF... I agree with your concerns about having multiple overlapping systems. During this beta-test, it seems that one of the AF's uses is that is has become a place for inappropriate posts to go, to siphon inappropriate posts or vandalistic urges away from the actual article, away from the main article-talk page and direct them somewhere off-stage. Shearonink (talk) 19:27, 26 January 2013 (UTC)
 * This seems to me to be a very important point, one I don't see anywhere else. Can we somehow measure whether page vandalism / unconstructive edits were reduced by the presence of the feedback box? EEng (talk) 01:22, 27 January 2013 (UTC)
 * Reddit already invented a far better way to do this. If you vandalize Reddit, you see your posts but nobody else does. There are vandals and trolls who have been furiously typing for years, happily believing that they are posting all sorts of outrageous material to Reddit when actually everything they have posted since the their first vandalism got flagged can only be seen by them. For a while there they were experimenting with allowing vandals to see posts from other vandals so they can get into long, drawn-out flame wars that nobody else can see. This is unlikely to work on Wikipedia, but it is something to think about. --Guy Macon (talk) 21:56, 26 January 2013 (UTC)
 * Something to think about?? Something to think about???? Talk about high-risk, high-payoff -- I don't care what the risk is, the payoff would be incalculable! We need a Wikipedia Manhattan Project. The first challenge is discussion separation, by which the 0.71% of discussion (D-235, it's called) that moves things forward can be separated from the other 99.28%, which is D-238 (though D-238, aka "depleted discussion", does find use in the manufacture of consensus-piercing shots). Discussion (or better yet, editors themselves) can be placed in high-velocity centrifuges, and... [here the page was torn off]
 * 1) Support, this would create some problems but resolve many others. Obviously later we can fine-tune it based on experience. —Ynhockey (Talk) 00:57, 27 January 2013 (UTC)
 * 2) Suppport. Both "feedback" and "talk pages" serve the same purpose - to allow suggestions to improve the article. We don't need two separate systems. What we need is a way to make it easier for our readers to use the existing talk page. Mitch Ames (talk) 03:18, 27 January 2013 (UTC)
 * 3) Support. Another benefit would be that editors could engage in discussion about the suggestion. A concern would be that much of the pointless feedback would clutter up talkpages, but a specialized bot could probably deal with that. (For example, it might remove any one-word comments.) Ypnypn (talk) 05:01, 27 January 2013 (UTC)
 * 4) Support Some sort of filtering mechanism would be needed to screen out the gibberish and outright abuse, but this seems like a good approach if some form of the tool is retained. Nick-D (talk) 07:19, 27 January 2013 (UTC)
 * 5) It would be an improvement. It still doesn't prevent people from asking the same questions again and again, for example. --Enric Naval (talk) 12:57, 6 February 2013 (UTC)
 * 6) Support.  You could even maintain a bit of the current workflow: an initial review/flagging pass before posting things to the talk page, with all retained comments being posted to a talk-page section.  The poster (either IP or username) would get a message based on how the comment was handled.  Sure, Flow or LT may improve discussions; but until that happens all useful feedback should be organized in a single place per article.  – SJ<font style="color:#f90;"> +  23:34, 9 February 2013 (UTC)

This is a problematic UI for a number of reasons. 1. It navigates people away from the article they are commenting on, so that they cannot refer to it. 2. It forces people to fill out a subject line, write in wiki syntax, and sign their posts, which they didn't have to do before and may discourage some feedback. 3. Without AFT5's curation tools, we can't easily identify and highlight the most useful feedback among less useful feedback - feedback would remain in chronological order. Obvious bad feedback can be reverted (if a page has enough watchers) but there is a continuum of usefulness among the feedback we retain. Dcoetzee 21:13, 30 January 2013 (UTC)
 * Comments
 * Any proposed UI could easily eliminate the subject line field, auto-sign, and there's never been a requirement for using wiki syntax in a comment. It also wouldn't necessarily require navigating away from the page. Think about the current comments box from AFTv5. Now imagine that instead of it updating its own database, it posted to the talk page. Most of your stated problems aren't real. --MZMcBride (talk) 22:40, 30 January 2013 (UTC)
 * With additional development effort some of the talk page issues could be eliminated, but you still end up with an unorganised permanently chronologically ordered set of threads, many of which are prematurely archived, and users who leave comments don't know enough wiki syntax to reply to requests for feedback, etc. AFTv5 also has issues with attaching conversations to feedback at the moment, but talk pages have always been broken and I don't like the idea of expanding their use. Dcoetzee 02:35, 2 February 2013 (UTC)
 * Trying to maintain multiple different places to look for feedback allows confusion. There's no reason one couldn't keep all of the curation tools you like as part of a system that updates a feedback page which gets transcluded as a talk-page section.  Once we transition to talk-page sections each having its own ID (say, as a Flow post), you have an obvious way to convert each feedback post + thread into a flow post + thread.  – SJ<font style="color:#f90;"> +  23:56, 9 February 2013 (UTC)

This plan would have added 35 feedback messages to Talk:Hamas in the last few months. Do you think that drive-by accusations of bias would actually have helped the editors at that article? Do you think that one-off comments from readers who know nothing at all about Wikipedia's policies or practices should appear on talk pages that ArbCom has placed under discretionary sanctions?

YouTube has accumulated more than 3,500 feedback messages in the last year. Can anyone honestly believe that adding 3,500 new sections to the article's talk page would be anything except a serious annoyance to the people trying to get work done? World War II picked up 1,800 comments during that time. That would be five messages on its talk page each day.

I understand the desire to get helpful comments onto a page where editors will see them (a one-click copy-to-talk button would be desirable), but I believe there are good reasons not to flood the talk pages of contentious or popular articles with feedback from random readers. WhatamIdoing (talk) 21:24, 4 February 2013 (UTC)
 * I see people suggesting a single section that could be {in,trans}cluded on talk pages. Clearly, high-traffic and controversial pages have their own needs - in protection, editing policy, and feedback.  Perhaps like your idea of a "click to post" button, the simple, short feedback could be less visible until someone chooses to filter it up to the talkpage.  you could show at the top of the feedback section how many unreviewed comments there are; as well as show those that had been posted by a second person to the page directly. – SJ<font style="color:#f90;"> +  23:56, 9 February 2013 (UTC)

View by Diego Moya
There are valid concerns about the validity of the feedback provided by the tool, but those are surmountable problems and shouldn't stop deployment of the tool. Here is my evaluation and suggestions to improve the quality of feedback.

Diego (talk) 22:01, 25 January 2013 (UTC)
 * The feedback form is a huge improvement for readers; being located in the article space, ocuppying a big target area (much bigger than the "Discussion" link), and having a good affordance are all heavy advantages for newcomers over the Talk page.
 * Editors above complaint about nobody watching the feedback. The solution for that is to notify new feedback directly in the Watchlist, intermingled with other edits, for editors that opt-in to it; so regular watchers can provide instant "useful/useless" evaluation.
 * In order to make feedback actionable and improve usefulness, there should be a second field in the feedback form requesting to provide an (optional but encouraged) URL to a reliable source. This way, even if the feedback text is unclear, editors will be able to act on the linked article and find whatever information is relevant, and better infer the intent of the one who provided the feedback. This will also create a backlog of recommended sources for the topic, something that is more difficult to extract from comments in the talk page.
 * To avoid cluttering talk pages, the feedback should not be directly included there. It's better to keep it on a separate page, but with a unified notification format as explained above.


 * Users who endorse this view




 * Comments


 * I'm intrigued that you regard the problems as being surmountable, but saddened that your approach seems to be greater effort to divert more editors from improving Wikipedia to processing AFT feedback. I don't dispute that if we have to process that feedback and we are prepared to try and divert more editor time to it then your watchlist idea might help. But for someone like myself who has circa more pages on my watchlist than I can track all the edits of this would just mean my watchlist was less useful.  Ϣere Spiel  Chequers 18:37, 28 January 2013 (UTC)
 * "divert more editors...to processing AFT feedback". Agreed...in my opinion, either that will happen or AF will probably just be largely ignored by the editorial community.Shearonink (talk) 18:45, 28 January 2013 (UTC)
 * Well there is a third scenario, it becomes an unmanageable backlog, both a millstone round the neck of the community in terms of a heap of extra work created for the volunteers and also an ongoing embarrassment in that as a volunteer you are associating yourself with an organisation that operates a poorly moderated graffiti wall. I have the advantage here in that WereSpielChequers is not my real name, but there are editors who do so in their real name and thereby link their reputation to this site. I would not want it publicly known that I was an admin on a site with a poorly moderated graffiti wall, and if we do just ignore the problem till something happens that forces us to stop thenwhen the brown stuff hits the fan, as it inevitably will, I pity the community spokesman who has to go on television and defend it.  Ϣere  Spiel  Chequers 10:41, 4 February 2013 (UTC)
 * Having a substantial backlog would be impressive given the proposal for auto-archiving. There are many substantive arguments to turn off AFT5; "if the cacky hits the fan, and people know we're admins, what on earth will happen to our reputation?" is not one of them. Can people point me to admins whose real lives were negatively impacted following the Essjay controversy, or the various situations that led to the BLP policy? Okeyes (WMF) (talk) 10:48, 4 February 2013 (UTC)
 * WereSpielChequers: I'm sure there's some ghastly, scurrilous, libellous and BLP-infringing stuff on talk pages. What are we gonna do about that? We need to make sure that kind of junk gets seen and removed, regardless of whether it comes through AFT or through the edit button. The moral risk we take from not properly patrolling that is the same regardless of how it was added. The solution is as a community, we figure out a way of fixing it, rather than seeing it only as an AFT5 problem. —Tom Morris (talk) 10:54, 4 February 2013 (UTC)
 * Yes there is, and on user pages as well. I'm one of the editors who spends time fixing some of that. My objection to AFT is that the disbenefits are inevitably greater than the benefits. An x% increase in work for the regulars would be OK if accompanied by a commensurate increase in new regulars. But this whole development is predicated on creating extra work for the existing editors whilst exacerbating our longstanding problem of the shift from the SoFixIt culture of our golden years to the templating culture that has replaced it. We need to be looking at the various ways in which people identify work for others to do and see how much of that we can shift back to getting people to fix things themselves, we do not need a development that is designed to get a whole new peanut gallery of commentators generating work for the existing editors.  Ϣere Spiel  Chequers 15:06, 18 February 2013 (UTC)
 * Hmm, disbenefit is a new word to me. I've always heard/used detriment. --MZMcBride (talk) 06:40, 20 February 2013 (UTC)

Sorry, but the tool doesn't allow discussion of the merits of the link. In conspiracy articles you would get flooded with useless links, for example here, and you couldn't engage into discussion with the user. --Enric Naval (talk) 13:05, 6 February 2013 (UTC)

View by Bensin
This was a WMF pet project from the start. As far as I see it the tool is a cheat to trick readers into editing. It's purpose, to increase the number of editors, may be good, but the method is dishonest and manipulative. Let readers be readers when they want to be readers. Let them become editors when they choose it for themselves. Not because we trick them into it.

This WMF comment makes me itch: "thanks to everyone for your thoughtful comments about Article Feedback v5. You've given us much to think about, as we plan our next steps for this reader engagement tool."

Correct me if I'm wrong, but isn't most comments on this RfC mostly negative towards AFT? Isn't that perhaps something to think about? What would it take to actually scrap this project?

Salvio said "please kill it with fire." I say we take off and nuke it from orbit. It's the only way to be sure... --Bensin (talk) 18:37, 27 January 2013 (UTC)


 * Users who endorse this view


 * 1) The whole concept is too deeply flawed to be worth trying to salvage. The WMF would be better advised to put some of those resources into important developments like improving the software so that more conflicting edits are successfully resolved without edit conflicts. There is plenty of low hanging fruit out there if the WMF has IT developers available, no need to do something contentious, risky and believed by some to be flawed in concept.  Ϣere  Spiel  Chequers 14:55, 18 February 2013 (UTC)

Not quite sure you need your own view here. Endorsing Greg's view and commenting in Fabrice's view would have a greater impact, I think. --MZMcBride (talk) 21:36, 27 January 2013 (UTC)
 * Comments
 * Well, after all the office hours sessions I attended suggesting improvements on this tool, I took this liberty. Read other views. Didn't find one I subscribed to. Added my own. Simple as that. Whatever impact you attribute editors have on the fate of this tool, I don't share your optimism. --Bensin (talk) 00:40, 28 January 2013 (UTC)
 * Heh, fair enough. Though I can say that I value my time more than most things and I've invested a significant portion of it into this RFC. If I thought it would be for naught, I wouldn't have invested the time I have or asked my colleagues to invest their time. As it is, barring some radical and unforeseen events, every version of this tool will be disabled on the English Wikipedia by the end of February 2013. --MZMcBride (talk) 20:19, 28 January 2013 (UTC)

View from Canoe1967
Suspend it. I didn't read all of the above posts, but it seems that too many are negative. I feel the biggest issue is policing it with a watchdog system that is already over-taxed. I have answered questions on many un-protected pages from templates, to categories, to files, that have nothing to do with the page they are on. Readers seem to find a way to get messages noticed without the use of another system to patrol. I don't know if WMF had input from editors or not, but they may wish to suspend it all together as an experiment that needs work yet. We could find other methods to get new editors that may be more effective and less taxing on patrol systems.--Canoe1967 (talk) 19:37, 28 January 2013 (UTC)


 * Users who endorse this view




 * Comments

A question from jc37

 * One question - at what point does a reader become an editor? To clarify: Why is editing a blog an AFT5 to be considered not to be an editor, but editing a talk page is? And do you get a sense of how ridiculous that sounds? And better, how it sounds like another way to create an artificial barrier between people, and reinforcing another case of "us vs them"? - jc37 23:19, 25 January 2013 (UTC)

I don't think this question is difficult to answer:
 * Discussion
 * when you click the "Edit" (or "Edit this page") tab, you're directed to &action=edit;
 * when you click a section edit link, you're directed to &action=edit&section=X; and
 * when you click "+" (or "+comment" or "+section" or whatever it is now), you're directed to &action=edit&section=new.

All of these turn readers into editors.

Contrast this with posting feedback, which does not involve any edit action. :-) --MZMcBride (talk) 20:24, 28 January 2013 (UTC)
 * So the action needs to be labelled "edit" to make one an "editor"? meh. more arbitrary dividing lines. Let's use the word "contributor" and toss out all the complexity for the sake of complexity.


 * Incidentally, this reminds me of something else that's killing Wikipedia: the continued development of gadgets and gizmos which treat symptoms and not problems. Just as AFT does.  Rather than helping people learn how to edit, we're making everyone more and more act in isolation of each other. Werespielchequers above states it rightly that we're becoming a tag and ignore it project. Maybe if we outlawed bots and the like we'd see people better understand and develop that sense of fellowship of helping each other. Instead, we're developing the idea that everyone is an enemy, a potential reverter in Wikipedia the video game. But too many people are staunchly entrenched in their own personal corner of the ever increasing bureaucracies and policies and guidelines and style choice etc. There are maybe a dozen fundamental principles for Wikipedia. I wonder what would happen if some "power that be" (JW? the WMF?) deleted all of Wikipedia namespace and said we're starting over. If nothing else, it definitely would be healthy... - jc37 00:03, 29 January 2013 (UTC)

AF Feedback linkage broken from my Watchlist
Don't know if this page is quite the right place to put this, but since this is an RfC and being somewhat unable to access our Watchlisted AF might have some bearing on editors being able to truly weigh-in with their AF experiences... The linkage from my Watchlist to its Article Feedback seems to be broken, instead of a full URL I am getting "special:ArticleFeedbackv5Watchlist?ref=watchlist" instead. Thanks, Shearonink (talk) 20:33, 28 January 2013 (UTC)
 * Lawlz, confirmed. Reported as 44439. --MZMcBride (talk) 21:35, 28 January 2013 (UTC)
 * Thanks for reporting this bug, Shearonink and MZ. We were not able to reproduce it on our end on most platforms, but believe that it is caused on some platforms by a caching conflict -- where a new value has been updated through PHP, while the browser has not yet cleared JS cache. We expect this to be cleared up shortly. Thanks for your patience. :) Fabrice Florin (WMF) (talk) 22:17, 29 January 2013 (UTC)

View by Aspheric
2. I have found the tool essential while making changes to cataract. it allows editors when technical expertise but little insight to tailor the article to the needs and interests of the readership, which in this case includes patients. poorly written or complex sections are quickly identified as are important gaps. The tool keeps the article relevant.

I would like to see some way of auditing the output i.e how many were happy/unhappy, what the most common words in the comments were. Aspheric (talk) 02:29, 29 January 2013 (UTC)


 * Users who endorse this view



You're running about 50-50, with some of the "unhappy" people looking for Waterfall, and some saying they didn't find what they were after but that the article was great. This is also a good example of an editor using that feedback, but not marking any of it as being resolved. (I just processed a handful of simple ones.) WhatamIdoing (talk) 21:36, 4 February 2013 (UTC)
 * Comments

Question from Pete Forsyth
What is the theory of change for the Article Feedback Tool? What guides the evaluation of its success?

It seems to me that this page should contain the answer: meta:Research:Article feedback/Final quality assessment But what I see there is a great deal of detailed information, without a clear and compelling argument for why it should be interpreted in a certain way.

I would expect to see something of the form: "We will consider the AFT a success if it clearly results in ... We believe these changes will support the Wikimedia mission because ... We see these potential downsides, which we will measure by doing... We will compare the upsides and downsides according to this model..."

Without an explanation like that, the data reported on the page is pretty meaningless to me -- and, I'd imagine, to many others. It does seem to me that the quantity and quality of feedback are the main things being measured; but that does not account for questions about how the software design itself affects current readership and collaboration. Are we trying to measure stuff like that? If so, what's the basic approach, in plain English?

tl;dr: The theory behind measuring success is insufficiently articulated. To the extent I can discern the thinking, it appears to be laser-focused on some details without more broadly measuring the impact of the existence of the system itself. What is that theory, and is it the right one? -Pete (talk) 08:31, 29 January 2013 (UTC)


 * Update, with more specifics: I think I have found the piece that is closest to what I'm seeking, from Fabrice's December 2012 blog post. It's a step in the right direction, but hopefully only the first step of several:

This tool is converting readers into new editors

Article Feedback appears effective in getting new users to contribute to Wikipedia. For example, 2.7 percent of readers who post feedback go on to create a new account, after being invited to sign up. And 3 percent of these new users go on to edit articles within 24 hours from signing up. At this rate, we project several hundred thousand new registrations per year on the English Wikipedia — resulting in many new contributors, which we hope can help reverse the current editor decline.


 * If we want to interpret that as good news, that requires accepting a number of assumptions that I haven't seen called out or explored in a transparent way, or perhaps not at all. Any number of assumptions should be backed by a solid and compelling theory, before it is considered a valid premise. For instance:
 * Is this causation or mere correlation? It could be mere correlation; maybe only the people most likely to engage click on AFT to begin with.
 * Or, maybe worse -- maybe it's actually reducing continued engagement among that population. An alternate possibility: Of the people who click AFT, 5% would have instead left a comment on the talk page if AFT didn't exist, and 5% of those would have gone on to create an account. But AFT introduced complexity or a lack of feedback that reduced the number to 3% and 2.7% (as opposed to increasing it) within this already highly-motivated group.
 * If measurement of the AFT's impact is built on a flawed or fluid set of premises, it won't yield anything useful; garbage in, garbage out. What we need is not more number-crunching, but a careful consideration of why we consider those numbers significant to begin with. -Pete (talk) 01:32, 30 January 2013 (UTC)
 * Uh. We did that :). As linked above, we directly considered and studied the hypothesis that we were cannibalising existing or new anonymous contributors and found it baseless. Okeyes (WMF) (talk) 01:44, 30 January 2013 (UTC)
 * Okay, I'm glad to know that this has been considered. I've read through a great deal of information, sat through presentations, and had lots of conversations, and I'm not clear on that. In my view, this should be the very first thing somebody sees when they go looking for information about why the AFT is a good thing; but (as Risker pointed out prior to this RFC) it's not on the front page, and (as my own efforts have convinced me) it's also not possible to find even with a fair amount of clicking and reading. So, two questions remain:
 * If buy-in from the volunteer community was desired from the beginning, why hasn't this reasoning been on prominent display from the beginning? (I hope this doesn't come across as accusatory -- I think it's a mere mistake, but it does get frustrating because WMF has made this same mistake in other big technical projects).
 * Because we didn't have the results at the beginning; this is research that has been done during the project's design and deployment. When it was done, it was announced, release and thrown around. It's worth noting that lots of different editors have lots of different concerns; cannibalisation is one of them, yes, but quality of feedback is pretty high up on the list as well, as is volume, prominence of design, so on, so forth. We did a lot of intensive research on all of these things, and it's often hard to point at one piece of work and say 'this, this editors will care about above all else'. The main project page links through to the research hub, where all of our research is available, for precisely that reason. As you said, it's your view: a lot of other editors have different concerns with the project. It is impossible to give all possible perspectives equal light and time without overwhelming the documentation. Okeyes (WMF) (talk) 02:45, 30 January 2013 (UTC)
 * You're not understanding (maybe my fault). What I'm looking for is the kind of basic framework that would encompass possible objections, and establish a useful way of communicating about them. When WMF has submitted grant proposals for projects, they have typically had mandatory sections like "theory of change" (to use Hewlett's term) in which the reasoning behind the project is laid out; and the "measures of success" section is not just a bullet list of stuff you will measure, but a thorough evaluation of why measuring that stuff will yield useful information. At the proposal stage.
 * I'd say it's more important to be upfront and engage on that level with the community, than with a funder. WMF managed that really well with the Strategic Planning process, and with the TOU rewrite. The organization knows how to do it. I believe if it had been done in this case, things would look very different right now. For starters, we would be starting from a position in which a lot of people had bought into the experiment from the beginning, and felt invested in its success. -Pete (talk) 03:20, 30 January 2013 (UTC)
 * Sure; I thought you were asking about the research, not about the project in question. I don't deny things could have gone better (they always can) and we should probably have gone into more detail on WP:AFT5, but I'm confused by the argument that we simply didn't set out the reasoning behind the project and what it was hoping to achieve. We clearly stated the two goals (getting feedback to improve articles, and providing an on-boarding ramp for readers): the second line of the page is "The goal of this project is to engage readers to participate on Wikipedia, inviting them to give feedback on articles they read, and to become editors over time." As an aside, I have no idea who Hewlett is or what his "theory of change" term means; if we're talking about transparency and making it easy for editors to understand what's going on I suggest we try to avoid using inside baseball terminology :). Okeyes (WMF) (talk) 03:41, 30 January 2013 (UTC)
 * Ah, I see -- I think we can wrap this up, reply outdented below. -Pete (talk) 07:11, 30 January 2013 (UTC)
 * Forgetting the process stuff -- what is the reasoning? Can you post a link here? I've looked. -Pete (talk) 01:57, 30 January 2013 (UTC)
 * What do you mean 'the reasoning', sorry? For the project, or...? Okeyes (WMF) (talk) 02:45, 30 January 2013 (UTC)
 * I found what you were referring to, near the bottom of Risker's section -- this link. You noted that it's in "research-ese" at the time, which really goes to my main point about having a substantive dialogue with the community about the value of the project; but I am glad to know that the question has at least been asked. -Pete (talk) 03:20, 30 January 2013 (UTC)

OK, it's coming into focus, from the page you linked on WP:AFT5. Four paragraphs in is the following (emphasis is mine):

WMF projects a significant amount of feedback when the tool is fully deployed on the English Wikipedia. As a result, they believe that this new tool can enable a new form of participation for Wikipedia readers, most of whom do not currently edit the encyclopedia.

I think that's at the heart of everything going on here. To my ears, that is a radical notion -- so radical that maybe I've just been resistant to hearing it before. Maybe it has been said elsewhere, and I just didn't hear it for what it was.

Built into WMF's theory a definition of a new form of participation. In order to agree that the results are good, we would have to agree that an increase in that type of participation is a good thing, in itself.

For me, all the complexity just faded away -- I just fundamentally disagree with that premise. An increase in the kind of one-way communication that AFT provides is not on my list of things that Wikipedia needs. I don't think it will help, and a tool that is designed mainly to promote it is not a tool I think we need. -Pete (talk) 07:11, 30 January 2013 (UTC)
 * Then I'm sorry you disagree. But, for the record; I've spoken to over 200 users, held a pile of office hours sessions (3 back to back, in one day, for Australasia, North America and Europe), announced what we're doing on mailing lists, the village pump, the administrators' noticeboard, and distributed a fairly regular newsletter to anyone and everyone who signs up explaining what's going on. And that's just the consistent stuff. I am happy to accept we could've run things better (this is always the case. Hindsight is 20/20) but, yes; it's been stated that this is what we were trying to achieve. I can't speak for if you saw the statement before or didn't, but we advertised our actions very widely. Okeyes (WMF) (talk) 07:30, 30 January 2013 (UTC)
 * OK. I've absorbed huge amounts of that communication, going back to the first version of the tool, and throughout its evolution. I've given it a good faith effort. I know that you and all your colleagues have too. I regret that it's taken so long to sink in. I'm not inclined to reverse-engineer how that happened right now, but I'm happy to finally understand why this project has always been so confusing to me. Thanks for your engagement here and helping me put it together. -Pete (talk) 08:48, 30 January 2013 (UTC)

View by Isarra
For a disclaimer, I haven't really been following this thing and only read parts of it, so I apologise if any of this is redundant or inaccurate in light of what is above or elsewhere. Just calling things how I see them.

AFTv5 works. It aims to give readers another avenue through which to become involved - to show them that they can contribute and to encourage them to become editors as well, and it does this. At the same time, it also provides valuable feedback from some readers that experienced editors can then apply to articles, feedback that they may not otherwise get.

AFTv5 also doesn't work. It has opened a floodgate of spam, crap, and single-word messages, placing them in a hard-to-reach queue where they more often than not stagnate, with useful stuff getting lost to a morass of useless. It gives those who take the time to provide feedback the false expectation that someone will be able to use that feedback and respond to it, when indeed often nobody can, and even if they do, there is little way for one to respond to the feedback itself.

But much of this should be relatively simple to address, and as I understand it the folks working on this thing have been doing exactly that, due to release a new version soon. I don't know what all is changed in that, though, so I'm basing the following off the currently deployed implementation. Basically, these are things that come to mind when I look at it:
 * Much of the useless feedback is only a word or two. Changing the representation of the feedback form itself so it is clear the comment is optional would probably result in a lot less of that. And if it's not optional, it should be.
 * The call to action telling those leaving feedback that they can fix things themselves comes at the end. It should come at the beginning. If they find an article unhelpful, they should be directed to try to fix it themselves, and only if they cannot do that, to leave a comment. This would be much more in line with what was, from what I understand, the primary goal of recent iterations AFT: letting people know that they can edit and contribute. This way, while people could still leave feedback, they should be less likely to do so, thus resulting in less items for reviewers to sort through while still allowing for useful comments.
 * Feedback should appear directly on talkpages and watchlists so it is easier for editors to find and deal with, and also so they can discuss and respond to it. This is, unfortunately, not the most feasible with current implementations of talkpages and even watchlists, but with development of Echo and Flow should help alleviate this... in the next couple years. Meantime someone suggested just having it dump a new section directly on the talkpage, which would also work, though it wouldn't be so easily patrolled, which would be unfortunate.
 * The feedback form should make it clearer what readers are looking at - pages are not company websites, and this is not a contact form or place to add telephone numbers or names or other such stuff. This is an encyclopedia. We should be very clear on this on the form itself.
 * AFTv5 needs better antispam and antinonsense filters - not necessarily just abusefilter filters, but... whatever it is it needs. Hopefully this has already been resolved with the next iteration because frankly I've got nothing.
 * People don't know what kind of feedback is even meaningful - they will suggest things without being specific enough for someone to act on it, or neglect to suggest things because they do not even realise they are a possibility. Checkboxes with some common issues could help alleviate this and also provide more manageable feedback - have options like needing proofreading or missing a lead or whatever is common, and then users could get data on which articles people said needed proofreading without even having to read feedback comments themselves.

That's what comes to mind when I look at it, anyway. But whatever the case I'd say the thing is worth giving a chance - it just needs more work before full deployment, but if the thing is scalable on the backend, it should be relatively trivial to take it the rest of the way on the front, where it affects us as editors and users. -— Isarra ༆ 08:36, 29 January 2013 (UTC)
 * Users who endorse this view
 * 1) A nice well-rounded opinion. In particular, we should improve the clarity of the feedback form (purpose, etc) and see if anything changes. — This, that, and the other (talk) 11:00, 31 January 2013 (UTC)
 * 2) Yeah, good ideas. &bull; Jesse V.(talk) 19:44, 4 February 2013 (UTC)
 * 3) --Guy Macon (talk) 05:12, 5 February 2013 (UTC)
 * 4) Not sure about dumping the comments on the talk page, yes to the rest of it. -- phoebe / (talk to me) 05:20, 6 February 2013 (UTC)
 * 5) "If they find an article unhelpful, they should be directed to try to fix it themselves, and only if they cannot do that, to leave a comment." Indeed, that's the whole point of "anyone can edit" and WP:BOLD. --Enric Naval (talk) 13:10, 6 February 2013 (UTC)

I strongly agree with "It has opened a floodgate of spam, crap, and single-word messages, placing them in a hard-to-reach queue". Most of those are solvable, but all available evidence suggests that antispam filters will not work. Millions of dollars have been spent on this problem by dozens of companies, and nobody has ever created a spam filter that actually works. If you create a filter that stops 99.99% of the spam, the spammers analyze the 00.01% that gets through, and a week later the filter becomes 10% effective because the spammers adapt. It is an arms race that Wikipedia cannot win. So why are our articles and talk pages not full of spam? Humans. Humans are the only method that works. --Guy Macon (talk) 05:12, 5 February 2013 (UTC)
 * Comments
 * Is there any evidence people-who-are-spammers have targeted AFT5? Okeyes (WMF) (talk) 07:08, 5 February 2013 (UTC)
 * A small amount (#269779 #844580...), but not much so far. I support what the WMF is doing about this (make the spam filters as good as possible and let humans clean up the rest). It is the oft-expressed opinion that we don't need the humans that I disagree with.
 * Spam often starts slow on new systems Twitter and cell phone text messaging are well known examples of someone thinking that spam won't be a problem based on early results, only to find the system clogged with spam later.


 * The first telegram in the United States was sent in 1838 and the US west and east coasts connected in 1861.


 * The first mass unsolicited commercial telegram was sent in 1864.


 * The first ARPANET email was sent in 1971.


 * The first email spam was sent in 1988 ("Make Money Fast" spam).


 * The first USENET posts were sent in 1980.


 * The first USENET spam was sent in 1994 ("Green Card" spam).


 * --Guy Macon (talk) 08:49, 5 February 2013 (UTC)

View by Andy Dingley
Any "open question system", where the commentator is allowed a free-text box to type into, is just an invitation for our regular baboon-flingers. This effort has not proven to be any exception.


 * Any comment open to anon editors should be restricted to radio buttons or sliders. These might be more open than previously: it might include specific buttons for "Needs images" or "I couldn't understand it".


 * If there is going to be free-text entry, limit it to editors with traceable identities, i.e. logged in accounts.

A comment on the previous system too - it had sliders for readability and trustworthiness, but nothing for basic accuracy.

Andy Dingley (talk) 12:37, 29 January 2013 (UTC)


 * Users who endorse this view


 * 1) I like this--though I think check boxes, not sliders. This eliminates the need to use edit filters, it can cover all common situations--and even suggest them, and build categories with those tagged in a particular way. Anyone who wants to say something unprovided for can be directed to the talk page, where  everything needing discussion will be in one place to be dealt with.  My guess is that those who really do have something to say will not be discouraged by having to go to the talk page, while many fewer people who might contribute something unhelpful will go to the trouble DGG ( talk ) 03:06, 30 January 2013 (UTC)


 * Comments

I'm not convinced that sliders are much use either for general public evaluation of encyclopaedia articles, far too open to people not liking controversial people rather than the articles on them. Like works fine for photos on Facebook that one shares with Facebook friends, but it really isn't a sensible way to run an encyclopaedia. Yes we could have separate buttons for whether people like the article or the subject of that article, but I'd have BLP concerns with doing so and it would be seen by some as violating our neutrality. It would be awkward running this in a major election, especially a US one and as for religions, flags and religious books, the opportunities AFT gives our trolls are scary. As for restricting this to logged in editors, yes it might help AFT function in ways that don't unduly burden the community, but it would make this less of a reader interaction tool than the talkpage as very few of those are semiprotected.  Ϣere Spiel  Chequers 18:26, 29 January 2013 (UTC)

AFTv4 had "star ratings". The resulting feedback was completely useless. We need specific information about how to improve articles, referring to specific portions of articles, not just common suggested improvements. The fact that open comments require moderation is not new or surprising. Dcoetzee 21:09, 30 January 2013 (UTC)


 * Is the purpose of feedback to identify which articles need work, or why particular articles need work? As an editor of a topic, I'm also a knowledgeable reader of that topic. Why isn't a problem for me, I can see that for myself. OTOH, which, articles I should focus on is something where its visibility is improved by the statistical view across many. I gain that much from v4.
 * I've read a lot of feedback. I've seen a fair bit of useful feedback that said "More images" on a topic where images are hard to obtain (and I was looking already). I've seen a little where feedback raised an issue I'd not otherwise looked at and then did go on to address. Almost entirely though, the ability to enter free text is just the opportunity to enter the text "poo". ATFv5 isn't giving me such useful feedback as even v4 did. Andy Dingley (talk) 22:51, 30 January 2013 (UTC)


 * The most useful feedback from AFTv5 that I've acted on usually pointed out very specific issues with articles. Generally these are issues that were not apparent to me upon reviewing the article, because the reader knew the subject area better than I did. Dcoetzee 02:32, 2 February 2013 (UTC)

Comments by Shearonink
Something new I just ran into today was where multiple users from a school IP were using AF at a historical biography article to chat with each other. Shearonink (talk) 22:55, 30 January 2013 (UTC)


 * JWZ's law. Any software system expands until it can also read email (or these days, chat). Andy Dingley (talk) 04:04, 31 January 2013 (UTC)

And here's another one that is new to me....found a single IP that over the past few days has been posting AF to different film articles that consists of nothing but actors' names and/or film titles with question marks. Shearonink (talk) 23:58, 31 January 2013 (UTC)
 * They are still going strong. I know that this is not quite within the scope of AF, but what are we supposed to do when someone is using AFT in this fashion?  Is it actionable, or do we just stand back and let them post away?  Shearonink (talk) 01:18, 3 February 2013 (UTC)
 * Well, presumably you could block them for misuse, but frankly with the way this RfC is going it may be a waste of effort on the admin's part :/. Okeyes (WMF) (talk) 01:46, 3 February 2013 (UTC)
 * I figured that they are at least learning to post and don't seem to be doing any real harm, but I wasn't sure if someone just keeps doing this without adding to the project, if post-fodder eats up meaningful server space without any meaningful return.Shearonink (talk) 02:20, 3 February 2013 (UTC)

Comments by Spasemunki/Clay Collier
Four questions were asked:

1) What should the scope of the article feedback tool be? I see no reason why the tool should not be included on all articles. Articles that have little user activity are in the most need of constructive feedback in many cases. The feedback tool creates a 'to do' list that can wait dormant until an interested editor comes along.

2) Are there sufficient resources to moderate and respond to all of the feedback? Certainly not, but this is hardly an objection. Why should suggestions for article improvement be moderated? Editors are perfectly capable of determining for themselves which suggestions they will follow up on, and which can be ignored. The 'useful' vote and the flag as abuse provide sufficient opportunity for incremental feedback and response. Everything does not need to be done at once.

3) How will abuse filters handle articles in which the subject's title contains a disallowed word (e.g., Blue-footed Booby + "boob")? I am suspicious of the utility of automated filters for an encyclopedia that is itself not content filtered. Who cares if the word boob appears in a feedback request? Wikipedia is not censored, useless feedback will be ignored. Trolls only keep coming back if you feed them.

4) Will the tool continue to only be a(n expanded) box or will it go more minimal? Will it go "above the fold" (e.g., File:Article-link-to-feedback.png)? The size and position of the feedback box seem perfectly well-suited to me- unobtrusively asking the reader at the end of the article whether or not they found what they are looking for. If the feedback box is at the top, lazy people will click 'no' if the info they want isn't in the lead paragraph. If there is too much feedback requested, it will discourage asking for the kind of incremental, small improvements that Wikipedia is best at providing.

We are planning for a future without an expiration date. --Spasemunki (talk) 03:59, 3 February 2013 (UTC)


 * Regarding your question "Why should suggestions for article improvement be moderated?" and your related comment that feedback can "wait dormant until an interested editor comes along", some material is prohibited on Wikipedia. This includes libel, personal details, violations of copyright, spam, medical/legal advice, and material that violates our living persons or banning/sockpuppetry policies. We cannot let prohibited material lie dormant until an interested editor comes along. And the current state-of-the-art in filtering technology is such that the best filters keep out 90% to 95% of the undesired material -- well worth doing, but not good enough to try to get along without human moderators. That's why the WMF has designed a system that uses human moderators --Guy Macon (talk) 11:52, 4 February 2013 (UTC)


 * Re your answers.
 * 1) We can't leave answers dormant, some need oversight. I found a "call for a good time" with phone number that had been up for several months.  It is clearly inappropriate for those to remain up a day, much less for two months.
 * 2) Editors are the moderators, and there are simply better things for editors to do.
 * 3) AFT is feeding the trolls. It is an unintended consequence, but it is what is happening.
 * 4) It is more likely, based on this RfC, that the tool will be gone completely.  GregJackP   Boomer!   12:43, 4 February 2013 (UTC)
 * Comment: GregJackP: Ditto, echo, like. Absolutely agree. I've already supported another opinion above, but I wish to cheer here. <font color="#595454">Fylbecatulous  <font color="#DB7093">talk  13:04, 6 February 2013 (UTC)

Comments by Phoebe

 * Small comment: It's not so hard to get through a pile of AFT comments -- easier than reviewing articles or edits -- but I, for one, forget to do it. It seems like it would help to have more tools for review, as someone brought up, and especially more visibility of it as a to-do task. Having MzM's suggestion of AFT5 as an optional add-in for those authors who want it seems like a nice bridging step regardless of whether we go to full deployment, since I think it would help editors get used to the workflow. Are there usability improvements we could make that would lead to more helpful comments?
 * edit after skimming the rest of the RFC: maybe the new version Fabrice alludes to above will take care of the review tools; seems like it will help, anyway. -- phoebe / (talk to me) 02:01, 6 February 2013 (UTC)


 * In general: I agree with the principle that reader feedback is valuable, but also that the feedback needs to be better integrated, and better integrated for the editing timescale of "eventualist" improvements -- one thing that's nice about AFT5 is any helpful comments do reflect the state of the article at the time, rather than the star-rating version of AFT (where you can't tell if someone thought it was poorly written 6 months ago or today, which I find totally irritating).


 * I think as a principle, we should aim to build tools that make it easier to give feedback on all articles (just like all articles are editable), which means that I support a goal of full deployment for some kind of feedback tool. But one size may not fit all; long-tail and front-and-center articles are not the same, and we might have to make tools differently for these cases. An article that gets 100 views a week is very different from one that gets 10,000, and I would much rather get full comments on the long-tail articles than on the celeb-of-the-week. -- phoebe / (talk to me) 01:45, 6 February 2013 (UTC)
 * Hi. I think one of the frustrating parts here is that these variables are known or knowable, but aren't being utilized. That is, it would be trivial to insert some code that says "if article_views_per_month > 1000 { don't show the feedback tool }" or "if recent_edits < 50 { show the feedback tool }" or a million other combinations and variables. And, as you and I have both said, a compromise measure, such as making the tool opt-in (i.e., making its use discretionary) for now is a workable solution. But the Wikimedia Foundation doesn't seem interested in compromise, to be honest. A lot of this discussion has felt like "we'll do it my way or we're taking the ball and going home." And stomping off home, ball in hand, they shall go, it seems, from the way the discussion is headed. I think it's unfortunate. I was re-reading Tom Morris' view after seeing you'd endorsed it. He makes some very strong points about the power of a scalable, sustainable, and sane feedback system. It's difficult to disagree with the ideal he puts forward, but as I see it, we're simply not there (yet) in reality. --MZMcBride (talk) 05:11, 6 February 2013 (UTC)
 * Well, while my intuition is that long-tail comments are more useful, simply because -- per Andrew West's excellent article in the signpost this week on popular articles -- most popular articles are popular for the same reasons, maybe that's not true; maybe all we can conclude is that popular articles will give us more comments. Anyway, I think it's worth some thought and maybe some stats. Regardless, I do certainly agree with Tom's vision of a sustainable feedback system. I don't have the same dismay that others commenting above seem to have about the nature of the feedback. Look, we -- those of us commenting on this RFC, right here right now -- are collectively the creme de la creme of Wikipedia editors: the people who edit a lot, understand it, do maintenance tasks, and care enough about the project to talk about it. And even we sometimes have a hard time figuring out how to improve articles. Is it any shock that casual readers do too?


 * Anyway, thank you for starting this RFC, though I have to say, I don't see the same tone in the discussion that you do; it seems like everyone has pretty polite and professional, including staff responding to calls to toss out (or burn with fire) a whole lot of work and what has been a concerted, good faith effort to try a large-scale effort in engaging our billion readers. -- phoebe / (talk to me) 05:43, 6 February 2013 (UTC)
 * Max; not at all :). More "we'll do it in the way we have resources to implement; if you would prefer to positively state 'we should take our ball and go home', we will acquiesce". I'm not going to lie: I think that it would be awesome if we could spend more time working on this to come up with something enwiki is more willing to consider a positive change - give some time for users to write third-party tools to semi-automate patrolling, build out some equivalent of ClueBot that can hit comments, look into more efficient ways of filtering - but the fact of the matter is that our resources are always going to be less than our to-do list would ideally like. I think it would be great to have AFT5 improved on and deployed; I think when that comes at the cost of having to push back on making a talkpage system that works, or an inter-wiki notifications system, or any of the other things that we have to do, the return for such an investment is a bit more questionable. Okeyes (WMF) (talk) 10:22, 6 February 2013 (UTC)


 * Yeah, re: resources. Just in case it wasn't clear from the above: I for one find the comments that the tool is useful (from Schrocat, Charles Matthews and others above), very very compelling -- compelling enough to keep the tool around. And way more compelling than saying "but there's spam!" This community knows a thing or two about dealing with spam :) -- phoebe / (talk to me) 16:12, 6 February 2013 (UTC)
 * Yes, too bad that both are made useless by AFT not using wiki pages... ;-) --Nemo 18:05, 7 February 2013 (UTC)
 * If editors are finding it useful, it's not useless! -- phoebe / (talk to me) 22:00, 27 February 2013 (UTC)
 * Lets say there is a village in the desert with very limited water supply. Along comes a visitor, lets call her Aftes, bearing a cup of precious water to the village. She proceeds to share the water with the village - and then drinks 5 cups of water and leaves. She has brought water into the village, therefore she cannot be considered "useless" in addressing the village's water supply issues? -- TRPoD aka The Red Pen of Doom  00:10, 28 February 2013 (UTC)

Compromise proposal by Noleander
I think a good compromise would be to keep the AFT5 user-interface (easy text entry at the bottom of each article) but send the feedback to the article's talk page (or a subpage of the talk page). That way, we'd still get the valuable feedback (apparently between 5% and 40% of the feedback is useful) but we would not be inventing a new system that parallels the talk pages. I can see a couple of ways to do this: If we think that 90% of the feedback will be bogus, then option B may be better. We could either: (B1) require editors to opt-in to get watchlist notifications of additions to the Feedback subpage (not many editors would see the feedback); or (B2) make notification the default, and require editors to opt-out (more editors would see the feedback but many of them might be annoyed ... until they opt-out). I realize that this suggestion was hinted at in some proposals above (e.g. Srleffler and Legoktm) so I apologize for any redundancy. Personally, I think that customer (i.e. reader) feedback is incredibly important; even if only 1 out of 20 comments are useful, the ATF tool is still worthwhile, so I'd like to see its defects remedied rather than have it eliminated altogether. --Noleander (talk) 18:49, 7 February 2013 (UTC)
 * Option A:
 * 1) Send the feedback to the article's talk page (as a new section titled "Feedback from blah blah")
 * Option B:
 * 1) Send the feedback to a dedicated subpage under the article's talk page, such as Talk:ArticleName/Feedback
 * 2) Add a preference for editors to opt-in (or opt-out) so they can choose whether or not to get notified of additions to the feedback subpage (of pages already in their watchlist)


 * If you ask me, article feedback holds great potential, as it could allow users to be able to voice their opinions on how the article is written, and point out errors in the article itself, with the addition of being able to draw attention to stubs or articles without any pictures of a person or citations or anything of that sort, so I believe that you are correct in your idea, Noleander. signing for User:173.58.94.199


 * Perhaps feedback that is marked as unhelpful could not be transferred to talk? Any blank or duplicate feedback also not transferred? Or, perhaps the 'talk' tab could just be renamed 'feedback' or 'comment on this article'? Abductive  (reasoning) 20:15, 8 February 2013 (UTC)


 * Users who endorse this view


 * 1) Strong Support. (I like B better, but both are good) This would address all of my concerns. It is important that I be able revert vandalism, and it is important that other editors be able to see what I did in the page history and undo it if needed. This proposal accomplishes all of that. --Guy Macon (talk) 21:30, 8 February 2013 (UTC)
 * 2) Support B, oppose A.  Similar to 'Plan D' section of talk page that has discussion as well.--Canoe1967 (talk) 21:34, 8 February 2013 (UTC)
 * 3) Support (Option A). Perhaps we should consider not calling it "feedback" at all, but "suggestions to improve the article". Do we really want unactionable feedback such as "good" or "bad" - with no suggestions as to how to improve the article? Or do we only want actionable suggestions. Is the purpose of the tool to measure what readers think about the article (good, bad or indifferent), or is to improve the article? Mitch Ames (talk) 06:37, 9 February 2013 (UTC)
 * A term like 'input' may work better?--Canoe1967 (talk) 17:06, 9 February 2013 (UTC)
 * 1) Support B, oppose A.  There are articles (see my comment below) where this would be unuseful.  I prefer a dedicated subpage, and perhaps any substantial discussions should be moved to talk.--Wehwalt (talk) 22:42, 10 February 2013 (UTC)
 * 2) Support B, oppose A [from the proposer] - The more I think about it, the more B seems superior to A.  If A were proposed to the community, it would draw a large amount of opposition (for reasons similar to those given against the current incarnation of AFT5).   Option B seems to strike a good balance between (i) not distracting busy editors; and (ii) gathering input from WP's customers. --Noleander (talk) 22:53, 10 February 2013 (UTC)
 * 3) Support but please shrink the tool to a "feedback" button on the banner. And please junk the ratings subsystem. --Jorge Stolfi (talk) 06:03, 15 February 2013 (UTC) P.S. The feedback section must be editable like the rest of the talk page.
 * 4) Support it belongs on the talk page, not in wherever land that you can't respond to.  How am I meant to let the IP (who is probably a dynamic IP, so replying on the user:talk isn't going to work) about what redlinks are, in response to this feedback. If it was automatically added to the talk page in a new section (at the bottom, not above the banners), then I'd happilly respond and explain?  Am I missing something, or can I not give an answer to feedback?  We don't need two separate "feedback" related sections.  Talk is enough. The-Pope (talk) 07:19, 17 February 2013 (UTC)
 * Comments

Comments by KConWiki
I think that the feedback tool has potential for helping us improve our wiki, especially assuming the idea of all bugs being shallow with enough eyes. However, I think that the box is aesthetically displeasing, and leaves a big ugly chunk of white space in the bottom right corner of each page. What do we think about making it shorter and wider, so that it is the same width as a template? KConWiki (talk) 22:21, 10 February 2013 (UTC)

Comments by Wehwalt
I like the feedback tool, but am concerned about the difficulty in communicating with the commenters. I'd like some way for them to be able to receive a reply. I am not terribly happy about comments being posted on talk page; the sheer volume of crappy comments I've gotten on John A. Macdonald convinces me that's a bad idea. I want to be able to reply, that's the key thing.--Wehwalt (talk) 22:40, 10 February 2013 (UTC)
 * I like this comment. Abyssal (talk) 16:07, 11 February 2013 (UTC)
 * I find the two word comments from unregistered users mostly unhelpful. There is no opportunity for a dialoge as they are unlikely to have a fixed ip address. I would much rather effort went into:
 * -- stopping unregistered users vandalising articles, I have recently reverted vandalism that has been posted unchallenged for months devaluing the articles and wp in general
 * -- curbing uncited edits which can take ages to verify especially when they are one digit chages to numbers or stats
 * Keomike (talk) 01:08, 17 February 2013 (UTC)

To talk and to reply
I just a few minutes ago noticed the "Feedback from my watched pages" link at the top of my watchlist, and I think it's a useful tool. Suggestions that the comments be sent to Talk, or that establishing mutual communication with feedbackers is difficult, are missing the point. The new tool is for helping us understand the ignorant, the uncommunicative, the unWikian, which is what most readers are. The good old talk page is for helping us talk to ourselves, and serves splendidly for that purpose. Jim.henderson (talk) 02:50, 12 February 2013 (UTC)


 * You said it. Hindustanilanguage (talk) 16:03, 15 February 2013 (UTC).

View by Fiddle Faddle
The article feedback scheme implies a level of customer service. Wikipedia has no customer service. It is a purely self service organ. The Feedback service's existence implies that someone is actually going to care enough to do something with the feedback and is going to make some sort of reply. We are not the world's customer service organisation for its ignorance. Look, for example at this set of feedback None of them even grasped what the article was about! Who is about to handle this type of alleged feedback? All deserves a reply? Fiddle Faddle (talk) 12:32, 16 February 2013 (UTC)

Comment by an IP
I think that the feedback tool is a great idea + a very good method of accessing a proportion of the readers who do not normally edit. My suggestion after browsing a lot of misplaced comments would be to allow experienced users to move feedback entries from 1 page to another, e.g. a large proportion of the suggestions at Help:Using talk pages belong on other articles. -109.148.25.255 (talk) 17:26, 16 February 2013 (UTC)

Comment/view by Stevage
As I understand it, the current model is: 1. Casual reader provides feedback 2. Feedback goes to central pool, forming a backlog. 3. A small dedicated team of feedback-responders process the backlog as fast as possible, getting burnt out.

Steps 2 and 3 don't make any sense to me. Why not this: 1. Casual reader provides feedback. 2. Feedback goes to a list of comments on the article (implemented perhaps as a section on the talk page, or even a subpage: Talk: /Feedback) 3. Editors interested in the article peruse the comments at their leisure, taking action or not at their discretion.

Most of the feedback I've seen is aspirational in nature ("it would be good if there was a picture of ..."). That's not urgent. It's just something to jot down on the todo list, and do something about one of these days. Stevage 11:04, 17 February 2013 (UTC)


 * This is slightly off-topic, but may be worth pointing out that in many articles there is no stable place for a "to do" list. This is because talk page threads are archived off into oblivion after a certain period, whether or not they have been actioned. Over time I have made qute a few small but worthwhile suggestions that are now effectively lost forever. 86.160.219.18 (talk) 21:19, 17 February 2013 (UTC)
 * That seems pretty easily solvable with a transcluded subpage? Stevage 23:25, 18 February 2013 (UTC)

Comment by TheRedPenOfDoom
If the decision is to keep the tool, it should not be put on new and controversial articles. Take a look at this train wreck --  TRPoD aka The Red Pen of Doom  23:16, 17 February 2013 (UTC)
 * Unless we see a dramatic shift in the next few days, ArticleFeedbackv4 and ArticleFeedbackv5 will be disabled entirely here in short order. --MZMcBride (talk) 13:42, 18 February 2013 (UTC)

View by Ukrained2012
First of all, TLDR all above)

I don't know... Looks nice to me, although certainly upopular for now (underprompoted?) Why not keeping if the project doesn't consume too much money from Mr. Wales) Happy edits, Ukrained2012 (talk) 04:59, 18 February 2013 (UTC)
 * The issue is not the financial cost of running it, though some of us would prefer that the financial cost of developing it had been invested in higher priority and less contentious things like reducing edit conflicts. The issue is the cost in volunteer time involved in this and how the AFT exacerbates the community's main problem - the drift from the SoFixIt culture of our heyday to the current "SoTemplateItForOthersToFix" culture of template bombing.  Ϣere Spiel  Chequers 19:27, 24 February 2013 (UTC)

Some comments by Soroosh129
I think feedback isn't welcomed on Wikipedia. First of all, this isn't a site like codeproject. Pages don't have any author or admin. It is changed by any user, all changes are welcomed. Why a user should waste time giving a feedback? Because he/she is too lazy to contribute? then his/her opinion doesn't matter. It is a web site that everyone should participate on the content of the article. What a feedback gives? It conflict to a certain level with the foundation of Wikipedia. Another reason that comes to mind for using feedback is that changing the actual article is hard, it shouldn't. It should be as easy as giving a feedback. also opinions doesn't matter to a certain level, the articles should be neutral.
 * I agree; it should be a lot easier. That's why we're working on a WYSIWYG editor. But until that's live, and even afterwards, making direct changes requires quite a substantial amount of effort. Okeyes (WMF) (talk) 12:19, 20 February 2013 (UTC)

Comments by SoCalSuperEagle
I believe that an article feedback system can be very useful if implemented correctly. Unfortunately, the current implementation is clearly not acceptable. As of right now, the ArticleFeedbackv5 submission interface simply asks readers if they found what they were looking for and asks for suggestions if any. So far, there have been far too many suggestions that are too vague to act on, and that certainly doesn't include all the blank, irrelevant, or other useless feedback. If we are to continue accepting article feedback, then the feedback submission interface will need a significant overhaul. Here's what I suggest: -- SoCalSuperEagle ( talk ) 19:57, 19 February 2013 (UTC)
 * The feedback box that appears below an article should ask readers if they were satisfied with the existing content and appearance of the article.
 * If a reader answers "no" and if the article's protection settings would not prevent the reader from editing the article, then the reader should be asked if he/she would like to make any necessary changes to the article. If the reader answers "yes" to the second question, then the article's edit window should automatically open and invite the reader to be bold.
 * If the article is protected from editing or if the reader declines the opportunity to edit for any reason, then the reader should be asked what type of change (addition, removal, or alteration of content and/or formatting) needs to be made to the article. After the reader selects what type of action should be taken, the reader should then be asked to describe the specific addition, removal, or alteration that needs to be made and the rationale for the action. This feedback would then be posted to a special subpage that's associated with the article for easy access.
 * The feedback interface should clearly inform readers who choose to submit feedback comments that vague suggestions may be considered unactionable, that the feedback system is not a chatroom, and that the WP:BLP, WP:Copyrights, and WP:NPA policies are applicable since feedback is publicly viewable.
 * How would we know if the reader had 'declined the opportunity to edit'? And it seems a bit unfair to be asking them to answer a question and, in response, open the edit interface. It's unexpected behaviour that smacks of getting them in under false pretenses. Okeyes (WMF) (talk) 12:15, 20 February 2013 (UTC)

Idea by Ypnypn
If in the end we decide to keep this, maybe it should automatically be disabled if there are more than 20 unreviewed pieces of feedback. This will keep the pile on any given article to a reasonable size. -- Ypnypn (talk) 20:04, 19 February 2013 (UTC)
 * It seems very unlikely that AFTv4 or AFTv5 will be kept here at this point, but I think this is a innovative idea that's worth exploring in future iterations of feedback tools. --MZMcBride (talk) 06:46, 20 February 2013 (UTC)

Refactored TODO list, by IP
Feedback and discussion are 2 different things. We should care about their difference because we need both of them. The assumption that feedback would be a substitute for discussion is just as wrong as the assumption that discussion would be a substitute for feedback. We should divide our attention over them equally favoring neither.

The misconception is that the feedback tool should only be for new users. This makes the feedback not worth reading. Feedback should refer to all of the stuff that comes from editors not necessarily planning to help write the article.

My magic trick would be to make a todo list that is heavily re-factored. I imagine to get to Featured Article status one would have to at least one time make a list of items that need fixing. People unfamiliar with wikipedia can post their feedback, it will end up at the end of this page where it will scream for refactoring. The newbie can tell his life story and how wonderful it is to have images with articles. Oh, bla bla bla, how unprofessional it looks not to have images with the article and that his brother always puts images with the articles he writes on wikipedia. ''GTG mum calls for dinner! bye!''

After I refactor his post it would say: * need images.

Make it an item worthy of todolisting. The new guy can see exactly how wikipedia works. He visits the next article and types "needs pictures". This is his graduation, he now knows everything he needs to be a productive wikipedian. If we want him to get more involved we have to put images with the articles the way he told us to. :-) 84.106.26.81 (talk) 14:00, 20 February 2013 (UTC)
 * A bot can post feedback if there are no images and humans can tell us where even further illustration would be helpful contextually.84.106.26.81 (talk) 14:48, 20 February 2013 (UTC)

Comment from BarkingFish
You're probably thinking "oh god, here it comes." :) But, no. The Feedback system is a very useful tool - it has its flaws, like most things on WP - people leaving f-bombs in feedback posts, or posting garbage in it - but it's essentially bloody useful, and my understanding is, we're leaning towards sinking it.  Why in the name of Mike would you do this?  Aren't we all about giving readers the chance to tell us what they think? Isn't that part of the idea of what we provide?  What the system needs is proper control, and more people to moderate it when abuse is reported in posts, so it's kept clear of crud.  <font color="red" size="2" face="Ubuntu">Fish <font color="blue" size="2" face="Ubuntu">Barking?  23:11, 20 February 2013 (UTC)
 * Agreed. -- phoebe / (talk to me) 21:59, 27 February 2013 (UTC)

response
While getting and responding to "feedback" is a central tenant of process improvement, the feedback that the feedback tool has gotten is that in its current form and functioning the feedback tool does not help the encyclopedia improve. -- TRPoD aka The Red Pen of Doom  23:23, 20 February 2013 (UTC)
 * I believe that the Tool has shown a yawning chasm between what Wikipedia is and what it can be for the average reader. Something has to be done to provide the casual reader who wants to communicate with the unseen editors to improve articles. Presently they really don't see the Talk tab for what it is. I hope that the outcome of this RfC is a retooling, not an outright elimination, of the Tool. Abductive  (reasoning) 23:31, 20 February 2013 (UTC)
 * This presupposes that there are editors willing and able to respond to readers' requests for improvement. For large numbers of Wikipedia articles this is not true. 86.181.174.63 (talk) 23:45, 26 February 2013 (UTC)

I like it
http://en.wikipedia.org/wiki/Talk:Hh_antigen_system at the top of this talk page is a link to feedback. There was some left and I referred to it in my argument to keep some info. that was ultimately deleted.(feedback in nutshell-they wanted more info. on interesting topic) Even-though the feedback is not currently being considered by other editors in my opinion-I think that it deserves a longer chance.24.0.133.234 (talk) 16:25, 21 February 2013 (UTC)