Template talk:Article history/Deprecation

It strikes me that ArticleHistory, when used correctly, deprecates a lot of templates: And probably a fair few others. I'm not at all inclined to miss them - they're cumbersome, take up a huge quantity of space at the top of talk pages (the existence of skiptotoc indicates we put too much on the top of talk pages) and aren't nearly as easily machine-readable as ArticleHistory. Given that we now have a robust and compact way of listing all these article milestones, doesn't it make sense to deprecate all these templates? I suspect there are three schools of thought: Of course I have no doubt someone will find an alternative viewpoint. Note that FAC, FAR, GAN, etc, are not on the list, as they cannot be incorporated into ArticleHistory. Discussion, please. Happy‑melon 12:30, 19 December 2007 (UTC)
 * 1) These templates should be deprecated, a bot should convert all instances of the templates into ArticleHistory, and the templates should then be deleted.
 * 2) These templates should be deprecated and a bot should convert all instances of the templates into ArticleHistory. However, they should not be deleted.  New instances of the templates should be converted by the same bot.
 * 3) The templates should not be deprecated.
 * Note, some previous comments can be found here. Geometry guy 12:58, 22 December 2007 (UTC)

Comments

 * I'm definitely in favor of depcrating the old templates. But that simply raises a very important question that you have glossed over - how are users supposed to interact with ArticleHistory? It is my impression from Sandy that people often create ArticleHistory templates that are broken. Deleting the old templates would, it seems to me, only accelerate this trend. Raul654 (talk) 16:31, 19 December 2007 (UTC)
 * A clarification: the rate of broken articlehistory templates has markedly decreased (finally), and mostly came/comes from the GA process.  Because anyone can pass or fail a GA, it pulls in many editors who may not be familiar with articlehistory, so they are the case most likely to need to fall back on templates. We almost *never* get FA template errors anymore; it's a complete rarity now, and ah/Gimmebot is functioning marvelously in the featured areas and wrt maindate, dyktalk.   Sandy Georgia  (Talk) 17:18, 19 December 2007 (UTC)
 * Must be my newsletter article :-)
 * I'm broadly in the school 2 (or possibly 2.5 :-) In other words, I'm in favour of a bot converting these templates into article history actions when they arise, but I'm not in favour of eliminating the templates. I'm against deprecating them (especially the GA templates) unless ArticleHistory is entirely automated, because manually adding article history actions is tedious, time-consuming, and error-prone. I do not know whether complete automation is possible, as some of the article history information is awkward to find. Gimmetrow may be able to comment here. Geometry guy 13:09, 22 December 2007 (UTC)
 * Gimmetrow's list is at Template talk:ArticleHistory/work, the workpage we used when converting FAs and FFAs to articlehistory. Just yesterday, I had to correct another faulty GA passing, using the old templates (not articlehistory), so those problems continue to be related to experience level/anyone can pass a GA rather than method (GA templates or articlehistory).  Other issues with some of these templates still in use:  there are still almost 1,000 article talk pages with the old  template.  Converting those to articlehistory would be an inordinate amount of work. Probably same for others, such as old peer reviews.  Sandy Georgia  (Talk) 13:39, 22 December 2007 (UTC)
 * I agree with the above in that it makes sense to not depreciate them, and actually encourage users to still use them (unless they're familiar with AH, with a bot running through to make the change to the more complex and delicate AH. (Of course, barring the creation of some tool (Javascript based?) that can understand and parse the AH template to provide a simple form-driven interface for adding in the right details :-)  --M ASEM  16:51, 22 December 2007 (UTC)
 * Ideally the other templates should just be signals to a bot to update article history, but we are still some way from that. Geometry guy 20:54, 22 December 2007 (UTC)
 * I think of "ArticleHistory" like I think of "cite web", while I find it useful and I think it presents a more condensed and ultimately easier to read final product, I also recoginze that it is cumbersome and uses esoteric language and is hard for the uninitiated to use. Heck, I have used it a hundred times, and STILL have to load up the instruction page for it each time.  While it should be availible for people to use, there should not be a stand one way or the other as to whether it is favored over the individual templates.  While users should not be discouraged from using it to clean up talk pages, NEITHER should users be forced to use it... --Jayron32| talk | contribs  03:38, 24 December 2007 (UTC)

Oppose deprecation of other templates because ArticleHistory is too difficult to use. --J.L.W.S. The Special One (talk) 11:36, 26 December 2007 (UTC)
 * But would you be averse to anyone that did know how to use it replacing instances of GA, oldpeerreview, etc, with properly-implemented ArticleHistory?? Happy‑melon 12:37, 26 December 2007 (UTC)
 * I think there has been a clear consensus here that those who have the expertise and the time should be encouraged to replace instances of the other templates with ArticleHistory actions. No one is averse to editors doing that, as far as I can tell. Geometry guy 12:37, 27 December 2007 (UTC)
 * Happy‑melon 16:52, 27 December 2007 (UTC)
 * If you are one of the few Wikipedians who know how to use it, go ahead and replace GA, etc. templates with ArticleHistory. --J.L.W.S. The Special One (talk) 04:53, 28 December 2007 (UTC)

ArticleHistory and GA

 * I'm not sure I made my point clearly; I disagree that the articlehistory template is complicated, although it does require reading instructions. The ongoing GA errors occur just as much with the old templates as they do with the new articlehistory; editors passing GAs often don't know how to use either.  There are not errors in the featured content area; they almost all come from the GA process.  The problem isn't the templates (old or new); the problem is that anyone can pass a GA without reading instructions.  At least there is less potential for error on GA with the old templates, but on the other hand, with the new method (articlehistory) at least we have a means of finding the errors. The GA errors show up at the time the article comes to FAC, which is why I end up having to fix them all. The only way I can think to stop these GA errors is to change the GA process.  Instead of letting just anyone change talk pages, have a centralized page where GA passes and fails are lodged, and let experienced editors update the talk pages; that is, get some quality control/training over the process. Sandy Georgia  (Talk) 16:08, 28 December 2007 (UTC)
 * But then GA becomes watered-down FAC, which we don't want in the same way we don't want WP:RFR (if it's ever implemented) to be a watered-down WP:RFA. The issue is, as you say, education - what, exactly, are GA reviewers doing wrong?  Not adding oldids and/or topics seems like a likely error, but are there any others?  Happy‑melon 16:52, 28 December 2007 (UTC)
 * You name it, I've seen it done :-) Mostly not adding oldid or date, so someone else has to go back and look it up in the diffs, which is VERY time consuming.  I'm not saying to make it like FAC; anyone can pass a GA, but make it so that someone knowledgeable does the talk page updating to avoid the errors.  Have them pass the GA on a central page, where someone else does the technical/logistical aspects.  Sandy Georgia  (Talk) 17:07, 28 December 2007 (UTC)
 * The only error I've seen is an incorrect topic (e.g. "topic=television") and the templates automatically traps many such mistakes. Everything else amounts to the omission of optional information (the date, the oldid and the topic). Also an incorrect topic is treated as a missing topic, so it isn't a big deal. If any editor wants to spend time filling in this missing information, they are free to do so. Furthermore, contrary to SandyGeorgia's assertion, we do have a means to find these "errors": Category:Uncategorized good articles and Category:Good articles without an oldid.
 * Changing processes to make templates run smoothly is backwards: we should be changing templates to make processes run smoothly. On that, my main point of agreement with SandyGeorgia is that GA, DelistedGA and FailedGA are badly documented. Geometry guy 17:24, 28 December 2007 (UTC)
 * I had a missing topic, oldid and date just yesterday (or was it the day before? - lost track now). Can someone at GA start cleaning out those errors regularly, so I don't have to be the first to hit them when I prep FACs for GimmeBot?  There are several current FACs in there now, which means someone has to look that data up before the FAC is closed; it is really time consuming, and it would be stupendous if someone at GA worked on it regularly.  I didn't realize you had that category, so I withdraw my suggestion for better centralization, and change it to a request that someone process these errors so that they aren't all left to me :-))  Sandy Georgia  (Talk) 17:31, 28 December 2007 (UTC)
 * FA regulars should stop attacking GA and start suggesting how the template can be made easier to use. --J.L.W.S. The Special One (talk) 00:54, 31 December 2007 (UTC)
 * I am afraid, Hildanknight, that I don't see any attacks there. As a GA regular myself, I don't read anything above besides some suggestions to improve the process where it has flaws.  Some (such as a centralized clearinghouse and multiple reviewers) aren't likely to happen, but others are good suggestions, and none is phrased in the form of an "attack."  Also, to respond to Sandy, the bots that work the GA's has been doing a much better job of catching the errors you note on recently promoted GA's.  There are problems with the older GA's, and these are being slowly addressed through quality sweeps.  I understand your concern with the problems there, but its simply a matter of too much to do and not enough hands to do it.  I personally fix the problems you note whenever I come across them, but there are just not many people to fix the older GA's... --Jayron32| talk | contribs  01:18, 31 December 2007 (UTC)

[←]Might I suggest that a decent step forward would be to indicate, in the respective template documentation, which parameters are mandatory? The first time I used AH, I didn't include the oldids because I didn't know them, and there's nothing to indicate they are required. Similarly, I suggest advertising Dr PDA's article history script on all these template documentation pages, as I was unaware of it the first time I used any of the templates and it makes finding the oldids much easier. Carre (talk) 17:49, 28 December 2007 (UTC)

(ec: reply to Sandy) There is actually a task force that is supposed to do this, but it is not very active at the moment. Anyway, from the GA perspective, the topic/date/oldid information is optional: it is nice to have it, but not a big deal if it is missing. I occasionally use the cats to tidy up a bunch of GAs, but, to be honest, it isn't a high priority for me. Actually, in ArticleHistory, missing dates, topics, links or oldids do not generate ArticleHistory errors either, although a missing date does generate an "invalid time" error. (This could be replaced with an "Unknown" date.) It would be great if these data were always filled in, but there are many things to do at WP, and adding links and oldids are not the best use of time in my view. Are any of these data essential for "prepping" GimmeBot? Geometry guy 18:11, 28 December 2007 (UTC)
 * Will need to hear from Gimmetrow on that; my goal has been (considering the considerable time and effort GimmeBot saves everyone and the fantastic job it does) to make Gimmetrow's job easier so that the bot doesn't stall on faulty talk page templates when it closes FACs. To that end, I review every new FAC to make sure the talk page templates are in order, so that GimmeBot/Gimmetrow have less issues to deal with at the other end.  Sandy Georgia  (Talk) 18:19, 28 December 2007 (UTC)
 * The AH system is set up assuming there will always be a date. If the template is just GA, the bot has no idea when that was added without going through edit history. It does look for edit summaries with "GA" or similar when the GA template has no oldid. (It could look through diffs but that seemed silly.) The oldid is optional in AH too, but the bot fills in missing ones when it processes the page if it can understand the date. (The #time function handles ISO dates but the bot doesn't have this feature yet.) Gimmetrow 18:21, 28 December 2007 (UTC)
 * Okay, we better make the date mandatory then, and generate an ArticleHistory error when it is missing. I guess the answer to Carre's question is that the action, date and result are mandatory, but the oldid and the link are not. Geometry guy 18:26, 28 December 2007 (UTC)
 * Result is optional for peer reviews, I believe. Gimmetrow 18:29, 28 December 2007 (UTC)
 * Peer review failed hehe! <b style="color:forestgreen;">Happy</b>‑<b style="color:darkorange;">melon</b> 18:32, 28 December 2007 (UTC)
 * Someone once used "not reviewed" as the result for a PR with no comments ;) Gimmetrow 18:36, 28 December 2007 (UTC)
 * I was thinking more along the lines of "result = helpful" or "result = waste of time" for peer reviews :) Geometry guy 18:38, 28 December 2007 (UTC)
 * Gimmetrow's response reminds me, I can't tell you how often GA passes include no edit summary, so that when you're looking for the oldid, Dr pda's script doesn't work, so you have to go back through the history, diff by diff, to find the GA pass. Sandy Georgia  (Talk) 18:42, 28 December 2007 (UTC)
 * Pain enough when you're doing it by hand. It's even worse for a bot: yeah, just get the last 100 revisions of Talk:Foo - that'll be a nice, oh, five megabyte API query!  Grrr....  <b style="color:forestgreen;">Happy</b>‑<b style="color:darkorange;">melon</b> 18:56, 28 December 2007 (UTC)
 * This is a good point: I'm going to be updating various GA guidelines soon, and have taken this on board. Geometry guy 18:58, 29 December 2007 (UTC)

Update

 * 1) Happy melon has listed some of the above templates (namely the ones which are currently unused) at Templates for deletion. Comments welcome.
 * 2) I've added an error check for a missing date to Template:Historyoutput (part of ArticleHistory). This threw up a dozen or so articles, which I've been fixing.
 * 3) Concerning dates for GAs, the issue here requires some work, because (much to my amazement) GA has never, in its entire history, provided a date parameter, in spite of the fact that DelistedGA and FailedGA both do! Conseqently, hardly any GAs have a review date! I've added a date parameter to the GA template today (and improved the documentation of the GA templates), but there are a lot of existing GAs to fix. Fortunately, almost all GAs have an oldid, so I imagine a bot could sweep through the GAs, fetching the review date using the oldid information, and adding it to the GA template on the talk page. Geometry guy 18:58, 29 December 2007 (UTC)


 * No real point adding a date if there's an oldid. It's essentially duplicate information (if done by a bot). And yes GA hasn't supported a date parameter, but GimmeBot has since, oh, about 10 months ago :) (Seriously, if it looks like, the bot handles it; there used to be quite a few of these.) Gimmetrow 01:24, 30 December 2007 (UTC)


 * Okay that's good. The templates accept the date in any form accepted by #time, but the documented format is "2007-12-30". Reviewers get used to using dates in the hundreds of other nomination templates, so this should be more reliable than an oldid.If we can also agree that topic parameters are optional (and hence remove this error check from ArticleHistory) then very few GAs should need prepping, since they'll almost always either have an oldid or date.
 * Meanwhile the TfD has closed, with 3 templates deleted, 2 deprecated, and the GAN guidelines have been updated. Geometry guy 11:14, 30 December 2007 (UTC)


 * Isn't the topic field more or less optional in the current code? I think the topic only generates a Category:ArticleHistory error if it's specified *and* it's not recognized by GA/Topic. If the topic field is missing or empty, I think it's only flagged if it's a GA or FFA/GA, and then only with Category:Uncategorized good articles. Gimmetrow 23:28, 30 December 2007 (UTC)