Wikipedia talk:Did you know/Archive 9

Concerns and commonsense
After a long time, I got round to updating the DYK as it has been due for over 10 hours. I updated DYK regularly for around 4 months, from mid-Dec — mid-Apr of 2006. I was the sole updating admin for a couple of months and I can't recall an occasion where worthy suggestions had to be passed over due to lack of admin presence. However, I was shocked today to see that 5-day old suggestions are still lying around and older days' suggestion were still cluttering the page. With around 3-4 admins, sincere and competent at that, being involved in updation, why has it happened? Let me proffer some explanations - any chore if it becomes drudgery, cannot be done easily. Don't ask me to provide a cite ;) for my statement. Updating DYK is daunting for any admin - apart from the multitude of technical things (protecting/unprotecting, archiving, resetting the time counter) and political things (balancing the length on main page, ensuring diversity, ensuring that no entry becomes stale), we also have the maintenance things such as updating talkpage and 2 user talkpages for each entry. While it helps in the updating admin's catering to our editcountitis, I don't see practical use. FAs, In the news, selected anniversaries - is there any intimation to the user/suggestor/initiator? No. Why on DYK? Assuming that an average of 3.5 updates are done in 24 hours and each update has average of 5 suggestions with 3 self-noms and 2 others' noms, the no. of talkpages I need to update are 5*3.5 talkpages + 5*3.5 usertalk pages(creators) + 2*3.5 usertalkpages(nominators) = 42 edits. For each update, add atleast an edit to template page, its talk page, image page and protection/unprotection, another 14 edits. So, 56 edits per day to keep DYK running? People may not mind so much if they are not intimated abt their article making it to DYK when compared to their article not making it to DYK, because the entry has become stale. To prevent entries becoming stale, we need to have more admins updating it in a simple way. See the earliest posts on this page about the proposed changes in DYK which I had followed and which were not objected to. Why not keep it simple? Why not have a manual DYK counter, esp. since the updation itself takes 20-30 min and the DYK refresh template would become stale by that much time? Why have a in-use template for admins when the probability of edit conflict on the template page is 0.01 and on template talk page is 0.03 (I quote these probabilities from my experience of ~200 updates, and edit conflicts of 2 and 6 respectively, which were fixed within 2-3 mins)? Are we interested in creating better mouse traps? Why not concentrate on pressing needs like archival bot or facilitating easier updation? Am sorry if it looks like a long-winding rant or if i appear a techno-phobe but I find it difficult to digest that worthy noms just don't see the light despite a longer and a daily DYK section just because it becomes progressively difficult for an admin to update DYK. Please understand that my rant is against the system in its current form. --Gurubrahma 18:01, 21 June 2006 (UTC)
 * I think it's simply due to the larger number of nominations that seem to be occurring lately. We don't have to use every nomination; only the most interesting ones should be used, and the template should be kept highly interesting and varied, since that is about all we have to offer. &mdash; BRIAN 0918 &bull; 2006-06-21 18:12
 * "Interesting" is subjective. So, unless there are objections on a nom (by any editor, including the updating admin) it should be a shoo-in for the DYK template. How wd ppl know why their suggestion hasn't made it if it wasn't commented upon but still doesn't make it to the template? --Gurubrahma 18:27, 21 June 2006 (UTC)


 * I do agree about keeping it simpler. We should probably ditch the in-use templates, both in the template and on the talk page. What else can we get rid of? The DYK-Refresh template, though interesting for checking the time, is just more wasted edits. Why not simply make it clear that you have to wait at least 6 hours, and provide a link to the template history to check when the last update was. Most people know how to add. :) &mdash; BRIAN 0918 &bull; 2006-06-21 18:18

How Lar uses time
I have done some thinking about where I spend time and how much. The process for me is:
 * 1) Mark that we're starting. (turn on the hidden box in the template talk, then go and turn on the hidden comment in the template, grab the articles at the same time so they can be archived.
 * 2) Archive. This is a big timewaster due to *mp vs * differences on the bullets... perhaps we should use *mp across the board? The automation that used to make the archive entries for you is broken and I don't know how to contact whoever did it. It's worse the larger the archive is, so keeping the archive emptied helps matters. (3-5 min)
 * 3) Pick a picture. Check it for freedom. Protect it. If it's commons, that means copying it down, uploading etc. (I'm working on uploading enough to stand for commons admin so I can protect there, which will save time) (5 min)
 * 4) By far the largest part of the time is spent on reviewing and selecting candidate articles, (which means checking the article for a number of different things, good length, fact is present in article, good refs/cites) stripping away comments and picking the best version of the hook, fiddling with the wording and grammar to try to avoid other editors making needless edits to fix small things, sorting into the right order, etc. (15-25 min)
 * 5) one final check of the template with the new items, remove the comment block notice, then save it, check main.
 * 6) This part goes fast now that I created js to automate: Plop notices on all the article talk pages. (Now a 2-3 minute job)
 * 7) This part goes relatively fast with automation as well: Plop notices on all the user talk pages. (5 min because I still have to paste article names into the dialog boxes, and you ahve to check to see if there are noms and auths)
 * 8) Double check main to see if there are any issues I missed the first time, see what Brian changed in the meantime.
 * 9) Remove the articles that were nominated. Check each one removed to see that I left the right things on the talk pages (in my contribvutions). 5-10 min
 * 10) unprotect or delete if from commons the most recent previous picture. 2 min
 * 11) Remove the notice from the template talk page and update the clock. 2 min

That adds up to a bit more than the 40 min it usually takes me but gives an idea of proportions.


 * Zocky's searh box script is ideal for dealing with the -> * issue. -- Cactus.man   &#9997;  13:59, 22 June 2006 (UTC)
 * I love that widget, it's very powerful. Last I checked it was still forcing the edit box width to not resize with the window width though. As soon as that's fixed, it goes back in...  + +Lar: t/c 15:07, 22 June 2006 (UTC)
 * It may be fixed now, the talk page intimates it is. Will test later.  + +Lar: t/c 15:23, 22 June 2006 (UTC)
 * It'd be easier just to let the *mp's pile up in the archive, and then put them in notepad and mass-replace. &mdash; BRIAN 0918 &bull; 2006-06-22 15:11
 * Good point. Or use them in the archive too? I dunno... they seem to work there OK.  + +Lar: t/c 15:23, 22 June 2006 (UTC)
 * On point 3, are you sure that just protecting the version on Commons would be a good idea (presuming you do get promoted to admin there)? With the way that images work, that'd mean that if any vandal uploaded an image of the same name to Wikipedia while it was displayed on the Main Page, it'd override the commons-linked image. GeeJo (t)⁄(c) &bull; 15:36, 7 September 2006 (UTC)

Refresh templates and automation, more thoughts
I find the DYK-refresh template and the normally hidden box warning that an update is in process to be tremendously useful. I was getting a LOT of edit conflicts all the time without these present. I would strenuously object to any change to this part, losing inuse and refresh save LITTLE time and prevent a LOT of problems. Traffic on this page is higher now than it was before.

I support additional automation where it makes sense. I'd love to see the archive processer fixed, it would save some time. I created js funcs that at least for me almost completely automate leaving the talk notices (they were pooh-poohed by Brian as not useful though) and I highly recommend them... I have ideas for more automation as well.

I have a final comment. I think we'll have more people helping out if we have a more collegial atmosphere here. For one, I find the atmosphere a lot less collegial since we had the mandatory reference flap, for reasons I think most of us know but which I will not spell out. Less collegial means less desire to do the work.  + +Lar: t/c 18:38, 21 June 2006 (UTC)
 * Care to point out where I "pooh pooh"'d your js as not useful? Or did you just assume that I would say something like that? &mdash; BRIAN 0918 &bull; 2006-06-21 18:44
 * Wikipedia_talk:Did_you_know Reads like pooh-poohing to me. (and others, confirmed offline at the time)  + +Lar: t/c 19:05, 21 June 2006 (UTC)
 * Where? I recommended you talk to Lupin. Then I detailed what I'm currently doing. Nope, no pooh-poohing, no matter what these mysterious "others" have to say. &mdash; BRIAN 0918 &bull; 2006-06-22 00:55

Updating admins are not idiots. I'd trust them to update the time properly when they can do more complex things like c-uploads, protects and archives. Only article talkpage template shd be compulsory, not others, as I mention in the 4th post of this page and the subsequent discussion with Ghirla. Let all other things be optional. --Gurubrahma 18:27, 21 June 2006 (UTC)
 * Disagree, see above.  + +Lar: t/c 18:38, 21 June 2006 (UTC)
 * Honestly, 40 minutes per update and at 3.5 updates per day is 140 minutes, i.e. 2hr.20min., not just worth it. For an admin who is trying to update it first time, he wd be bewildered with the DYK-refresh and all that. I don't know abt ur automation, but overall, it doesn't seem to save much time. For someone who works from a slow internet connection like me, using inuse templates on the template page and talk page take more time. Also, the red box is no guarantee, ppl. can still add while you are updating. --Gurubrahma 18:54, 21 June 2006 (UTC)
 * It is not a guarantee but it vastly lessens conflicts. It's very easy to use and is worth the time. The place to find time savings is in the selection process itself which is where the majority is. My automation saves me over 10 minutes over doing it by hand, I find.


 * Well, everybodys working methods vary. I see no harm in keeping the hidden box warning for the talk page that an update is in progress and the UpdatingDYK for the template itself, although I seldom use them myself. I also find the DYK-Refresh useful as well, no reason to get rid of it, although it does seem prone to breaking from time to time. Haven't had time to test Lar's javascript tabs yet, but every time saving device helps. -- Cactus.man  &#9997;  12:47, 22 June 2006 (UTC)

which templates to keep

 * I'd keep the article-creator-notification templates, just to keep the self-nom folks coming back (it's always nice getting an "award" on your talk page). I think the nominator template is less necessary, simply because most of the nominations are self-noms. &mdash; BRIAN 0918 &bull; 2006-06-21 18:47
 * No one prevents the editor from adding the template to his user talk page if he wishes to. I dunno if you read the exchange between me and Ghirla above. DYK medal is a better award, anyways. --Gurubrahma 18:54, 21 June 2006 (UTC)
 * I would never give myself an award or notice of any sort. We would have to change norms if we want people to put notices on their own pages after their article is selected.  + +Lar: t/c 19:05, 21 June 2006 (UTC)
 * I do it just so I can keep track of the articles I nominated, since I've been listing them on my user page. &mdash; BRIAN 0918 &bull; 2006-06-21 19:28

Why don't we go back to the old process of simply listing the latest entries under "The following users need to be notified" and let a non-admin or anyone else handle it? &mdash; BRIAN</b> 0918 &bull; 2006-06-21 19:02
 * Not where the bulk of the time is spent, see my time above. Maybe yours differs? No savings and worth doing.  + +Lar: t/c 19:05, 21 June 2006 (UTC)
 * So, I'll simply list them under that section, and you can go through them with your automation and update them as you please. Or don't, and someone else will. Just because it's not the bulk of the time doesn't mean it's not good to save a little time. &mdash; <b style="color:black;">BRIAN</b> 0918 &bull; 2006-06-21 19:26
 * Let's have one process that we all use, knowing that you do things one way and everyone else another is just a recipe for havoc.  + +Lar: t/c 19:38, 21 June 2006 (UTC)
 * G pointed out on my talk page that one size may not fit all. If people on slow connections want to do less, or do it differently, and we can come up with a documented and clear alternative process (which doesn't result in confusion about which was and wasn't done) I for one am fine with it, can't speak for anyone else of course.  + +Lar: t/c 21:53, 21 June 2006 (UTC)

Maybe we should have something like a Template:Did you know/Next, so that the update can be prepared at a more leisurely pace, and possibly with the participation of non-admins.--Pharos 19:20, 21 June 2006 (UTC)
 * I envision Polish/Russian revert wars, among others. :) &mdash; <b style="color:black;">BRIAN</b> 0918 &bull; 2006-06-21 19:27
 * my worry too  + +Lar: t/c 19:38, 21 June 2006 (UTC)


 * (2X edit conflict) Aside from the revert wars issure, that's not a bad idea (permanant semiprotection would prevent most pointless edits); it also stops the influx of posts to Wikipedia talk:Main page/Errors which follow every DYK update. One unrelated point though. Would protecting the image on Commons stop someone else uploading a different image to Wikipedia in the same name as the Commons image? smurrayinch e  ster( User ), ( Talk ) 19:29, 21 June 2006 (UTC)


 * I don't think so. I've noticed people forgetting to temporarily upload the file to Wikipedia in order to protect it. &mdash; <b style="color:black;">BRIAN</b> 0918 &bull; 2006-06-21 19:31


 * That's not what he meant. Some have claimed that protecting the commons image is enough. I thought it was. I may have been wrong. But if you still have to upload to WP, protecting on commons doesn't help. + +Lar: t/c 19:38, 21 June 2006 (UTC)
 * You would have to have some sort of "blue page" on Wikipedia (if only of text) and protect that too. However, you would not have to upload the image locally.--Pharos 19:41, 21 June 2006 (UTC)
 * Thanks for clearing that up. &mdash; <b style="color:black;">BRIAN</b> 0918 &bull; 2006-06-21 19:42


 * I don't think we should overplay the danger of revert wars in such a situation (though it is a serious concern). Higher levels of technocracy like this tend to discourage the over-zealous; as long as there was guideline against self-promotion, we shouldn't have too much of a problerm.  And it would be still admins who make the final decision.  Anyway, I think it's worth a trial period.--Pharos 19:53, 21 June 2006 (UTC)
 * I'm willing to try a trial if we can hash out exactly what the process is, and get consensus first, before making radical or possibly disruptive changes. can you bang out what you have in mind? Maybe in a new section? (PS I added a lot of level 3 heads to reduce conflict and make it easier for dialup users to work with)  + +Lar: t/c 21:16, 21 June 2006 (UTC)


 * I agree with Brian, keep the article creator notification template, but the nominator one is probably less important. Nominators can always watchlist the article talk page and spot the notice there, but it's good in my view to give the article creators a nice pat on the back with a notice.


 * I'm not too sure about the idea of a Template:Did you know/Next. Quite apart from the possible revert war issue it would make the updating process too cumbersome and almost unworkable. Getting agreement over which 5 or 6 entries to use could become a complete nightmare. The nomination page already has plenty of comment on many entries, let's continue to trust the updating admin to use judgement and common sense when the time for an update comes around. -- Cactus.man  &#9997;  13:03, 22 June 2006 (UTC)

what problem are we trying to solve here
I think we need to step back here... the problem Gurubramhma brount up was too many noms not getting selected. But a week ago we had a drought, and we deliberately slowed the update rate down significantly to compensate... So what's the real problem here? Maybe it isn't what we think it is. Also... While I'm all in favour of process improvement, one thing I've learned in my 25+ years of professional experience is, don't optimise the stuff that doesn't matter. Find where the bulk of the time is spent, and optimise that. The bulk seems to be in the picking the articles itself process, at least for me. Perhaps others can post their timings to the same level of detail to see if their bulk of time is where mine is. If others can efficiently pick noms faster than 15-25 min, I want to adopt their process. If others spend more than 2 min on posting notices to article talk pages, lothers should adopt MY process, and so forth. Sparring around about what we should do may be pointless without measurements  + +Lar: t/c 19:38, 21 June 2006 (UTC)

brian0918's time
If anyone cares, my time for this last update was 30 minutes. 3 minutes to add updating notices and archive. 13 minutes to select new entries [including proofreading an article with typos in Word, and adding comments for nominations I didn't use]. 7 more to protect the image, add the entries into the template, proofread, and save. 3-4 minutes to update the refresh clock, download and upload the image from commons to wikipedia (then add the licensing and info). And 3-4 minutes to post notices on the articles'/creators'/nominators' talk pages (5 articles, 5 creators, 2 nominators), and then clear out the "staging area". &mdash; <b style="color:black;">BRIAN</b> 0918 &bull; 2006-06-22 01:11
 * Thanks Brian! How do you pick the new entries so fast? That's where I dump a LOT of time, I find that it takes a fair bit to read the articles, check to see how they are referenced, etc. And I'm still not getting perfect results, someone pointed out that one of the articles I picked this AM may not be factually correct... The image protection time at 7 min (plus another 3-4 ?... wasn't totally clear) might be a place to look for savings... I can get it done a bit faster, but not much, and it sounds like Cactus.man also doesn't get it done really fast... I think I am going to try to see if that can be automated better... but I have my doubts as it spans two sites, etc.  + +Lar: t/c 15:12, 22 June 2006 (UTC)
 * I think it goes quickly because I spend other time looking through all the entries. "13 minutes" is just how much time I spend in updating. When I said 7 minutes, I meant 7 minutes to: protect the image, add the entries into the template, proofread, and save. I had forgotten to upload the image from commons, so I then had to go back and do that. It could still be done more quickly. The only part that goes quickly is the talk page notification. &mdash; <b style="color:black;">BRIAN</b> 0918 &bull; 2006-06-22 16:29

Cactus.man's method and time
As suggested, my method and approximate times follow. I don't use the warning templates or the staging area, I just prepare everything in my text editor (blatant plug) in advance and then bung it in the template, deal with the talk page and fire out the notifications. Typical order and timings:
 * 1) Review articles and images for selection. Check articles for quality, length, date of creation, correct author, stated fact is mentioned etc. Copy and paste to text editor. (15-20 mins).
 * 2) Prepare draft template entry in text editor, paste into template and preview, compare length versus current template content and likely size on the main Page. Paste old entries to text editor for archiving later. (5 mins).
 * 3) Protect the selected image (1 min - 5 mins (if uploading from commons))
 * 4) Save the new items to the template and purge to update the Main Page.
 * 5) Update the talk page refresh clock, add old items to the archive, remove promoted items from the suggestions area, unprotect or delete the previously featured image (5 mins).
 * 6) Place DYK notices on article talk pages, creators talk pages and nominators talk pages (5 - 10 mins).
 * 7) Done, have a cup of coffee (5 mins).

Hope that helps. -- Cactus.man  &#9997;  13:40, 22 June 2006 (UTC)

Russian articles?
I don't know if it's only me, but I've noticed that everyday, there's at least one article that is related in one or another way to Russia; why is it like this? etc... --Shandristhe azylean cat 21:28, 21 June 2006 (UTC)
 * ...that although Ernst Neizvestny's work had been denounced by Nikita Khrushchev as degenerate art, he was commissioned to sculpt Khrushchev's tomb?
 * ...that Grand Duchess Maria Pavlovna of Russia (pictured) was noted as a patroness of Schiller, Goethe, and Liszt?
 * ...that due to protests and financial problems, the Saint Petersburg Dam was one of the Soviet Union's most notorious long-term construction projects?
 * ...that the Alexander Column (pictured) on Palace Square in Saint Petersburg, despite its weight of 600 tons, is set so nicely that no attachment to the base is required?


 * We pick from the noms presented, and there have been a large number of articles involving Russia (as well as ones involving Poland) presented. I don't think it's an intentional bias, it's just what the pool has in it. (someone else told me in IRC that it was funny that there were so many train articles picked... :) ) If it's a serious problem, let's discuss but my suggestion would be to nominate more articles from other areas... Hope that helps.  + +Lar: t/c 21:51, 21 June 2006 (UTC)


 * Also, because Ghirlandajo finds cool pictures for the main page. :) Note that while there may be a lot of articles on Russia/Poland, there is usually never more than 1 of each in any given set of DYK articles on the main page. &mdash; <b style="color:black;">BRIAN</b> 0918 &bull; 2006-06-22 01:43


 * It's because Ghirla (Russia) and Piotrus (Poland) put up heaps of stuff relating to their country. Back in May when I had more time for article writing, there was about an Australian one up every day or two.Blnguyen | rant-line 01:52, 22 June 2006 (UTC)


 * Brian and Blnguyen are exactly right. Fundamentally, the answer to "too many X" is to add more Y, Z, A, B and C, not restrict X. As long as we aren't overbalancing it's not a bad thing to select good articles.  + +Lar: t/c 14:36, 22 June 2006 (UTC)

Lists in articles
Joseph Krumgold, which is on DYK now, has well under 1000 characters/bytes (if you ignore the list of works). Should we be including such lists in determining whether an article is long enough? &mdash; <b style="color:black;">BRIAN</b> 0918 &bull; 2006-06-22 13:46
 * I think it's reasonable to evaluate lists as part of the article, as they are kind of content. I saw your comment to that effect on the nom, thought about it, looked at the article, and felt it was a good enough article (including the list) to warrant selection. IIRC we have selected actual LIST articles in the past, but I may be mistaken. I'm interested in what others think, and if it's a bad practice, would be happy to see it added to the guidelines, but I fear instruction creep as it's rather a small point... Thoughts?  + +Lar: t/c 14:32, 22 June 2006 (UTC)