Jump to content

Wikipedia talk:Did you know: Difference between revisions

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Browse history interactively
← Previous editNext edit →
Content deleted Content added
VisualWikitext
Line 109: Line 109:
:*Not to worry, there will be plenty of opposition to a suggestion of 1 set/day with 300+ active nominations. For me, I'd want to see total nominations at fewer than 150 and a single-figure nomination rate before entertaining the extreme suggestion of a single daily set. [[User:BlueMoonset|BlueMoonset]] ([[User talk:BlueMoonset|talk]]) 05:15, 18 September 2014 (UTC)
:*Not to worry, there will be plenty of opposition to a suggestion of 1 set/day with 300+ active nominations. For me, I'd want to see total nominations at fewer than 150 and a single-figure nomination rate before entertaining the extreme suggestion of a single daily set. [[User:BlueMoonset|BlueMoonset]] ([[User talk:BlueMoonset|talk]]) 05:15, 18 September 2014 (UTC)
:::MBS, you're a smart guy (or gal -- never did know which). But you keep missing the point. It doesn't matter how many noms are awaiting approval -- all that matters is the size of the approved reserve, because we can't put noms-awaiting-approval on the MP. [[User:EEng|EEng]] ([[User talk:EEng|talk]]) 02:03, 19 September 2014 (UTC)
:::MBS, you're a smart guy (or gal -- never did know which). But you keep missing the point. It doesn't matter how many noms are awaiting approval -- all that matters is the size of the approved reserve, because we can't put noms-awaiting-approval on the MP. [[User:EEng|EEng]] ([[User talk:EEng|talk]]) 02:03, 19 September 2014 (UTC)
::::EEng, I'm not missing the point. I'm just aware of more variables. We can vary the size of the sets up or down, which allows for fine tuning; if we had (say) 45 approved hooks in the reserve, we could still easily fill two sets a day, since approvals will continue to occur. At the moment, you're one guy (or gal?) with a new theory that's suddenly the be-all and end-all, while there's years of DYK experience that you're apparently unaware of. [[User:BlueMoonset|BlueMoonset]] ([[User talk:BlueMoonset|talk]]) 07:12, 21 September 2014 (UTC)
*Just now 287 noms under review, 95 in reserve. [[User:EEng|EEng]] ([[User talk:EEng|talk]]) 02:03, 19 September 2014 (UTC)
*Just now 287 noms under review, 95 in reserve. [[User:EEng|EEng]] ([[User talk:EEng|talk]]) 02:03, 19 September 2014 (UTC)
*After the next update (in order to be comparable to the numbers above) we'll have 290 under review, 92 in reserve. [[User:EEng|EEng]] ([[User talk:EEng|talk]]) 00:07, 20 September 2014 (UTC)
*After the next update (in order to be comparable to the numbers above) we'll have 290 under review, 92 in reserve. [[User:EEng|EEng]] ([[User talk:EEng|talk]]) 00:07, 20 September 2014 (UTC)

Revision as of 07:12, 21 September 2014

SKIP TO THE BOTTOM


Did you know?
Introduction and rules
IntroductionWP:DYK
General discussionWT:DYK
GuidelinesWP:DYKCRIT
Reviewer instructionsWP:DYKRI
Nominations
Nominate an articleWP:DYKCNN
Awaiting approvalWP:DYKN
ApprovedWP:DYKNA
April 1 hooksWP:DYKAPRIL
Holding areaWP:SOHA
Preparation
Preps and queuesT:DYK/Q
Prepper instructionsWP:DYKPBI
Admin instructionsWP:DYKAI
Main Page errorsWP:ERRORS
History
StatisticsWP:DYKSTATS
Archived setsWP:DYKA
Just for fun
Monthly wrapsWP:DYKW
AwardsWP:DYKAWARDS
UserboxesWP:DYKUBX
Hall of FameWP:DYK/HoF
List of users ...
... by nominationsWP:DYKNC
... by promotionsWP:DYKPC
Administrative
Scripts and botsWP:DYKSB
On the Main Page
Main Page errorsWP:ERRORS
To ping the DYK admins{{DYK admins}}
Shortcut



This is where the Did you know section on the main page, its policies and the featured items can be discussed. Proposals for changing how Did You Know works were being discussed at Wikipedia:Did you know/2011 reform proposals.

Increase to 3 sets a day?

The backlog is now over 300 noms with 54 hook approved, and there is no backlog in the queue. I say it's time to increase the number of sets per day to 3, as 300+ noms has typically been the point when an increase would be made. Thoughts? —Bloom6132 (talk) 09:27, 10 September 2014 (UTC)[reply]

Cwmhiraeth posted something similar here a couple of days ago in Hook accumulation. It probably is about time, especially with the final push for the Wikicup underway, and the normal post–Labor day increase in nominations. BlueMoonset (talk) 15:16, 10 September 2014 (UTC)[reply]
  • I was asked to look into this, and although there hasn't been much discussion I agree with Blue and Bloom. Just don't recall where the bot's adjustment page is... — Crisco 1492 (talk) 00:51, 12 September 2014 (UTC)[reply]
    • The critical figure is the number approved, not the number of noms. I agree that 300+ is excessive, but experience has shown that 54 hooks is only enough to fill two prep areas! There is no way that we can sustain three per days with that approval rate! Hawkeye7 (talk) 08:51, 12 September 2014 (UTC)[reply]
      • 54 approved = 2 prep areas? Do you mean 2 sets? 54/6 = 9, so assuming some duds there should still be at least 7 or 8. — Crisco 1492 (talk) 15:03, 12 September 2014 (UTC)[reply]
        • Haven't we been putting 7 hooks in a set? 54/7<8. I think we should wait a little while before going to 3 sets per day. We will likely end up going to 3 sets for a while and soon find ourselves unable to fill prep sets far enough in advance to put them into the queue before the bot is telling us the queues are empty. It is annoying seeing the page run sporadically. Let's wait until the backlog is larger so we can keep on our schedule.--TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 17:56, 12 September 2014 (UTC)[reply]
  • Yes, it's been 7 hooks per set, but that can easily be reduced to 6. And with a backlog of over 300 (now 335 to be specific), I don't see why anyone should be advocating waiting longer to increase to 3 sets a day, especially when there are 71 approved hooks already. That's enough to build 10 sets of 7 hooks each. At the rate we're going at, it'll take 5 days for that to be used up (not to mention the additional approvals that would take place within those 5 days). Rule of thumb has always been – reduce to 2 sets when under 200 noms; increase to 3 sets when over 300. The time for an increase is now. —Bloom6132 (talk) 19:25, 12 September 2014 (UTC)[reply]
  • The number of noms should not determine the throughrate, as Hawkeye explains, so if that ever was a rule of thumb it shouldn't be. I too think that 2 sets a day remains an appropriate rate. Nikkimaria (talk) 01:20, 13 September 2014 (UTC)[reply]
  • We have me, Crisco, Bloom and BlueMoonset for increase to 3, and Tony, Hawkeye and Nikkimaria preferring to stay at 2 for the time being. Much as I'd love to tweak, I think it might be prudent until a couple more opinions have weighed in. Cas Liber (talk · contribs) 09:06, 13 September 2014 (UTC)[reply]
  • Per WP:NOTDEMOCRACY, your vote counting is irrelevant to the current situation. Hawkeye didn't actually specify whether there should be 2 or 3 sets a day, and stated his concern when there were only 54 approved hooks. There are now 62 of them, not to mention the fact that all four prep areas have now been filled. That, along with the 3 queues, covers at least the next 3.5 days, if not more. —Bloom6132 (talk) 15:09, 13 September 2014 (UTC)[reply]
  • I think it is pretty clear that Hawkeye7 said 3 per day was too much. I don't think WP:NOTDEMOCRACY is relevant here. We are trying to determine a collective consensus on what to do and there is certainly not a consensus to move to 2. It really would be no more helpful for me to bluelink CONSENSUS that it is for you to bluelink NOTDEMOCRACY. There is just not a consensus to go to 3 right now.--TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 16:14, 13 September 2014 (UTC)[reply]
  • Per Hawkeye: The critical figure is the number approved, not the number of noms. I agree that 300+ is excessive, but experience has shown that 54 hooks is only enough to fill two prep areas! There is no way that we can sustain three per days with that approval rate! Unless I'm blind, nowhere in that statement does Hawkeye say he's in favour of staying at 2 sets a day or that he's opposed to three sets a day. Taking his quote within the context it was given, he says we can't sustain 3 sets a day "with that approval rate". Since then, the number of approved hooks has increased to 64 (and would've now been over 80 approved had not all 4 prep areas been filled today). That's an increase of more than an entire set. —Bloom6132 (talk) 16:46, 13 September 2014 (UTC)[reply]
  • And your claim that NOTDEMOCRACY is not relevant is simply inaccurate, since the way you derived "consensus" is by counting votes. You previously stated, That still only makes 5 ayes and 4 nays (with TRM below).) If that's not vote counting, then what is? —Bloom6132 (talk) 16:50, 13 September 2014 (UTC)[reply]
  • Why the delay? I still see hooks being pulled here and there at the same pace they were back when there were 3 sets a day. Hence, the claim correlating slow pace with quality control is iffy at best… —Bloom6132 (talk) 19:30, 14 September 2014 (UTC)[reply]
  • I wasn't talking about solely the main page. If you haven't noticed, there's been threads like "Queue1 problems" and "Pulled Perce Wilson from Queue5". The end result is exactly the same – inferior noms by inexperienced users. Slowing down the entire DYK process with slogans like "No need to rush" repeated ad nauseum is like putting a band aid on a bullet wound. —Bloom6132 (talk) 21:19, 14 September 2014 (UTC)[reply]

"If hooks get pulled prior to main page …" – they shouldn't have to pulled at all. The fact that they are even promoted to the prep areas and then to the queues in the first place is the real problem. Insisting on 12 hour cycles is merely a temporary solution to a much bigger problem (another band aid on a bullet wound). We all know you want to completely overhaul the DYK system. That's fine. But repeating your opinion ad nauseum each time doesn't make it an iota more correct. Create an RFC instead and we can start some meaningful, constructive reforms from there. And BlueMoonset confirmed what I had been suspecting all along – 3 sets a day is indeed the norm. So why don't we go back to 3 sets and see how it pans out? We can always change back to two when we run low on hooks or when there's an increase in the number of pulled hooks. For someone who staunchly advocates necessary change, it's hard to comprehend how you can simultaneously be so resistant to it, especially when this change here to 3 sets has become very much necessary. —Bloom6132 (talk) 22:20, 14 September 2014 (UTC)[reply]

I think two a day is appropriate too. No need to rush, and the quality is increasing slowly but surely as more people have time to visit the queues before they get to the main page. The Rambling Man (talk) 09:09, 13 September 2014 (UTC)[reply]

  • Our purpose here is to choose a rate of consumption (hooks per day moved to main page) and a nominal number of noms approved-but-not-yet-run-on-MP (including those in prep or Q -- let's call this the approved reserve) -- matched such that, while the approved reserve will always fluctuate, the probability of its reaching zero on any given day is very small.
  • Right now we've got 340 noms on the board, 80 in "approved reserve". For this discussion only the 80 has any significance. (If anything the 260 "excess" tells us large numbers of noms are stalled for unknown lengths of time.)
  • A sample of 28 hooks promoted in the last week suggests that the average time from nomination to approval is (believe it or not) about 19 days, with a standard deviation ("give or take") of about the same amount. (Those who can remember their elementary statistics without painful PTSD relapses should take a moment to think about what that last bit implies.)
  • Here's the complete data (days from nomination to final tick) -- it's a highly variable process: 0,0,1,2,2,3,3,3,5,5,6,4,7,10,11,20,22,23,26,31,33,32,35,38,44,41,56,62
  • How to translate this data into the right rate of consumption, approved reserve, etc. (not to mention what we consider an acceptable probability of running dry on any given day) is complex. However, if we move to consumption of 3*7=21 hooks per day, then the current reserve of 80 is only 4 days' worth. From the data, only about 25% of hooks are approved in 4 days or less. Combining this with the recent nomination rate of about 9 per day means that....
  • Wait a minute.... The rate of nominations (9 per day) over the last month has been insufficient to sustain even 2*7=14 promotions per day, much less 3*7=21 promotions per day. That alone tells us that we shouldn't go to 3 sets per day, and that even 2 sets per day may be risky.
  • I'll complete the above lecture when 30 days of nominations rises to the 21 needed to sustain 3 sets per day. Until then 2 sets per day is the absolute max unless we want to find ourselves running out sometime in the next few weeks. Then that stupid DYKbot starts bugging us, and then my violent fantasies of bot-o-cide begin to return, and...

[My own idiotic comments struck -- see below. Lecture to resume presently...]

EEng (talk) 16:14, 14 September 2014 (UTC)[reply]

  • EEng, your figure of 9 nominations per day is about half what it should be. Since the beginning of August, there have only been 7 days where the nominations have gone below 14 per day, and 6 have exceeded 21 per day. (Note that I'm not counting the days that are still filling under Current nominations, but even there only the two most recent days are below 9.) If you're basing your count on the List of DYK Hooks by date, that table doesn't count already promoted or rejected nominations, so it's useless for determining the rate of nominations.
  • In fact, the average rate of nominations in August was 17 per day. At a promotion rate of 14 per day, the number of waiting nominations will increase, as indeed it has been doing. For the first seven days of September (and the September 7 total of 18 nominations can still increase over the next six hours), the average is currently 18.85 nominations per day. If we stay at 14 promotions per day, our backlog will continue to increase, and it's enormously high already. If we increase to 18 or 21 per day (in three sets), then our unprecedented backlog will start to decrease. We can always back down to 14 per day later if we start running out of approved hooks. BlueMoonset (talk) 18:18, 14 September 2014 (UTC)[reply]
Feel free to grab this trout by the tail and give EEng a quick whack!
Arggh! I'm an idiot! I did exactly the stupid thing BMS guesses above that I did. Please, I beg everyone, help me expiate my sin! Whack me with the trout provided at right.
BMS, where did you get the full noms-per-day figures? Are they hidden somewhere? Also, can you give separate figures for (still in review vs. approved vs. rejected)?
I'll resume the above lecture after my ego has recovered somewhat (though the conclusion remains that 2 sets/day is the maximum prudent rate of consumption -- I'll just have to use some ∑s and δs and ∫s and stuff). EEng (talk) 18:45, 14 September 2014 (UTC)[reply]
EEng, I used the brute-force method for nominations: took T:TDYK into the editor and counted the number of templates per day. A bit of a pain, but it works...so long as someone hasn't decided to clean out the promoted/rejected ones in order to reduce the loading time of the page. (This was last done on July 27, so it wasn't an issue for August or September.) I did have to go back into the page history to check August 2, which had been deleted entirely since all its hooks had been promoted/rejected, but that was an easy diff. Unfortunately, I can't give separate figures for the still in review vs. approved vs. rejected, and the one problem with the brute force method is that it doesn't know about special occasion hooks that were removed from their original date. There aren't a lot of them, so they wouldn't have significantly changed my numbers. BlueMoonset (talk) 19:54, 14 September 2014 (UTC)[reply]

Just to clear up one point BMS makes: the number 300 has absolutely nothing to do with this. That large number means only that there are a lot of noms stuck awaiting approval, but the only thing to do about that is crack the whip on reviews. Unless and until those noms end up in the "approved reserve" (as I call it), this has zero to do with the appropriate "consumption rate" (as, again, I have called it). [Lecture resumes]

  • BMS' nomination figures clarify things marvelously. His number of 17 noms/day includes a few that will end up rejected/withdrawn, but on the other hand it looks like nomination rates have been going up recently, in desperation let's just call it 17 noms arriving per day. That's somewhat more than what's needed to sustain 2 sets of 7 = 14/day, but not enough to sustain 3 sets of 7 = 21/day. So an alternation between 2 sets/day and 3 sets/day is the right strategy. But of course we knew that already.
  • The question, then, is when to trigger the switches back and forth. If we switch to 3/day right now, the approved reserve will start at 4 days, and rapidly become 3, then 2, then 1 -- then suddenly we'll be having an emergency debate on going back to 2/day again. That's no way to run a railroad.
  • Let me suggest two really nice, round numbers which are easy to remember, and safe against the approved reserve running dry -- but not excessively so: 100 and 150. That is:
  • Keep running 2 sets/day until the approved reserve hits 150 or above.
  • At that point switch to 3 sets/day, until the approved reserve hits 100 or below, at which point we go back to 2 sets/day.
  • Notice the trigger points are such that drift from 7 days' reserve up to 11 days' reserve during the 2/day period ("on the way up"), then from 7 days' reserved down to 4 days' reserve during the 3/day period ("on the way down").
  • I know these numbers will seem excessive to the "rushers", but the key number in the above is the minimum 4 days' reserve. I submit that 4 days' reserve is really the least we should want to plan to have ever.
  • This yields a very stable process with changes to the # sets/day only every few months, and we would never, ever run out of hooks. If we can agree on the 100/150 trigger points now (at least provisionally) then we don't have to have a debate every time -- when the approved reserve hits a trigger point, whoever notices that can just change the bot parameter, no questions asked.

Informal tracking of the stats

  • Not to worry, there will be plenty of opposition to a suggestion of 1 set/day with 300+ active nominations. For me, I'd want to see total nominations at fewer than 150 and a single-figure nomination rate before entertaining the extreme suggestion of a single daily set. BlueMoonset (talk) 05:15, 18 September 2014 (UTC)[reply]
MBS, you're a smart guy (or gal -- never did know which). But you keep missing the point. It doesn't matter how many noms are awaiting approval -- all that matters is the size of the approved reserve, because we can't put noms-awaiting-approval on the MP. EEng (talk) 02:03, 19 September 2014 (UTC)[reply]
EEng, I'm not missing the point. I'm just aware of more variables. We can vary the size of the sets up or down, which allows for fine tuning; if we had (say) 45 approved hooks in the reserve, we could still easily fill two sets a day, since approvals will continue to occur. At the moment, you're one guy (or gal?) with a new theory that's suddenly the be-all and end-all, while there's years of DYK experience that you're apparently unaware of. BlueMoonset (talk) 07:12, 21 September 2014 (UTC)[reply]

Resuming normal, non–lecture-format debate

  • No need to delay. The whole purpose of a backlog is to build up enough noms so that we don't run out. Unless this wasn't the case before August 2012, having three sets a day has always been the norm since I've been nominating DYKs. Reducing to two sets is only utilized as an emergency measure when we get below 150–200 noms. Now that we're at over 300 noms (and 81 approved), the backlog has clearly served its purpose already. Time to return back to normalcy. —Bloom6132 (talk) 19:39, 14 September 2014 (UTC)[reply]
  • Yes, if "normalcy" is posting garbage to the main page and getting dozens of error reports and allowing hooks to be re-written on the fly while contributors sleep. Slow it down. Current is good. Your version of "normal" is simply not good. The Rambling Man (talk) 19:55, 14 September 2014 (UTC)[reply]
  • As I replied earlier above, I still see hooks being pulled here and there at the same pace they were back when there were 3 sets a day. It makes no difference whether it's 2 or 3 sets. The end result is exactly the same – inferior noms by inexperienced users. Hence, your claim correlating slow pace with quality control is iffy at best… —Bloom6132 (talk) 20:26, 14 September 2014 (UTC)[reply]
  • The numbers do not agree with your observations. Compare the one-month periods between 11 June and 11 July (3 sets/day), versus 19 July and 19 August (2 sets/day), at Wikipedia:Did_you_know/Removed: 44 pulls for a month of 3 sets/day, 5 pulls for a month of 2 sets/day. Granted there are likely other factors contributing to this, but it would seem there is in fact a very strong correlation in recent months between higher run rate and higher pull rate. Nikkimaria (talk) 21:57, 14 September 2014 (UTC)[reply]
  • There's no point in comparing stats when you skew them to your favour. You conveniently left out April 2014 (which was 3 sets a day and only 13 pulls) and February 2014 (also 3 sets/day with just 14 pulls). And there would be no pulls for a month of 0 sets/day. So the stats you provided are basically meaningless, because it's obvious the more sets run, the more pulls there will be. Correlation does not imply causation. —Bloom6132 (talk) 22:19, 14 September 2014 (UTC)[reply]
  • Which is of course why I said only that there was a correlation. I chose the stats period from looking at the history of rate changes - selecting a period of one month starting the day after the most recent change to each rate - before looking at the removals list, so your accusations of deliberate skewing are out of line. Please dial it back a bit. Nikkimaria (talk) 22:43, 14 September 2014 (UTC)[reply]
  • "[Y]our accusations of deliberate skewing are out of line" – let's see. You claim that the "numbers do not agree with [my] observations". Then, you choose a one-month period with 3 sets that is a complete anomaly (there's no other way to describe it) to all the other DYK months in 2014. But you don't disclose the two months with 3 sets that do agree with my observations (and rebut the anomalous month example you gave). When I (correctly) call this out as skewing the stats in your favour, you call my accusations "out of line". Reality check – the only thing "out of line" here is your picking and choosing stats that suit your (and TRM's) argument but don't give the full perspective of the actual situation. It presents an incomplete and inaccurate one-sided picture which, if I hadn't probed deeper into these stats, would have misled others to believe that your argument was correct. Clearly that's not the case. —Bloom6132 (talk) 23:40, 14 September 2014 (UTC)[reply]
  • "I still see hooks being pulled here and there at the same pace they were back when there were 3 sets a day. It makes no difference whether it's 2 or 3 sets" as a claim is not compatible with "it's obvious the more sets run, the more pulls there will be", nor is it compatible with recent data. If you want to look at larger trends to get the "full perspective of the actual situation", it is indeed obvious that more sets per day equals more pulls, not to mention more late-update warnings. Nikkimaria (talk) 02:53, 16 September 2014 (UTC)[reply]
  • Actually Bloom6132, I make my observations based on the number of ERRORS and pulls I'm seeing. Subjectively it seems that the quality has increased substantially and that I rarely have to attend to major DYK errors on the main page. This is because there's more time for some of us to wander past the queues because the mad rush to cycle DYKs round has slowed a little. There is absolutely not one single advantage to rushing these to the main page. So you would claim "backlog!", I would suggest this is a good investment in the future of DYK. Keep allowing us time to make sure the quality is on the increase, and continue the education of those reviewers who clearly have something to learn about quality control in Wikipedia. The Rambling Man (talk) 17:04, 16 September 2014 (UTC)[reply]
Speaking as only a sometime contributor to DYK (I have maybe about 50 or 100 — we temporarily had the record for most hooks in one article at one time — I thought a whole day was good. Running them through just to get them out of the line doesn't do the articles justice in terms of visibility. Who (and how many) is looking at these articles on the midnight shift {{yes, I know we are global)? I daresay that most people don't constantly monitor the main page. I understand the need to move the queue, but that can't be the only consideration. 7&6=thirteen (☎) 19:25, 14 September 2014 (UTC)[reply]
If this were some sort of perfect world, every article would get its own little feature on the main page. It isn't. DYK is set up to give a brief, hopefully memorable mention for a variable number of hours on the main page. It isn't ideal that it can't be a whole day, or even necessarily half a day—I've had some hooks that were up for six hours, when we had to go to four sets a day—but the hook is there and can be seen by anyone who stops at the front page for that period. If we don't "run them through" at the rate they're coming in, then eventually we have 400, 500, or more hooks backed up—300 was a rare event in my experience until very recently. Something has to give if we aren't to have an ever-inflating backlog: either we have to bar the door to newcomers, or increase the flow of hooks going to the main page. Three times a day was the standard when I first started coming here, occasionally varying to two or four times when necessary. We write articles, nominate them for DYK, and take our chances as to when and how long the nomination will be seen. BlueMoonset (talk) 20:58, 14 September 2014 (UTC)[reply]
There's no rush. 12 hours for a hook is a fine amount of time as it allows most of the world an opportunity to see it and those who contributed to enjoy the fact they can see it on the main page. The reserved turnover rate also allows queues and prep areas to be patrolled more thoroughly and reduce the errors (and major embarrassments that DYK has provided Wikipedia in the past). Slow and steady, remember? The Rambling Man (talk) 21:05, 14 September 2014 (UTC)[reply]
My comment was not meant as an implicit criticism of the administration of DYK. I simply wanted to suggest that there are countervailing considerations, and y'all have to do the best you can to weigh them in the balance. But assuredly there is a balance and a trade off no matter what you do. That's all. 7&6=thirteen (☎) 22:06, 14 September 2014 (UTC)[reply]

I suspect the nomination increase is caused in part by the Stub Contest that is running in September, as the rules for stub expansion make every entry eligible for DYK as a 5x expansion (unless they've appeared before), and if the competitors are new or reviewing each other's entries as QPQ the number of unreviewed hooks would go up (I've done extensive research into this by imagining it as a cause; it's a poissin distribution with a standard deviation of a T-test; there, statistic that!). I don't see any reason not to go to three sets a day if the queues and prep areas are kept full, but since it seems that swapping between 2 and 3 sets a day requires days of debate rather than a flick of a switch, it is probably best to stay with two. Belle (talk) 11:56, 15 September 2014 (UTC)[reply]

Worst hook in the history of DYK

"...that ketchup was originally prepared using mushrooms as its main ingredient, and was sometimes referred to as mushroom ketchup?" currently sitting in Prep 4 has to be the worst hook I have ever seen. Seriously. And according to the ketchup article, is actually factually incorrect. The Rambling Man (talk) 11:28, 13 September 2014 (UTC)[reply]

  • You and I both know (well, I should hope that we do) that if the original hook that was promoted is reinstated, it is likely that the hook will be pulled anyways owing to concerns over its accuracy. Might exacerbate the situation, but I don't doubt that it will happen. — Crisco 1492 (talk) 13:59, 13 September 2014 (UTC)[reply]
  • The ketchup article needs to be edited for accuracy, not vice versa. Mushroom ketchup is written entirely from sources. I have copy edited the ketchup article for clarity, also adding inline templates requesting further verification within its lead, where clarification is needed. What part of the ketchup article contradicts this article at this time? Please review and compare both articles, as the supposed contradiction stated above isn't qualified by specifically stating how. NorthAmerica1000 14:07, 13 September 2014 (UTC)[reply]
  • Sources which may, themselves, be inaccurate. — Crisco 1492 (talk) 14:32, 13 September 2014 (UTC)[reply]

Perhaps redo the original hook as ..."that ketchup was originally prepared with mushrooms?"--¿3family6 contribs 14:22, 13 September 2014 (UTC)[reply]

Well said. 7&6=thirteen (☎) 14:24, 13 September 2014 (UTC)[reply]
Sorry, but no - there are sources that state that ketchup was originally made in the far east, from fish sauce - and explain the etymology. [1][2] Given such contradictory evidence, we shouldn't be making definitive assertions about this, especially on the main page. AndyTheGrump (talk) 14:29, 13 September 2014 (UTC)[reply]
And yet another source... yeah, questionable assertion at best. — Crisco 1492 (talk) 14:37, 13 September 2014 (UTC)[reply]

Thanks to users for qualifying the stance of inaccuracy. I wasn't specific enough in devising the hook, which made it inaccurate. A simple way to modify the hook is to reword it, adding accurate qualification with, "in the United Kingdom". See example below. NorthAmerica1000 14:47, 13 September 2014 (UTC)[reply]

How is a statement that mushroom ketchup was originally made from mushrooms even remotely interesting? AndyTheGrump (talk) 14:50, 13 September 2014 (UTC)[reply]
That can be worded better. Here's something below. NorthAmerica1000 14:57, 13 September 2014 (UTC)[reply]

7&6=thirteen (☎) 14:59, 13 September 2014 (UTC)[reply]

  • Early preparations prepared... gosh. How about
 — Crisco 1492 (talk) 15:07, 13 September 2014 (UTC)[reply]

There we go!--¿3family6 contribs 15:09, 13 September 2014 (UTC)[reply]

Homemade mushroom ketchup

I prefer ALT5 too, it seems factually correct and doesn't contradict the ketchup article (which User:Northamerica1000 has stated needs fixing). But in future, and as AndyTheGrump has reinforced above, I hope reviewers remember that a hook along the lines of "... that A of B is B from A?" is in no way interesting. The Rambling Man (talk) 18:02, 13 September 2014 (UTC)[reply]

ABBA tomato sauce? Martinevans123 (talk) 18:08, 13 September 2014 (UTC)[reply]

DYK Toolbox needs an edit

Orlady, Crisco 1492, Mandarax, BlueMoonset - I'm hoping one of you can edit Template:DYK tools. I'm a little hesitant I might not get it correct on this one if I try. In a nutshell, Dispenser's tools for Dablinks and External links are not functional and may not be in the foreseeable future. There is a Village Pump discussion about replacements. Therein, it suggests Extension:Disambiguator does the same thing as the Dablinks tool. Can any of you make the change on the DYK toolbox? — Maile (talk) 22:02, 13 September 2014 (UTC)[reply]

Thanks. — Maile (talk) 14:04, 14 September 2014 (UTC)[reply]

Removed from Prep 1

@Shhhhwwww!!, Cwmhiraeth, and Hawkeye7: I have removed "* ... that the first Bay Church was made of bamboo and nipa and was built along the lake shore of Laguna de Bay, Philippines?" from prep 1. The hook was sourced to one unavailable page and one wordpress page that had copied the info from Wiki Pilipinas. Worse, it looks quite likely that the info is simply wrong, and that people have been confusing the St. Augustine Parish Church (Laguna) with the San Agustin Church (Manila). This doesn't seem fit to be on the main page. And that, sadly, seems to be happening quite often again these days. Fram (talk) 09:57, 15 September 2014 (UTC)[reply]

  • No one claimed that the Manila church was ever nominated (unless Hawkeye was referring to that, no idea what he meant). The churches are two different churches, but it seems as if somewhere along the way the nominated Bay Church has appropriated the early history of the much more important Manila church. Note that already the second sentence of the article, "It served as an old Franciscan mission town in 1578. " is wrong, as it was supposedly Augustinian at the time, and only became Franciscan in 1737, or nearly 150 years later... Sourcing and fact checking of the article don't seem to be up to main page standards. Fram (talk) 14:56, 15 September 2014 (UTC)[reply]
  • Thank you for pointing out several errors. I hope so that the comments was also written on the article's talk page to address it immediately for revision. It was an honest mistake to mixed up the Franciscans and Augustinians in the lead paragraph. Changed the second sentence to "It was first administered by Augustinian priests and later transferred to Franciscans." Changed also the references to more credible one like the book of Huerta and the Historical Marker of the church. If it does not meet the DYK criteria, well then, close the nomination and fail it. We'll respect it. --Carlojoseph14 (talk) 15:29, 15 September 2014 (UTC)[reply]

DYK is almost overdue

In less than two hours Did you know will need to be updated, however the next queue either has no hooks or has not been approved by an administrator. It would be much appreciated if an administrator would take the time to ensure that DYK is updated on time by following these instructions:

  1. Check the prep areas; if there are between 6-10 hooks on the page then it is probably good to go. If not move approved hooks from the suggestions page and add them and the credits as required.
  2. Once completed edit queue #5 and replace the page with the entire content from the next update
  3. Add {{DYKbotdo|~~~}} to the top of the queue and save the page

Then, when the time is right I will be able to update the template. Thanks and have a good day, DYKUpdateBot (talk) 10:09, 15 September 2014 (UTC)[reply]

Why approve one hook and promote another? Pulled one from queue...

Template:Did you know nominations/Mahikeng Airport

@Jakec and Hawkeye7:, one hook was approved at the nomination, but another one taken for the promotion. This means that an unpproved hook was about to hit the main page. One I have trouble to find in the sources, I have to say (the "regain" part, which means it once was an international airpport). In any case, it isn't useful to have different checks balances when the promotor decides to take whatever hook he likes anyway... Fram (talk) 15:07, 15 September 2014 (UTC)[reply]

When I saw such a thing, I just changed prep, using the approved hook. --Gerda Arendt (talk) 15:17, 15 September 2014 (UTC)[reply]
In the promoter's defence, this was changed after it was promoted [3] and if the hook fact is supported I think the change makes a better hook than than the original. Belle (talk) 15:20, 15 September 2014 (UTC)[reply]
The promotor didn't use the hook that was promoted, then someone else changed the hook in the queue to yet another one; perhaps we can just abandon the whole review process, and give everyone the right to put things straight into the prep areas? The hook that was in the queue may be better and may be correct and sourced, but we have a process to decide and determine such things, to avoid too many mistakes from reaching prep, queue and main page. Fram (talk) 15:35, 15 September 2014 (UTC)[reply]
I only asked for the right to replace an unapproved hook by the approved one. --Gerda Arendt (talk) 15:39, 15 September 2014 (UTC)[reply]
I wasn't responding to you, but to Belle, apologies for any confusion. Fram (talk) 15:57, 15 September 2014 (UTC)[reply]
Gerda Arendt, if it's your own nomination or your own hook, then you don't have that right. Please ask here if there's an issue with a hook promotion from your own nomination; I'm sure someone will be happy to check and make an appropriate fix. Thanks. BlueMoonset (talk) 16:53, 15 September 2014 (UTC)[reply]
I exchanged one image, - should I have bothered someone else? --Gerda Arendt (talk) 17:18, 15 September 2014 (UTC)[reply]
Yes. Please do next time. Thanks. BlueMoonset (talk) 17:38, 15 September 2014 (UTC)[reply]
You seem to be serious about bothering another person for using the cropped version? --Gerda Arendt (talk) 21:30, 15 September 2014 (UTC)[reply]
  • When Hawkeye7 promoted this, neither hook had been struck out. Reviewer Jakob never mentioned the original hook on the review, which made the review incomplete and nobody catching that fact. The review mentions no details, such as a check for close paraphrasing/copyvio. "Sufficiently expanded, meets core content policies. ALT1 was supported by sources, so let's go with that." is a pretty surface skimming review. Nominator Nathan121212 never mentioned a preference on a hook after the review was done. More than one unstruck hook is promoter's choice. But no way should anybody just jump in to a Prep area or Queue and create an entirely new, unapproved and unreviewed hook. As pointed out above, it was EEng who changed it in Prep1 with no discussion on the template or anywhere else. — Maile (talk) 17:50, 15 September 2014 (UTC)[reply]
I changed the hook from
... that Mahikeng Airport was an air force base and is now trying to become an international airport?
to
... that efforts are underway to help Mahikeng Airport, a former air force base, regain its status as an international airport?
because (as my edit summary stated)
airport "trying" is a bit too anthropomorphic, and per article it was once an intl airport before
i.e. the promoted hook was borderline illiterate, and didn't match what the article said. I have said over and over and over and over that the final, perfect, all-it-needs-is-to-be-copied-to-prep version of the one-and-only, positively this-is-the-one-to-use hook should be explicitly agreed upon on the nom page (maybe with the required participation of a "senior reviewer") before the nom is closed, so that things like this can be handled where they should be handled -- on the nom page. But no one seems to want to do that, so often they've gone to prep before anyone's noticed some problem.
I make some change to maybe 1/3 of hooks while they're in prep, with almost zero pushback and plenty of Thank-You pings (not to mention the many, MANY changes and fixes I make on hooks in noms still in progress -- with even more Thank-You pings, and lots of "Thanks for a much better hook" as well) -- you want me to log a REMOVED and reopen the nom page every time? This was one of the few that wasn't strictly a copyedit or grammar/mechanical fix, but it changed a hook that didn't match what the article says, to one that does. (Maybe this one I should have REMOVED and reopened -- so sue me.) If you don't like that, then I hope you'll join in getting procedures changed so that hooks are final before the nom is closed.
EEng (talk) 18:17, 15 September 2014 (UTC)[reply]
FWIW, and IMHO, the original hook was a tragedy so EEng's choice was a vast improvement. The Rambling Man (talk) 18:37, 15 September 2014 (UTC)[reply]
Gratefully (and somewhat ashamed) I agree. ("Borderline" damn I'm getting close.) Nathan121212 (talk) 18:48, 15 September 2014 (UTC)[reply]

Every author, however modest, keeps a most outrageous vanity chained like a madman in the padded cell of his breast.

Logan Pearsall Smith (1931). Afterthoughts.

Don't you mean ashamedly, you borderline illiterate? ;) But seriously, I was too harsh. Let's just say The airport is trying sounds very odd. <begin musing> Strangely, had it said "Berlin is cracking down on..." I wouldn't have objected. And we routinely hear that "Cuba refused to recognize..." and so on, and that sounds OK. I think it's something to do with the fact that we don't think of an airport as an active entity -- if it had said "the XYZ Airport Authority is attempting..." that would sound right. And yet so would "XYZ Airport no longer allows small planes to land". Sometimes the ear knows what the mind can't really explain. <end musing>
In closing, let me use this discussion as an excuse to trot out one of my all-time favorite aphorisms -- see the box.
EEng (talk) 20:04, 15 September 2014 (UTC)[reply]
"the airport is trying" = 259,000 Google hits?? Martinevans123 (talk) 21:45, 15 September 2014 (UTC) [reply]
The Rambling Man says my hook is better. You dare to disagree? EEng (talk) 03:21, 16 September 2014 (UTC)[reply]
  • All I can say about promoting the wrong hook occasionally is that the prep area assembly process involves first selecting seven or more nominations, then I have to come back and move them to the prep area. At that point I can only see the markup, which can sometimes makes it hard to find the hook at all. But this is not part of the review process, it is the prep area assembly process. If the wrong hook is promoted, or even if someone thinks that another alternate hook would be better, then anyone is free to swapo the hooks. This is not overriding the reviewer, it is part of the prep area assembly process. Moreover, The rules encourage the kind of copyediting grammatical tweaks that EEng is talking about. They also recommended trimming back necessary verbiage. So long as the meaning is not changed. Hawkeye7 (talk) 20:11, 15 September 2014 (UTC)[reply]
Well, I for one think we should draw the line at trimming back necessary verbiage. And sentence fragments. EEng (talk) 21:25, 15 September 2014 (UTC)[reply]
  • Often, though, the meaning is changed while DYKs sit in the queues, but mostly this is a good thing. The benefit of the slow down in DYK promotion to main page is that some of us get a chance to look over the hooks properly before they're posted. DYK errors are decreasing, quality is increasing. Keep it going, and don't be afraid to modify or pull hooks that seem dubious. Better to pull now, ask questions later than to let erroneous material get to the main page. The Rambling Man (talk) 20:14, 15 September 2014 (UTC)[reply]
Well, while mods in prep/Q are often our only recourse for now, I repeat it would be far preferable for hooks to be completely finalized "in public" -- by which I mean on the nom page. But such a process would have to include some provision for making sure the select few have a final opportunity to fix the problems most don't see. EEng (talk) 21:25, 15 September 2014 (UTC)[reply]

Old nominations needing DYK reviewers

I've compiled a new set of the 36 oldest nominations that need reviewing; at the moment, 278 of 352 nominations are unapproved. Thanks as always to everyone who reviews.

Please remember to cross off entries as you finish reviewing them (unless you're asking for further review), even if the review was not an approval. Many thanks! BlueMoonset (talk) 20:14, 16 September 2014 (UTC)[reply]

What I've Been Looking For needs review

What I've Been Looking For, about a Disney song, was nominated July 29. There's been chatter on the nomination template, but nothing in the way of a real review. It's not listed above as an older nomination, but it sure needs a review. — Maile (talk) 14:07, 17 September 2014 (UTC)[reply]

ok hang on. Cas Liber (talk · contribs) 12:28, 18 September 2014 (UTC)[reply]
Thanks for doing this. — Maile (talk) 15:10, 18 September 2014 (UTC)[reply]

Template:Did you know nominations/Cross-border Terminal, Tijuana International Airport

Cross-border Terminal, Tijuana International Airport has a stack of text created as it was imported from Tijuana International Airport, however that material was created not long before. So even though the material was imported...it is still new and hence qualifies...right?...or not? Cas Liber (talk · contribs) 12:25, 18 September 2014 (UTC)[reply]

I see what you're getting at, but it seems legit to me. The content was added, and then on second thought, moved to a new article. I might have some concern (but limited for sure) in the case when older content was moved into a new article, but in this case I don't see an issue. Maury Markowitz (talk) 15:46, 18 September 2014 (UTC)[reply]
I agree that it qualifies as new: we've had cases before where a large amount of material was added to an existing article, and then removed into its own new article with a couple of days (well within the "new" deadline). I see some other issues with the article that need to be dealt with, but in terms of newness and size, it definitely qualifies. BlueMoonset (talk) 20:44, 18 September 2014 (UTC)[reply]

Fix the tools?

The DYK review page currently list three tools in a little box on the right: DYKcheck, Dup detector, QPQcheck. None of these actually "work". The first takes you to the description page of the tool, not the tool. The second takes you out-of-page and then requires you to to all of the URLs and kinda already know if it's a dup, and the third asks you to type in the username, which is on the page you just left.

So, is there any reason that DYKcheck can't go directly to the page and invoke the tool?

And the same for QPQcheck?

Is there any purpose to Dup detector at all? If I think it's a dup, don't my eyeballs do the same thing as this with a lot less typing?

Maury Markowitz (talk) 15:43, 18 September 2014 (UTC)[reply]

Maury, they're helpful links. DYKcheck takes you to a page that explains what the check does and how it can be used: very helpful for the new DYK reviewer (there are more all the time). Duplication detector takes you to the tool's wmflabs page, and I've never had to type at all: copy the article url, copy a source url, paste each onto the form, and bingo: a list of matching strings of words. Eyeballs take far longer to scan both documents, especially if the source is at all long, and especially if there are a lot of sources to check. It has found copyvios and close paraphrasing more times than I care to say, including cases where it's close paraphrasing not on the cited source but a copyvio of a source cited elsewhere in the article.
Once DYKcheck is installed, the user has it in their Tools section of the left-hand Wikipedia menu, so you invoke it whenever you're looking at an article you want to check. (It's two clicks from the nomination template, not one, but you should be looking at the article anyway as part of your review.) For QPQcheck, right now, that tool seems to be hit-or-miss: it shows that I have 34 DYKs, but doesn't seem to see any of yours, not even the one on your Talk page. (It may not be able to deal with your talk page archiving scheme, which strikes me as a weakness of the tool.) BlueMoonset (talk) 20:40, 18 September 2014 (UTC)[reply]
"DYKcheck takes you to a page that explains what the check does and how it can be used"
Which is completely useless, it doesn't actually do anything. Why does it not simply run the tool?
"copy the article url, copy a source url, paste each onto the form"
A perfect description of why this tool is useless. BM, how does one find if there is a copyvio if you need to know the page that it vios? And even if you torture logic to try to answer that, why doesn't it at least fill out the first URL?
These need fixing, not apologies. Who do I talk to? Maury Markowitz (talk) 13:00, 19 September 2014 (UTC)[reply]
Thanks Maile! Maury Markowitz (talk) 15:06, 20 September 2014 (UTC)[reply]
There are technical limitations here that would make such an implementation flaky at best. The main limitation is that DYKcheck can't be run by an anonymous editor. This wasn't true years ago, but somewhere along the way MediaWiki changed so now users must be logged in to use the script. Aside from this limitation, it is possible to tweak DYKcheck to allow a link on the nom page to invoke it. This sort of a tweak isn't trivial though. To be clear, DYKcheck is a script and not a tool run on an external server. With a tool on an external server it's much simpler to create links to invoke it from wherever regardless of user state. Shubinator (talk) 03:35, 21 September 2014 (UTC)[reply]
And just so those editors know what you are talking about, it's when a reviewer clicks on a tool in the upper right hand tool box of the Template:Did you know nominations, you believe clicking on the links should be all the reviewer should have to do. No fill in the blanks or anything else. It should just take the reviewer to the complete information. This is the DYK toolbox. — Maile (talk) 18:41, 20 September 2014 (UTC)[reply]

Removed one hook from Queue5

Template:Did you know nominations/Martin Manulis

@Cbl62, Cwmhiraeth, Hawkeye7, and Casliber:

  • ... that Martin Manulis was the producer of Playhouse 90, which was voted the greatest television series of all time in a 1970 poll of television critics?

This isn't supported by the source, [4]. Yes, I could have corrected the hook. And then I would have done what I object to in others, changing facts in hooks while bypassing the whole review process. I have no objection against making hooks better without changing the meaning (e.g. adding wikilinks, changing some punctuation or capitalization, whatever). But a wrong hook should be sent back to the nominations page. Fram (talk) 07:30, 19 September 2014 (UTC)[reply]

I think it is supported in this source [5] which is where I saw it. Cwmhiraeth (talk) 07:49, 19 September 2014 (UTC)[reply]
Then you think wrong. The hook is similar to a statement in the NYTimes article, but not the same. That you overlooked the difference in your review is not good. That you overlooked the difference even after the hook was pulled for being wrong makes me wonder whether you should review any hooks. For crying out loud, the hook is one sentence, 26 words. Fram (talk) 07:58, 19 September 2014 (UTC)[reply]
Aah, yes it is "editors" and not "critics" - changed now, given I reworte the hook, does someone else want to assess? Cas Liber (talk · contribs) 08:22, 19 September 2014 (UTC)[reply]
Don't you think editors are critics of one another's works? ;-) Cwmhiraeth (talk) 09:28, 19 September 2014 (UTC)[reply]
Some of our editors clearly not. Fram (talk) 09:31, 19 September 2014 (UTC)[reply]

Two hooks

Template:Did you know nominations/Monsieur Léonard started as a double nom. We arrived at separate statements about the two people. I have no idea how that may be processed further, --Gerda Arendt (talk) 22:45, 19 September 2014 (UTC)[reply]

Gerda, the promoter who picks one of the two hooks to promote should strike it but not close the template (should not substitute it or fill in "passed"); the promoter who picks the remaining hook to promote should close the template in the normal way. However, I found an issue with ALT2b, and have proposed a slightly changed ALT3b to deal with it; you'll want to review it because I had to supersede your AGF tick with a ?. BlueMoonset (talk) 23:30, 19 September 2014 (UTC)[reply]

Queue 5

I've noticed that in the lead article of this queue, there's an unreferenced paragraph in the bottom of this section. There's also another unreferenced sentence in the final section of that article. Since DYK rules suggest one citation per paragraph, these should be fixed before it moves to the Main Page. Pinging people involved: @Peak Player, Oreo Priest, and Maury Markowitz:. --Jakob (talk) 12:04, 20 September 2014 (UTC)[reply]

Your concerns have been addressed.Peak Player (talk) 13:44, 20 September 2014 (UTC)[reply]
And it's not a rule anyway. Queue outraged comments.... NOW! Maury Markowitz (talk) 15:15, 20 September 2014 (UTC)[reply]
This is true. Supplementary Rule D2: "The article in general should use inline, cited sources." I've always taken this to mean than an unreferenced sentence or paragraph is okay, but an article unreferenced except for the hook is not. Hawkeye7 (talk) 19:11, 20 September 2014 (UTC)[reply]
D2 does continue, A rule of thumb is one inline citation per paragraph, excluding the intro, plot summaries, and paragraphs which summarize other cited content. So it depends on what isn't sourced, but the rule of thumb gives guidance as to the expected level of sourcing for DYK. BlueMoonset (talk) 02:59, 21 September 2014 (UTC)[reply]
Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia_talk:Did_you_know&oldid=626443858"