Jump to content

Wikipedia talk:New pages patrol/Reviewers/Archive 8

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 5Archive 6Archive 7Archive 8Archive 9Archive 10→Archive 15

Who is doing the reviewing?

Well it looks to me as if about 90% of the reviews are being done by 10% of the Reviewers - and it's generally the same reviewers. So what are the other 360 reviewers doing? Kudpung กุดผึ้ง (talk) 15:08, 12 June 2017 (UTC)

Nothing. I suspect it's similar to AFC, which I've noticed that people are very active for a short time, then get bored/tired/sick of it and go on to other things. We drop about 15 editors a month from the AFCH list for inactivity; maybe it's worth implementing a similar "lack of activity" boot here as well (at least so the numbers aren't as skewed). and I suspect that list would have a similar activity percentage if it weren't trimmed regularly. Primefac (talk) 15:48, 12 June 2017 (UTC)
Primefac, that'd likely require an RfC since it is a user right, and would also play into the report's false idea that we want less people reviewing. We have other things that will need an RfC soon, and in my mind it's not worth expending so much of the community's energy on the user right when there are other things in the process that need to be dealt with. TonyBallioni (talk) 16:03, 12 June 2017 (UTC)
Fair point, wasn't thinking about it in terms of PERMs. Refactored comment above. Primefac (talk) 16:05, 12 June 2017 (UTC)
I agree with Primefac's non refactored opinion above. However, I also agree something should be done regarding inactive reviewers, and getting active reviewers. —usernamekiran(talk) 16:48, 12 June 2017 (UTC)
I posted some stats on the NPP Analysis and Proposal page, with a table of active reviewers and reviews per month. This was in the mid 900s up until November 2016, when it went down to the mid 300s. But there's currently around 350 unique people who take at least one review action a month. DannyH (WMF) (talk) 17:05, 12 June 2017 (UTC)
@Kudpung: As I've said before, working AfC and NPP results in pushback. The more pages I review, the more chattering I get from some unhappy party because I nominated something for deletion, turned an article back into a redirect, or declined a draft. Surely that's not just me. The backlog is self-regulating because the editors ensure I can't simply click-through to eliminate it; I only keep pace with it (if that). Maybe you like getting arguments from other editors; I don't. I want them all to go away so I can edit in peace. If I could just hand out blocks anytime I wanted to, I'd have the backlog gone by the end of the week. Chris Troutman (talk) 04:47, 13 June 2017 (UTC)
@Chris troutman: I agree it's a pain in the butt to see your edits reverted by angry editors, get insulted, hounded and generally harassed. The trouble is that if you don't reply and follow up then new pages reviewing is pointless and discredited. It's easy just to stay on the uncontoversial articles and add a "more references tag" and move on to the next one. And it's not just the newcomers that don't understand what you're trying to do sometimes editors with tens of thousands of edits say things like "Keep NASA astronauts are inherently notable," from a guy with 20k edits or this one from a a guy with 76k edits a a whole list of rights including autopatrolled, pending changes, rollbackers when I said that you can't just presume that that the sources exist they have to be cited in the article itself otherwise notability is not proven his reply was "No, that is not correct. It is sufficient for the material to exist. The actual condition of the article is immaterial. Wikipedia:Deletion is not cleanup"...how to counter an editor who is here for over 11 years and very very active but refuses to recognise the GNG? Domdeparis (talk) 06:29, 13 June 2017 (UTC)
  • DannyH (WMF): '....a table of active reviewers and reviews per month. This was in the mid 900s up until November 2016, when it went down to the mid 300s.' - Could you please refrain from constantly using that statistic? It's clear to anyone that with the introduction of the new user right, the sharp rise since July was actually halted dead in its tracks and the curve flattened out and even began to drop until I announced my retiremement from actively 'managing' NPP. Thank you. Kudpung กุดผึ้ง (talk) 08:06, 13 June 2017 (UTC)
Backlog May 2017. Graph: Kudpung
Backlog May 2017. Graph: Kudpung
The question that's being asked is how many people are reviewing pages? The answer is about 350 people a month -- down quite a bit from what it was six months ago, but significantly more than the 40 that you suggested at the top of the thread. DannyH (WMF) (talk) 16:57, 13 June 2017 (UTC)
DannyH (WMF) What I said at the top of the thread was: Well it looks to me as if about 90% of the reviews are being done by 10% of the Reviewers - and it's generally the same reviewers. So what are the other 360 reviewers doing? - To suggest that 350 are patrolling is deceptive when we all know from a script that the WMF itself has published that most of those 350 only did single numbers of patrols. Kudpung กุดผึ้ง (talk) 11:51, 14 June 2017 (UTC)
The occasional missing fortnightly figures don’t change the shape of my graph one iota. On July 13 the backlog was just over 5,000, by November it was 16,000. Anyone can see that 16K is a much bigger number than 5K. There is also no denying that the figure remained fairly flat for a while - with even a modest decrease - but then from February to 10 May it had risen again to a new all-time high of 22,000. Whether it's a logarithmic curve that joins the dots, a bar chart, or just expressed in words, it isn’t going to change the facts. With its 300 employees, why hasn't the WMF got a department of dedicated expert statisticians, who uninvolved with the issues they are quarrying figures for, can come up with neutral reports uncoloured by their Wiki-political convictions? Kudpung กุดผึ้ง (talk) 04:51, 15 June 2017 (UTC)
I have to agree with Kudpung. The depth of this aspect of the analysis is extremely shallow and lacking in transparency. The metric 'reviewers per month' is not very useful at all. I think it's unfortunate that WMF is not more helpful in this regard.- MrX 12:07, 15 June 2017 (UTC)

Some stats

I did some digging around after coming across this question on Quora. I came to a conclusion that only about 16.7% articles on Wikipedia have been marked as reviewed. Please see the calculations here. For those too lazy, there are about 5,422,965[1] articles on Wikipedia and only about 905695[2] have been marked as reviewed. Don't know if it helps but thought it wise to share it here as well. Yashovardhan (talk) 19:14, 12 June 2017 (UTC)

User:Yashovardhan Dhanania - Is that because articles created before reviewing was implemented, or implemented in its current fashion, are listed as not reviewed, when in fact they are de facto reviewed (grandfathered)? When was the current concept of reviewing implemented? Robert McClenon (talk) 20:41, 16 June 2017 (UTC)
Ofcourse @Robert McClenon: This is only the list of articles marked as reviewed using the page curation tool. I got this data from MusikAnimal's tool which tells the no. of reviews by month. The other articles (created earlier) seem to have been grandfathered in but are not marked as reviewed de facto. Yashovardhan (talk) 12:05, 18 June 2017 (UTC)

References

  1. ^ WP:STATS
  2. ^ added everything at MusikAnimal's. "New page reviews by months". WMF labs - Quarry. {{cite web}}: |author= has generic name (help)

Looks like WP is suffering the after effects of unsustainable development. Hopefully, WMF will give serious consideration to some of the suggestions that have been submitted to help reduce the backlog. Atsme📞📧 00:20, 14 June 2017 (UTC)

What are those suggestions? Also, how do you define unsustainable development? If article pages are being created at a rate that is faster than they can be reviewed, or if more article pages are being created than are being watchlisted and maintained effectively, what is the definition of unsustainable development? Robert McClenon (talk) 20:41, 16 June 2017 (UTC)
Robert McClenon, yes to either or both. See the 1st paragraph in Sustainable development and substitute "natural resources" with "human resources", and "ecosystem resources" with "free editing". Unsustainable is when a system can no longer meet human needs without undermining the integrity and stability of the natural systems, but in our case it would be can no longer meet the demands of growing without undermining the integrity and stability of the project. Articles are being created (or upgraded/changed) faster than we can meet the demands of reviewing, citing, validating GNG, copyediting, etc. so we either (1) slow it down so we can meet the demands with the resources we have on hand, (2) upgrade/write the technology that can handle more, or (3) find a way to recruit more editors to competently & efficiently handle the ever-increasing load. Pick two. Atsme📞📧 00:26, 18 June 2017 (UTC)
User:Atsme, User:Kudpung - I don't see any of the three happening. In particular, option (2) isn't plausible, in that it implies that technology can take the place of reviewers. The technology that we have handles the load of expanding the encyclopedia, but how can technology deal with the burden of spammers and promoters? Unfortunately, what will almost certainly happen is that quantity will surpass quality, and that the encyclopedia will continue to expand, but will become increasingly cruddy. Robert McClenon (talk) 17:08, 18 June 2017 (UTC)
User:Atsme - You say that "Hopefully, WMF will give serious consideration to some of the suggestions". Where are the suggestions? What are feasible suggestions to reduce the backlog? The backlog can be reduced either by slowing the creation of new articles or by speeding up their reviewing. The only plausible suggestion to slow the creation of new articles that I know of would be something similar to [[WP:ACTRIAL], limiting the creation of new articles to semi-established users, and the WMF once rejected that, presumably because quantity is foremost to them. I don't know of a way to speed up the reviewing other than to make it less effective. Robert McClenon (talk) 17:20, 18 June 2017 (UTC)
Robert McClenon, your last sentence is basically the same as what I referred to as "unsustainable development", and yes, it is quite obvious that, as you stated, "quantity will surpass quality" which in essence is what "undermining the integrity and stability" means. My apologies for not making it more clear. I strongly believe the problem can be resolved by changing the way we accept and publish new articles - just throwing out some ideas - but couldn't we make it policy that all new articles by new editors will remain in draft space until they have been properly reviewed and cited? I also would not oppose the deletion of any article that is not referenced with inline citations. Article creators should be responsible for doing it right, and not depending on others to complete what they started. Reminds me of the "idea person" who has an unlimited supply of ideas but no time to follow through, much less confirm/validate it as feasible. I also believe we could do a better job of clarifying GNG so we don't have to belabor relentlessly in AfD. There are a number of good remedies - the issues arise when editors get bogged down in the bureaucracy while trying to convince WMF of the project's inevitable fate if we don't make some basic modifications. --Atsme📞📧 18:26, 18 June 2017 (UTC)
No apology required. On the one hand, I would agree with any reasonable scheme to restrict either the moving of articles into article space or at least their indexing until they are reviewed. On the other hand, I have no reason to think that the WMF is likely to go along with such idea, at least not until the mainstream media begin to report that readers are dissatisfied with the proliferation of crud. Until that happens, the WMF has every reason to favor quantity, which can be measured, over quality, which is subjective. I agree that article authors should be expected to find the references rather than let others do the cleanup. It occurs to me that the phrase "Deletion is not cleanup" is part of the problem, in that maybe deletion should be a form of cleanup, but who are you and I against so many spammers and clueless creators and the WMF? Robert McClenon (talk) 18:53, 18 June 2017 (UTC)

(edit conflict)@Atsme and Robert McClenon: The solution requires a combination of items on the 'to do' list, none which are particularly difficult to comprehend or difficult to build. They could all have been easily addressed since I first brought the dilemma back into the forefront in July last year. Although one employee has made a very brave attempt at gathering and producing some stats, Wikipedia talk:New pages patrol/Analysis and proposal has been published in the wake of rising fear that the community will implement ACTRIAL - which as a local Wikipedia it is perfectly entitled to do. The WMF essay is just procrastination in an endeavour to stave off what is probably inevitable. The paradox is that while the Foundation insists that everything possible should be done to achieve the goal of being a a reliable and trusted encyclopedia, factions in the Foundation refuse to allow the tools and measures that would make it happen. Time to read this, and if it sounds familiar, do read it nevertheless - it's been expanded and updated and will answer a lot of questions and lists a lot of solutions. Kudpung กุดผึ้ง (talk) 19:21, 18 June 2017 (UTC)

NPP Noticeboard

I noticed someone post on WP:NPP/N on my watchlist today and it reminded me that we don't really use the page. Looking at the top of the page, it looks to have been designed as a dispute resolution arena for reviewers who might be having competency issues. Looking through the archives, the last time it was used for that seems to be May 2015 and before that, May 2014. Other conversations there tend to not get much feedback because this talk page has become the de facto project talk space for NPP.

The only reason I am bringing this up is that because NPP/N doesn't appear to be closely monitored and isn't being used for its intended purpose, it could be confusing for people who are new to the effort and they might post something that should be seen by more eyes there instead of here. Looking through it, it looks like every thread that was started there in the last year would have been better handled on this board. In the interest of having a place that is unambiguously the central discussion board for NPP, I think it might be best to official retire NPP/N. TonyBallioni (talk) 04:22, 21 April 2017 (UTC)

I'd support that proposal. Looking at the head of this page the graphics point to 9 pages we're expected to keep an eye on. This (IMHO) is too many, issues get raised at multiple pages and it takes too long to find which NPR talk page holds the thread you want to cite in your latest discussion. A smaller number of more broadly targeted pages would be more approachable. For example the Tutorial, School & Help seem to cover pretty similar areas. Cabayi (talk) 06:29, 21 April 2017 (UTC)
Same. We have the same issue at AFC, with at least five talk pages on which to keep tabs. It makes having centralized discussion difficult (and as mentioned, sometimes results in discussions in the wrong place. Just because we can have nine talk pages, doesn't mean we need them. Primefac (talk) 11:38, 21 April 2017 (UTC)
I also agree. - MrX 15:23, 21 April 2017 (UTC)

Proposal

Based on the above conversation and that there seems to be at least enough agreement for a formal proposal, I propose the following: NPP/N be marked as historical and the noticeboard link on the navigation template be retargeted to link to Wikipedia talk:New pages patrol/Reviewers TonyBallioni (talk) 15:45, 21 April 2017 (UTC)

  • Support as indicated in my comments above. Cabayi (talk) 16:05, 21 April 2017 (UTC)
  • Support as proposer. TonyBallioni (talk) 16:16, 21 April 2017 (UTC)
  • Support ~ ONUnicorn(Talk|Contribs)problem solving 16:21, 21 April 2017 (UTC)
  • Oppose I argue for the reverse: mark "reviewers" as historical and move conversations to noticeboard. The term "noticeboard" has meaning on wiki and it's counter-intuitive that we close a noticeboard because conversation formed somewhere else. Chris Troutman (talk) 16:33, 21 April 2017 (UTC)
    • I'd be fine with either. I think keeping the centralized forum here would be easier because it is currently active and has high participation (70/70 watchers visiting vs 50/73). The noticeboard hasn't been used for its stated purpose in two years, and I think the NPR user right means that it is less likely that you'll have situations that need to be escalated to formal dispute resolution. My philosophy is that its easier to build brick paths where people walk than it is to force people to change the path they take, but whatever the consensus is, so long as we fix the problem of having two boards used for the same thing, I'd be happy. TonyBallioni (talk) 16:52, 21 April 2017 (UTC)
  • Oppose the concept of removing it; we have the occasional problematic patroller, particularly if they are too fast with the CSD button on articles one minute old...need someplace to complain. Montanabw(talk) 02:37, 22 April 2017 (UTC)
And the difference in posting that complaint at NPP/R rather than NPP/N is...? Cabayi (talk) 09:36, 22 April 2017 (UTC)

This thread has gone stale, and I see a rough consensus to mark as historical WP:NPPN, if I don't hear any objections to implementing this without a formal close, I will probably implement it in the next few days. TonyBallioni (talk) 00:43, 17 June 2017 (UTC)

  • @MrX and Usernamekiran: I was about to mark NPPN as historical, and then I saw that MrX had brought someone there for a review. I still think the above proposal is a good idea, but thought I should get you alls feedback since you are both involved with active dispute resolution there. TonyBallioni (talk) 14:05, 19 June 2017 (UTC)
  • Yes, that was my point in making the proposal: the pages are very confusing as set up now, and having a single unified discussion board makes the most sense IMO. I think resolving the issue that is currently there makes sense because no need to confuse that, but once it is done, I think it makes sense to mark the page as historical per the rough consensus here. TonyBallioni (talk) 14:12, 19 June 2017 (UTC)
  • Sorry, I forgot. Based on the rough consensus here, I'm going to remove it from the tabs. We should archive everything on the page and mark it as historical as soon as possible.- MrX 14:34, 19 June 2017 (UTC)
  • Not a problem. I also added a discussion tab to the tabs page (I see you replied there, but just noting it here for others). Thanks for your help implementing this. I think we've covered all the basis in terms of marking it as historical and moving discussions here or to archives. TonyBallioni (talk) 14:52, 19 June 2017 (UTC)

Draftify...

The NPP tutorial notes that there is a 'Move to Draft' feature in Page Curation.But I seem to miss that! Any help?Winged Blades Godric 09:36, 19 June 2017 (UTC)

Thanks!Winged Blades Godric

For some articles; maintenance tag's date, and page creation date dont add up

Hi,
In some instances, I realised that these two dates differ. As in, maintenance tag's date precede the page creation date. In some cases I saw the difference in months, and in few cases the difference was 1-2 or even 3 years. These pages arent that old. In the history, there is no indication of being moved from draft-space, or being renamed. I am just puzzled by this issue. Anybody knows what is it about? Thanks. —usernamekiran(talk)(log) 19:51, 4 June 2017 (UTC)

Usernamekiran, chances are someone copy/pasted the tag from another article, either the creator when they copied a "template" to work on, or a well-meaning newish editor who copied a tag from another page. You'd be surprised how often I see that sort of thing in the Draft space as well. Primefac (talk) 20:47, 4 June 2017 (UTC)
@Primefac: yes, that is possible. Also, I noticed there are lots of articles that had (appropriate) maintenance tags, but were not reviewed. I tried to find same users who "patrol" the new articles, but I never came across the tags put by same user. Do you think this would a good idea to find NPP/reviewers? What do you think @Kudpung: sir? —usernamekiran(talk) 21:05, 4 June 2017 (UTC)
What happens is simple: New users copy the code of an existing article to use it as a template to create a new article and simply forget to delete any maintenance tags (or don't know how to). There's no harm in creating a new article this way - I did it myself in my early days when I was creating new articles about various Rhône wines. Among my other very early blunders were: signing my edits, and not kowing how to move a page whe I had made a typo in the title. These all demonstrate how 11 years later, the Wikipedia still fails miserably to give new users even some basic information. THe WMF adamantly resists addressing this situation (under the constant pretexts of: not enough funds, not enough personnel, not enough time), which I certainly feel should be the responsibility of a paid team and not the volunteers.Kudpung กุดผึ้ง (talk) 23:48, 19 June 2017 (UTC)

Extension updates are now live

After a 1 week delay due to unrelated software issues, the aforementioned updates to the PageTriage extension are now live on English Wikipedia:

  • You can now get a list of the top reviewers for the past day, week, or month from the API (so you don't have to scour the logs for this).
  • Fixed T165891 - Special:NewPagesFeed shows users as blocked that aren't currently blocked
  • Fixed T165738 - Number of pages in filtered list is not updated
  • Fixed T44254 - List filters keep getting reset

Kaldari (talk) 20:04, 1 June 2017 (UTC)

Thanks, Kaldari. Is the page supposed to display a chart or code text? d.g. L3X1 (distænt write) )evidence( 21:13, 1 June 2017 (UTC)
All I see is code. I can understand it though. —usernamekiran(talk) 21:28, 1 June 2017 (UTC)
The API only provides JSON data. Perhaps someone could build a bot or Tool Labs tool to convert it into a regularly-updated table or chart. Kaldari (talk) 21:34, 1 June 2017 (UTC)
  • This is excellent, Kaldari, thank you very much. The API provides the same information as a very valuable tool made by Scottywong (retired) which was never moved to Labs. It certainly needs a bot to keep it updated and something to display it nicely. It clearly reinforces some of the claims regarding the performance of the Reviewers. Kudpung กุดผึ้ง (talk) 19:51, 3 June 2017 (UTC)
@Kaldari:--At the API sandbox; the last field for pagetriagestats--topreviewers displays Include the top 10 reviewers over the given timeframe.That should be prob. changed to--Include all the reviewers over the given timeframe.Cheers!Winged Blades Godric 16:14, 5 June 2017 (UTC)

@Kaldari:.- Can you offer some insight as to why there is a (large) discrepancy between the stats in the list of the top reviewers and patrol logs? For example, I patrolled 458 articles in the past month, but the top reviewer stats show 179. Also pinging Kudpung and MusikAnimal who may have some thoughts about this.- MrX 17:49, 19 June 2017 (UTC)

According to usernamekiran in the subsection below, apparently the pagetriagestats do not include page reviews were not done with the page curation tool. The question is why? It kind of makes those stats totally worthless.- MrX 19:20, 19 June 2017 (UTC)

JavaScript viewer available for stats & top reviewers

I spotted this thread in passing, and saw that it would be quite easy to convert the raw API into something a little more presentational, so threw together a quick script to do that. It's User:Murph9000/pagetriagestats-topreviewers.js. It's an early first release type script, works nicely enough for me, but no warranty, there might be the odd bug lurking. It probably isn't compatible with any version of MS-IE (but will probably work ok in MS Edge), as MS-IE's JS support was never up to date and I've probably used some newer stuff somewhere. Load it in your common.js using the following (or equivalent): 

if ( mw.config.get( 'wgCanonicalSpecialPageName' ) === 'Blankpage' ) {
	mw.loader.load( '/w/index.php?title=User:Murph9000/pagetriagestats-topreviewers.js&action=raw&ctype=text/javascript' );
}

Then visit Special:BlankPage via a special URL to run it:

Output is live, although it suggests to the server and browser that the API data be cached for up to 5 minutes, to minimise server impact from frequent visits to the page. If you are not comfortable with JavaScript stuff, this might not be a good solution for you.

Murph9000 (talk) 19:45, 5 June 2017 (UTC)

Oh.. yeah.. works great. [[File:Confused_dog_staring_at_computer.jpg|thumb]] TimothyJosephWood 19:53, 5 June 2017 (UTC)
Nicely accessible. There appears to be a miscount somewhere *possibly in the data) -- I did ~50 reviews in the last month (either 31 days or May). ~Hydronium~Hydroxide~(Talk)~ 09:04, 7 June 2017 (UTC)
Hydronium~Hydroxide, thanks. My script just wraps formatting around the data provided by the API, so any things like that are indeed in the data coming from the server. I don't really perform any processing on the data itself, and don't control the underlying database queries used to generate it. Looking at phab:diffusion/EPTR/browse/master/includes/PageTriageUtil.php, the "last-month" data is for the rolling past 30 days (i.e. not aligned to calendar months), and similarly for "last-week" and "last-day" — they are all for the time period immediately before "now" (with up to 60 minutes of server-side delay on "day", and 24 hours on "week" and "month", due to server caching). Murph9000 (talk) 12:11, 7 June 2017 (UTC)
Thanks Murph9000. Kaldari, for some reason daily lists me correctly, but weekly and monthly don't. I review almost exclusively from the end of the list, where reviews have been made in the last week/month, but the pages were created 5+ months ago. Are the queries looking at page creation time rather than review time? If so, then there's others who also edit from the back, so there might be some level of underreporting. ~Hydronium~Hydroxide~(Talk)~ 06:45, 18 June 2017 (UTC)
I'm also seeing inaccuracy for week and month totals. I guess it's a question for WMF.- MrX 21:18, 18 June 2017 (UTC)
Update: Now I'm seeing inaccuracies in the last day stats.- MrX 22:30, 18 June 2017 (UTC)

@MrX: I dont remember when or where, but me and Kudpung sir had a similar discussion. I think the stats include only the activity from the page curation tool. When I was granted the flag recently, I observed some discrepancies too. While on desktop, if I tag an article for some issue using Twinkle then twinkle also "marks it as patrolled". I think this is excluded from Kaldari's stats script. In mobile version, the page curation script doesnt load up. Instead, at the bottom of an article there is an option: "Mark this page as patrolled". In case the curation tool's script isnt loaded on desktop browser, the option is visible there as well. I think these activities arent counted in Kaldari's stats script.

On other note: I still prefer Special:NewPages, as it provides filters of 5,10, and 15 days (tagging new/very recent pages seems unjustified to me unless they are hoaxes, promotional, attack pages, or regarding a subject with no notability at all). But after restarting the browser, if I open a page from the old feed, then curation tool doesnt load up. So I review few articles from new feed, and then switch back to the old feed. —usernamekiran(talk) 19:08, 19 June 2017 (UTC)

Thanks usernamekiran that makes sense. Unfortunately, it also make those stats completely useless for anyone who uses both Twinkle and the Page Curation tool. I'm mystified as to why the API would only count reviews from page curation tool.- MrX 19:15, 19 June 2017 (UTC)
We can look into this but the API is called "pagetraigestats", pertaining only to data generated by mw:Extension:PageTraige (aka Page Curation). Sorry we didn't make this clear. To get the full stats we'd need to run a query like quarry:query/19672 which unfortunately is quite slow — MusikAnimal talk 21:06, 19 June 2017 (UTC)
Thanks, that query and the one I forked off of it (last-day) seem accurate. I would rather have slow and accurate than fast and useless.- MrX 21:56, 19 June 2017 (UTC)

Top reviewers stats not accurate

Tracked in Phabricator
Task T168335

@MrX and Usernamekiran: I found the problem with the top reviewer stats, but I'm not sure what the best way to fix it is. I filled a Phabricator bug (T168335) so that it can be discussed further and resolved. Thanks for the bug report. Kaldari (talk) 19:39, 19 June 2017 (UTC)

Thanks Kaldari Did you see my request about the stats? It is a very simple one. Actually, most of the details are already up there, one needs to just put them in one place. I dont know how to access them, or I would have done it every month. Can you tell me where to get the stats I requested? Do you have any idea regarding that MrX? —usernamekiran(talk) 19:48, 19 June 2017 (UTC)
Thanks Kaldari for your response and for logging the bug.- MrX 19:54, 19 June 2017 (UTC)

Bot report of top reviewers

The "pagetriagestats" API was not intended to show anything other than stats on mw:Extension:PageTriage (aka Page Curation), but I've created a quick bot task to show the top reviewers by raw number of patrols (so Page Curation, Twinkle, manual). See User:MusikBot/TopPageReviewers/Report. It updates hourly. If you can offer some rough consensus I can file a quick BRFA and get the report written to the more centralized Wikipedia:Database reports. Pinging @MrX, Usernamekiran, and Kudpung:. Hope this helps! — MusikAnimal talk 22:05, 19 June 2017 (UTC)

  • The stats are likely to be skewed anyway because the Foundation has not wished to patch the TW patolls into Page Curation. Righly so (for once) , because Page Curation should be the default method for patrolling pages - that's what it was created for. I still fail to uunderstand however, why it is so impossible for the WFM (or anyone else) to come up with the same kind of excellent tools created by Scottywong on the tool server for reporting on NPP. I know that MisikAnimal has done a lot of work on X-Tools but I think issues like these should be financed by the Foundation rather than be left to the volunteers - of course the downside is that once something becomes a mandate of the WMF, it stands a good chance of never being done: WP:PCSI. Kudpung กุดผึ้ง (talk) 23:36, 19 June 2017 (UTC)
    • If you read the above, you can see that the SQL stats reflect all types of page reviews, regardless of the tool used. This is how it should be. Perhaps the original intention was for the Page Curation tool to be the default, but the reality is that patrollers use the tools that work best for their needs. That is not surprising given that there has been little to no improvement to Page curation tool for the past four years, and its design was lacking to begin with.- MrX 23:56, 19 June 2017 (UTC)
You don't honestly expect the WMF to roll out a completed software package, do you? It's been stated elsewhere (diffs available) that the Foundation will only develop what they want to develop, and they stop development when something more interesting for them comes along, but which is very much less important than getting on with the business of controlling for quality encyclopedia content. I think MusikAnimal, who has done lots of things in his private time that should be the direct responsibility of the WMF, knows what I mean. Kudpung กุดผึ้ง (talk) 00:39, 20 June 2017 (UTC)

User: Hayman30

Moved from Wikipedia:New pages patrol/Noticeboard § User: Hayman30

I am requesting a review of Hayman30 (talk · contribs · deleted contribs · logs · filter log · block user · block log) who is patrolling new articles quite aggressively and not responding to requests to be more careful and less hasty. This user routinely nominates articles for deletion within a minute of their creation and making judgement errors in nominating articles for speedy deletion. In the past 24 hours, four of their CSDs have been declined and I've asked them three time not to be so hasty in nominating articles for speedy deletion. So far, they have not responded, and have not changed their approach.- MrX 11:24, 19 June 2017 (UTC)

 Checking... the background, and incidents. —usernamekiran(talk) 11:46, 19 June 2017 (UTC)
  • The user was notified successfully about improperly using unref section template on March 20. Later on March 28 he was notified for improperly adding unref article template; to which he didnt reply.
  • There are few other incidents related to maintenance tags, and speedy deletions.
  • To summarise: user appears to be editing in goodfaith, but he also appears to lack an in depth understanding of wikipedia policies/guidelines. —usernamekiran(talk) 12:22, 19 June 2017 (UTC)
@MrX and Usernamekiran: Welp, those unref templates stuff is pretty old I'd say. I've changed my approach when tagging articles in the last 20 minutes, I've waited for at least 10 minutes before tagging them. And of course I'm editing in good faith, I'm just not that experienced, we learn all the time. Hayman30 (talk) 12:42, 19 June 2017 (UTC)
@Hayman30: Hi. Yes, we learn all the time. But personally, i think you are still inexperienced for new page reviewing, and should take a break from it for a while. In the meantime, you can work on different level to gain more experience. But this is just an opinion of mine, not a suggestion.
Also, your interaction with user:Ss112 also intrigues me. —usernamekiran(talk) 12:51, 19 June 2017 (UTC)
@Usernamekiran: How that might be? Hayman30 (talk) 13:08, 19 June 2017 (UTC)

@TonyBallioni, MrX, and Hayman30: The discussion been moved here. Hayman, how what might be? —usernamekiran(talk) 14:22, 19 June 2017 (UTC)

@Usernamekiran: I mean how does my interaction with Ss112 intrigue you? Also, I'm not a new page reviewer, in case you thought I am. Hayman30 (talk) 14:26, 19 June 2017 (UTC)
I don't necessarily think that Hayman30 should stop reviewing new articles, but he should be considerably more cautious and less hasty. I was new to reviewing once, and I made some of the same mistakes. I listened to the feedback, observed what other reviewers were doing, and adjusted my approach. What I recommend for Hayman30 is the following:
  1. Don't nominate any articles for A1, A3 or A7 unless they are at least 15 minutes old.
  2. Don't nominate any articles for G11 unless they are at least 15 minutes old, or are so obviously spammy that not a single reasonable person could disagree.
  3. Only nominate articles for A3 is they have no text, only an external link, or only a repeat of the title.
  4. Nominate far fewer articles for G11. If an article contains mostly factual information and some promotional content, don't nominate it for G11.
  5. If an experienced editor gives you feedback about your reviewing, acknowledge that you understand it and don't continue to make the same mistakes.
  6. Don't mark article for A7 deletion unless you are sure that vast majority of editors would agree that the article contains no claim of significance
  7. If you're not certain whether an article should be marked A7, do a Google news and Google book search. If the person, organization, web content or event appear in at least three sources, don't tag it A7.
  8. Help with detecting copyright violation by Google searching entire sentences or using Earwig's Copyvio Detector
  9. If an article is really rough and has no (or poor) sources, but may be a notable subject, consider moving it to Draft: space and nominating the resulting redirect under WP:CSD#R2
As you get more experience and receive less corrective feedback, you can move from 15 minutes to 10 minutes for items 1 and 2 above. Once you stop receiving CSD decline messages, unreviewed messages, and corrective feedback for at least a few months, you can seek to become a page reviewer at WP:PERM/NPR.- MrX 14:30, 19 June 2017 (UTC)
  • I completely agree with MrX. He has provided very good suggestions.
    In some cases, the reviewer should wait for even 30 minutes before nominating them. In case the article is created by new user, it will take a little time to create/improve article. 15 minutes is sort of very little.
    @Hayman30: I am aware you are not a "new page reviewer", but if you are reviewing new pages, it makes you new new page reviewer even if you dont have the flag.
    As MrX said above, you should do it with caution, and as you said, the process of learning is ever ongoing. :-) —usernamekiran(talk) 18:28, 19 June 2017 (UTC)

Unfortunately, there is no way we can force patrollers (or even qualified Reviewers) to read the instructions, and I would suggest Hayman30 read them - See WP:NPP Kudpung กุดผึ้ง (talk) 00:30, 20 June 2017 (UTC)

Also, I would like to request Hayman30 to not delete CSD tags in borderline cases.While going by the rulebooks your de-tagging of Finding Shoes For Odd Feet was correct, per standing conventions we let them remain on such cases.Winged Blades Godric 05:49, 20 June 2017 (UTC)

Possible glitch in Curation tool

I have seldom observed oddities with the tool. A few hours ago, I placed refimprove tag through the tool. It showed up in that particular window/tab, but fortunately, I was viewing the page creator's activity in other tab, and opened the article again, and there wasnt the refimprove tag. It showed in the history as added, but it wasnt removed from the page as per the history itself; yet it wasnt on the article. Here is the history of the article. —usernamekiran(talk) 22:30, 21 June 2017 (UTC)

Kiran, it was added with this edit. You then used reFill to fill the references in one minute later which removed the tag. My suspicion is that you had previously loaded the non-tagged version of the page in reFIll, and then saved it accidentally edit conflicting with yourself and removing the tag. TonyBallioni (talk) 22:35, 21 June 2017 (UTC)
Yes, this might have happened. Thanks for the explanation Tony :-) —usernamekiran(talk) 22:40, 21 June 2017 (UTC)

July 15 push

Moved from Wikipedia talk:New pages patrol/Noticeboard § User: July 15 push

Hello! I posted this on IRC, and it seems to be good, so I will notify people here. On July 15 (a Saturday), we are going to do a push. What will happen is that people with the patroller permission will patrol for as long as they can, and people without it will tag and clean up for as long as they can. No numerical goal should be imposed, as we don't want people to patrol too quickly. So, everybody, lets all try and do our best on July 15. RileyBugz会話投稿記録 15:39, 4 June 2017 (UTC)

Sounds like a good idea. To minimize edit conflicts and multiple people reviewing the same page at the same time, maybe we should split it up somehow, like usernames starting A-L patrol the front, M-Q the middle, R-Z the back. Do you think that's a good idea?~ ONUnicorn(Talk|Contribs)problem solving 17:56, 4 June 2017 (UTC)
I think that that would be ok for people that would like to coordinate. It would also be good to encourage those who are good in their subject area (on Wikipedia) and the policies related to it to participate in their subject. But, I don't think that we should advertise this as something where people should only choose articles in specific subjects or letter ranges. Thanks! RileyBugz会話投稿記録 18:02, 4 June 2017 (UTC)
Good! I get the front! I'll help! Saturday is a great day for me! :) J947(c) (m) 00:35, 5 June 2017 (UTC)
That is great! I will probably be mainly doing redirects and such, as I am coming back from a vacation that day, and patrolling redirects is easy on mobile. It would be nice if we could have these like every month or so, too, but lets focus on alerting people about this. RileyBugz会話投稿記録 01:40, 5 June 2017 (UTC)
Well, then some sort of mass-message needs to be sent.Winged Blades Godric 15:48, 19 June 2017 (UTC)
@RileyBugz and Winged Blades of Godric: if there are no objections to the idea of a review-a-thon (some people at AfC have said negative things about them in the past if I recall), I can send a mass message the week of. TonyBallioni (talk) 16:05, 19 June 2017 (UTC)
Pinging Kudpung and DGG.Winged Blades Godric 16:35, 19 June 2017 (UTC)
  • I know my own limits. I can patrol for at most an hour at a time without decreasing my accuracy, and during that time I cannot properly patrol more than 10 or 15 articles--unless I try to find totally obvious ones in which case I can perhaps do 20. Possibly those with more experience and skill than I can go a little faster; possibly those with more energy and concentration than I can do it for somewhat longer. But anyone who goes much faster or much longer is likely to be overestimating their own abilities, or working to a lower standard.
if we do this project, I will work where I think I'm most needed, which is reviewing the work of those who seem to have an excess of self-confidence. When I review rapid work, I usually find than at least 10% of what should be deleted or marked has been missed, which makes problems for everybody later. And much worse, about 10% of the deletions are likely to be unjustified, which is practice means that about 10% of potentially valuable new contributors are likely to be lost to us forever. DGG ( talk ) 17:15, 19 June 2017 (UTC)
@DGG: Yeah, so I think that in the mass message we should definitely note how patrollers should only do what they can. I think that it would also be good to note that if they find they are burned out from page patrolling, then they should just do cleanup on articles and leave the patrolling to others. RileyBugz会話投稿記録 17:29, 19 June 2017 (UTC)
perhaps even it should include some idea of the time and number expectations. People tend to treat pushes of this sort a a contest. DGG ( talk ) 18:18, 19 June 2017 (UTC)
Well, I don't think that that would be conducive to quality or quantity. If we set a number or time expectation, then people will likely just do the expected time and/or number. Also, we might then get hasty patrolling as people will reach for that limit. I also think that that could discourage those who are not new page patrollers from participating, which is something that I think would help patrollers enormously. I think that we should, although, say that we are trying to reduce the backlog as much as possible, as that might help some people participate. RileyBugz会話投稿記録 18:26, 19 June 2017 (UTC)
About limits: Unless we set these, the reviewers will actually be doing too much. If it's a suitable limit, people ought to reach for it. Perhaps the better word is "expectation" They should rather be encouraged to do several short sessions--and not all on the same day. (of course, looking on the bright side, if this does go through, it will give me ample material for my re-reviews and I may even be able to get some statistics of the harm that is done by the sort of reviewing that this proposal encourages.) . DGG ( talk ) 05:17, 20 June 2017 (UTC)
  • Just my 2₵:
  1. With over 400 Reviewers, I think the mass message should be restricted to the message list we already have TonyBallioni was already given Mass Message rights for this and similar purposes).
  2. It's time to start issuing the barnstars I created months ago. They've never been used by he the new co-ords and there's now a backlog to issue.
  3. Note: The WMF has for some odd reason clearly inferred that quality and depth of reviewing is not necessary. In fact, in their opinion, NPP could be as fast and superficial as possible - they've even suggested automatic patrolling of all articles left unpatrolled after 30 days. I find this odd, particularity considering that the Foundation's goal to make Wikipedia as reliable and accurate as possible in order to get a better reputation in the face of all the negative publicity. I do not believe the WMF is right and it can't be stressed enough how important it is to correctly identify commissioned works (paid editing), clever spam, COPYVIO, hoax, and attack pages and tag them for deletion with the correct criterion. Also, unless the creators are informed of other tags, their articles are likely to remain perma-tagged forever - particular attention to BLPs.
  4. The articles in the backlog are generally those that are more difficult for reviewers with less experience to evaluate. Kudpung กุดผึ้ง (talk) 23:00, 19 June 2017 (UTC)
I do think that it would be good to inform non-reviewers of this so they can help with clean-up, which is what a lot of the work will probably be. RileyBugz会話投稿記録 23:30, 19 June 2017 (UTC)
I don't, but it's a matter of opinion. Bearing in mind that 1) all the patrolls have to be double checked anyway by authorised Reviewers, just creating more work for the reviewers, 2) this type of maintenance work is not one for new and inexperienced users. They should really be concerning themselves (if they must) with vandal patrolling; and 3) as mentioned above, there is no way of forcing the unqualified patrollers to read the instructions.
The 400 reviewers should be able to do this between them, but sadly, as it turns out, less than 10% of those who have the New Page Reviewer right have actually used the permission they asked for, meaning that 90% of the patrols are done by a handful of well established editors with the reviewer right and admins like me and DGG who have it anyway, inferring that many rights holders were possibly hat collectors (as evidence by the requests at WP:PERM) and as a consequence none of the appeals to action by the newsletters have had any effect. Kudpung กุดผึ้ง (talk) 00:05, 20 June 2017 (UTC)
To be fair, Kudpung, I think partially due to your efforts on the newsletters we are on track to be at less than 20k soon.
RileyBugz, I'd be fine with a WP:VPM or WP:WPAFC posting about any potential review-a-thon if we did it (along with encouraging people to apply for the user right per Kudpung's point). I would personally feel uncomfortable sending a mass message about the activity to people who are not on the reviewer list. TonyBallioni (talk) 00:13, 20 June 2017 (UTC)
Well I don't mean sending newsletters to them. I mean posting a general notice. Also, I meant encouraging non-reviewers to do clean-up work, not anything else. RileyBugz会話投稿記録 00:15, 20 June 2017 (UTC)
I didn't say anything about sending newsletters to them either. However, when it come sto cleaning up, I have been practically alone (with some help from DGG) for the last 6 years cleaning up the 'work' of wannabe patrollers, and to be quite frank, its left me highly skeptical. about letting newbies and inexperienced users anywhere near NPP. The whole idea of creating the New Page Reviewer right was to get some quality into the process. Kudpung กุดผึ้ง (talk) 00:23, 20 June 2017 (UTC)
Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia_talk:New_pages_patrol/Reviewers/Archive_8&oldid=1143892794"