Wikipedia

Wikipedia talk:New pages patrol/Reviewers/Archive 7

< Wikipedia talk:New pages patrol | Reviewers
Latest comment: 7 years ago by SantiLak in topic Guidelines for granting suggestion
Archive 1←Archive 5Archive 6Archive 7Archive 8Archive 9Archive 10

Cheating way to cut down on backlog: remove extraneous stuff

Latest comment: 7 years ago42 comments16 people in discussion

The backlog contains all unpatrolled pages including

-user pages and subsets i.e. User:L3X1/CSD_Log, User:L3X1/CBNG, User:L3X1/satanic_rituals_keep_out_admins
As this is where many editors make drafts of new pages which will then go to be properly reviewed and tagged after being moved to namespace, there is little point in wasting Patroller Time with pages in these categories.
-user sandboxes
Again, anything being done here will either be moved to namespace or is will have to be checked again because it would have to be patrolled after every edit. If there are no Copyvios, there may be some tomorrow after all!

None of the above really belong in the backlog. I have stamped "ok" on some 60 wikiEd user pages back to back that had "This editor is a student in this class". Pages that don't need patrolling just add to the number which depresses the team. d.g. L3X1 (distant write) 21:20, 12 May 2017 (UTC)

Is the number cited above all pages or just articles? I often just skim through the feeds for the other namespaces, and if anything doesn't look like it should be there, have a look, but don't bother patrolling those that are fine. There's absolutely no need to patrol things like AFD's, tool-generated user talk pages, or talk pages containing only WP banners. A suggestion was raised at the Village Pump to have a bot handle certain cases of non-mainspace pages after a certain length of time, which sounds like a good idea. That said, I'm not in favor of automatically patrolling user pages/subpages, especially of non-extended-confirmed editors. I find a lot of spam, attack pages, and hoaxes there. – Train2104 (t • c) 21:36, 12 May 2017 (UTC)
Are you referring to my 60 back-to-back or the multi-thousand number above this section? d.g. L3X1 (distant write) 01:31, 13 May 2017 (UTC)
The ones in the thousands. – Train2104 (t • c) 05:19, 13 May 2017 (UTC)
I believe it is all pages, but I'm not entirely sure. d.g. L3X1 (distant write) 23:49, 13 May 2017 (UTC)

Extraneous stuff: If the backlog contains pages that are not articles in mainspace, why did the backlog suddenly being to grow rapidly in June 2016? Why was there a lull from November through mid February, and why is it suddenly growing fast again The stats I post here occasionally do generate a lot of comment but still nobody appears to be actually doing anything about it. In any case, it's now time to stop asking these questions and plough ahead with a concrete solution, otherwise, with Wikipedia now in sharp decline anyway, there soon won't be an English Wikipedia worth wasting our time editing and trying to improve.

Backlog: 21,240. Kudpung กุดผึ้ง (talk) 11:24, 5 May 2017 (UTC)
Backlog: 21,327. Kudpung กุดผึ้ง (talk) 08:28, 6 May 2017 (UTC)
Backlog: 21,476. Kudpung กุดผึ้ง (talk) 01:11, 9 May 2017 (UTC)
Backlog: 21,693. Kudpung กุดผึ้ง (talk) 06:19, 10 May 2017 (UTC)
Backlog today: 22,003. Kudpung กุดผึ้ง (talk) 03:24, 16 May 2017 (UTC)
Backlog: 22,204. Primefac (talk) 03:25, 21 May 2017 (UTC)
OK, here is a concrete solution: Have bots do some of the patrolling. Here are some reviewing tasks that could be handed over bots, or at least assisted by bots:
  • Move unreferenced articles more that 10 days old to draft space.
  • Check for copyright violations and mark them for deletion.
  • Find G4 candidates and nominate for deletion.
  • Automatically review articles that have at least one tag and have been edited by at least three extended autoconfirmed editors.
  • Triage articles into categories (low, medium, high risk of being undesirable). This could be done by evaluating the number of editors of an article, number of other articles edited by the article creator, article title appearance in Google search, etc.
  • Many more such rule sets could be defined to automate a significant portion of the the NPP queue.
Personally, I'm getting burnt out being a reviewer and seeing the queue continue to grow. If the community and WMF will not support a bolder approach to handling the net increasing influx of garbage articles, and instead continues to repeat the new editor retention canard and make other excuses, then I suspect other reviewers will give up as well.- MrX 16:54, 16 May 2017 (UTC)
MrX I don't know the technical aspects about how bots are made and operate, so If my ignorance is showing I'm sorry: I think that while the bot may cut down on some growth, human NPPers will still have to go cleaning up after them. I have worked in the shadow of ClueBotNG, and am in awe of the way it operates, but I have seen it make false positive, and fail to pick up vandalism.
For instance, I don't understand how you would make a bot to detect copy vios. It could operate the same way we do, copy/paste suspicious text into Google, but that isn't foolproof against slight changes and the fact there are a finite numbers of ways to right something. It would be highly annoying if dozens of good articles were yanked/flagged because 8 words strung together happened to find a match on some other page out on the web. The parameters would have to be defined sharply enough to leave good pages alone, but not so sharply it could onyl detect the most blatant (>90% of article exactly matches word for word no breaks another website, e.g).
G4. I would prefer they get flagged for human review, as sometimes articles are deleted for other reasons beside notability, and can stand to be legitamtley recreated. Also, salting takes care fo the most egregious cases. I would like there to be a CSD criteria for previously deleted article by CSD, it is annoying beyond belief to have a poor article's CSD declined because it doesn't align exactly witht he criteria, even though it has been deleted multiple times in the past. The other problem si us ordinary users can't view past deleted articles, so we and the bot have no idea how bad the previous versions were, and whether that should reflect on the new article. d.g. L3X1 (distant write) 02:17, 19 May 2017 (UTC)
We used to have a bot that automatically checked for copyvio. It's actually quite easy. I don't know why it no longer operates. Kudpung กุดผึ้ง (talk) 02:20, 19 May 2017 (UTC)
@Kudpung: We now have a website called CopyPatrol which shows possible copyright violations. Unlike CorenSearchBot, it checks all edits, not just new pages. —MRD2014 📞 contribs 14:36, 20 May 2017 (UTC)
@L3X1: As Kudpung wrote, we had a bot (CorenSearchBot) that would check for copyvios, and my recollection was that it was very effective. I have no idea why it is not operating now. Perhaps because Coren has been inactive for about a year. Bots are fallible just like humans, but I would rather see a few false positives than have copyright violations remain in articles indefinitely. I guess I have an opposite view of G4. If the material is substantially similar (including sources) to material in an article that was previously deleted at AfD, then it should be deleted again. I imagine that a CorenSearchBot with admin rights could manage that. There is a bot that tags recreated articles on their talk pages, but it tends to lag and seems to miss some.- MrX 15:58, 20 May 2017 (UTC)
@L3X1 and MrX:, I don't actually know when Coren stopped working - it's one of those things which one doesn't notice until it's mentioned, but, yes, it was indeed very effective. AFAIK, none of the even regular NPPers check each new article for copvio; I do sometimes but only when a page appears to be an obvious candidate for the additional control.
CopyPatrol is a very important development (incidentally with striking similarities to the New Pages Feed interface), but it's aimed at Recent Changes. I believe new page patrolling to be an equally important priority, if not more so because it concerns retaining good faith new, new article creators. CopyPatrol was developed in excellent collaboration± with some of our top admins and editors which is the way the most new software should be developed, but it is still being used only by the same 13 volunteer co-developers/testers and it is not widely known about (one of them, I can't remember who) developed a useful .js that all NPPers should know about.
In contrast, we have over 400 authorised New Page Reviewers and it still remains an absolute mystery why between them they are unable to reduce this 20,000 backlog. WE asked for 90 days before articles became indexed for Google, but now even this is not enough and some inappropriate articles and other junk is now leaking into the Internet at large which defies the WMF's grand aim of having a perfectly reliable encyclopedia by 2030. Now if that same team could focus their attention on NPP and do a similar job of working on the required upgrades to the New Pages Feed and the Curation dashboard, we might get more people interested in using it instead of Twinkle. Among other things, we would then get back the urgently need leaderboard (which was taken off) and some data resources (which we don't have) to provide stats and follow up. FYI: @Calliopejen, Crow, Diannaa, Doc James, EdJohnston, EdJohnston, MER-C, MRD2014, Sam Sailor, TonyBallioni, Sphilbrick, Bluerasberry, DannyH (WMF), and MusikAnimal (WMF): Kudpung กุดผึ้ง (talk) 02:01, 21 May 2017 (UTC)
Copypatrol includes new pages. MER-C 02:34, 21 May 2017 (UTC)
Thanks User:MER-C was about to say the same thing. I guess the question is can we integrate copypatrol and NPP better? What we need to do is increase the size of the community tech team so they can help address more of the major issues facing our projects. Would be good to mention this in the strategy process. Doc James (talk · contribs · email) 03:08, 21 May 2017 (UTC)

FWIW, NPP was what led me to get involved with Copypatrol. If anyone else was familiar with the EJustice issue, you can basically see my views on plagiarism on their talk page. Its bad writing that most people with a university degree should be able to pick out the second they read an article, and with tools like Copypatrol and Earwig, it becomes very easy to deal with (except for those of use without the bit who have to manually transclude Template:copyvio-revdel, which is a pain.) TL;DR: I wish more people at NPP would use common sense here, and help with copyvio. If something doesn't sound like it is written in the voice of Wikipedia, just put it through Earwing and do a Google search of the worst writing in the article. Odds are, it was lifted from somewhere. TonyBallioni (talk) 03:20, 21 May 2017 (UTC)

Glad to hear NPP is sending people to CopyPatrol! :) The technology behind CopyPatrol should make it easy to create a bot to tag the pages with copyright issues just as CorenSearchBot did. For this I've created phab:T165951. We have the data to not only check new pages but every edit. Does it make sense to tag pages after any edit that has a strong chance as being a copyright violation, or should we just do new pages? MusikAnimal (WMF) (talk) 09:53, 21 May 2017 (UTC)
The problem is, MA, that NPP is NOT sending people to CopyPatrol. Outside the Developer/tester group nobody is using it and few people have even heard of it. In fact for a New Page Reviewer to use it would double their work load. It's actually easier and quicker to paste a snippet from a suspect text into Google, and as previously mentioned, most new page patrollers, particularly the unlicenced ones, won't bother to do that anyway. COPYVIO will never be addressed from the New Page Review queue until CopyPatrol actually tags new articles as Coren used to, or at least shows an alert in th New Pages Feed alongside the other red alerts. It should be a clickable red link that sends the patroller to the entry in the CopyPatrol list for further action. Kudpung กุดผึ้ง (talk) 12:07, 21 May 2017 (UTC)
The Copypatrol workload is being dealt with promptly and daily by a small cadre of editors. There is essentially no backlog; items reported yesterday will be cleared today. You can see who is working on this task by looking at the Leaderboard. Coren no longer edits, and his bot became redundant, since Copypatrol checks both new articles and all added text over a certain size threshold. It's still important for new page patrollers to check for copyvio as the bot will not catch everything. — Diannaa 🍁 (talk) 13:00, 21 May 2017 (UTC)
I agree it seems CopyPatrol users are more or less keeping up with the backlog and catching many things NPPs are not. So I suppose we need not worry about sending more people to CopyPatrol. That being said, is it at all worthwhile to create a bot to tag pages, that goes off of CopyPatrol data? After all, Dianna will be there to take care of it before too long :) MusikAnimal (WMF) (talk) 13:27, 21 May 2017 (UTC)
A bot tagging the pages would be redundant as the items are being inspected already in a timely fashion via the copypatrol system. However I do think we need a few more experienced helpers there, as some days I am the only one, which makes for a very long day! — Diannaa 🍁 (talk) 13:46, 21 May 2017 (UTC)

The advantage of a bot or closer integration between copypatrol and page curation (red text like we have now for orphans, citations, etc.) is not necessarily helping out with Copypatrol (though I'm sure Diannaa would appreciate it.) The advantage I see would be that it would get more NPP eyes on bad article quicker. Even if something is fine from a copyright point of view, a lot of the false positives you get on new pages are unacceptable for the encyclopedia in other ways, or are just plain bad writing. Basically, my view is that false positives on Copypatrol are often a red flag for other issues. Highlighting them would put more eyes on them more quickly. TonyBallioni (talk) 14:02, 21 May 2017 (UTC)

  Comment: I agree with L3X1. If not all, (non-article) pages created by extended confirmed users should be excluded. —usernamekiran(talk) 14:44, 21 May 2017 (UTC)
I agree we need more autopatrolled users. APAT says A suggested standard is the prior creation of 25 valid articles, not including redirects or disambiguation pages. But how would we encourage people to apply to become autpatrolled? d.g. L3X1 (distænt write) )evidence( 15:39, 21 May 2017 (UTC)
Nominate them yourself. TonyBallioni (talk) 15:41, 21 May 2017 (UTC)
No. I came across one APAT user in the long past, and one in the recent past. The recent one has created 100+ articles. More than half of them have notability issues, and should be deleted/merged. But somehow his articles went unnoticed even before he became an APAT. Instead of more APAT users, we need more reviewers. And like Tony said, one should nominate good candidates. —usernamekiran(talk) 15:58, 21 May 2017 (UTC)
TonyBallioni Do you know of any log, leaderboard, or WMFtoolslab where I can see the top current article creators? d.g. L3X1 (distænt write) )evidence( 17:18, 21 May 2017 (UTC)
@L3X1: See Wikipedia:Database reports/Editors eligible for Autopatrol privilege. We are going to work to improve the report by looking at statistics other than just raw article count, in an effort to surface users who are more likely eligible to be autopatrolled. MusikAnimal (WMF) (talk) 20:12, 22 May 2017 (UTC)
@MusikAnimal (WMF): that's so great to hear, it'd incredibly useful to have a version that incorporated some more variables beyond articles created (perhaps, articles deleted; whether the user is active; history of blocks; or other factors). Question in the meantime: do you know if it's possible to have the bot run an updated report in current form? I saw last week that it tried to run but just blanked the page; and I had also seen a discussion on a talk page (I now can't recall where) that maybe it's getting automatically cancelled because it takes more than three hours to run, but if it were possible, it'd be great to have a current version even of this more basic list--AFAICT, the last report is from March 30 and at this point quite a few of the folks listed as "eligible" actually now have autopatrol, so it'd be so useful to have a refreshed list of those who don't yet. Understood if that's not technically feasible right now tho! Innisfree987 (talk) 03:50, 25 May 2017 (UTC)
@Innisfree987: Behold, the new and improved report! Let us know what you think at Wikipedia talk:Database reports#Editors eligible for Autopatrol privilege. MusikAnimal (WMF) (talk) 18:54, 30 May 2017 (UTC)
  • I think we need more gnomes in general, and admins, or admin-gnomes in these areas. Apart from the general NPP tasks, the copy patrol part is time consuming on the research, then when action is required (revdel, histmerge, or outright G12), that goes into the admin to-do-queue. As Doc James alluded to, I can think of a couple of tech gadgets that would speed up CV patrol, such as a tool link to automatically front-end the current page into Earwig in a new browser tab. (I do that manually now and usually get FireFox warning that any new tabs will slow down my experience, but I digress...), or a tweak to the NPP browser that lets you focus on a specific namespace (for the User page example above). Regarding the OP, I would not want to automatically mark user sub pages without eyes on them, as there is/was a rather Beansy reason for this, at least last year (may have been fixed by now). Just my ramblings... CrowCaw 18:06, 21 May 2017 (UTC)
I agree that user subpages should not be automatically patrolled -- just have a look at my deletion log. MER-C 07:14, 23 May 2017 (UTC)
I never thought much about user pages and sub pages until it was mentioned yesterday so I selected user pages in the Curation filter and had a look. Among the first 20 or so, so many were so bad I summarily deleted about 5 of them. Kudpung กุดผึ้ง (talk) 07:56, 23 May 2017 (UTC)
  • I care about this issue but do not have so much time to address it. Here is where I am:
  1. I just reviewed 10 new article drafts. I passed some and declined others, and wrote some comments. It took me about 20 minutes. I can commit to do at least this much every week for the next few months.
  2. I read these discussions from time to time and am aware of a range of issues about articles for creation. If I had to summarize all the problems, I would say that there is a lack of leadership here. Despite the long-term guidance of some of Wikipedia's most practical and insightful policy makers in this space, there is a need for some paid-staff research in this space to ever sort through the 3-5 major challenges and the 15-20 smaller challenges which are inherent in the infrastructure of this review process. The kind of leadership required is someone to do the complicated task of listing all problems, getting opinions about the consequences of what will happen if the situation does not change, and cornering some of the high-profile decision-makers (especially key WMF staff) to either grant some of their special powers to help address what is happening here or otherwise divest their power to community volunteers to resolve the problem in a sensible way. It is not sustainable to continue the same behavior and expect different results in the future, which is the current plan.
  3. The fundamental problem here is that for various reasons, volunteer reviewers' time is being wasted because of a combination of inadequate software, the fact that this process repeatedly puts volunteers into stressful social situations without providing them with the canned text extricate themselves from customer service to send implacable complainers on to somewhere else, and some wiki community isolation here in which volunteers lose time for sadness because despite the value of their contributions being here has a way of feeling thankless in the broader wiki community despite everyone depending on this service. I think that volunteer labor can address this process but some infrastructure development is necessary to leverage more impact from volunteer contributions. Lots of people here can name wanted infrastructure.
  4. I think that this project has the reputation of the Wikimedia movement hostage. If this project gets out of control, then it could get negative media attention. When negative media attention happens, then whatever else the WMF is doing gets derailed and at great cost and in a scramble the problem has to get emergency patching. There are some fundamental values conflicts brewing here right now, like for example, there have been previous conflicts in which the community has proposed "no article creation by new users" or similar things. While there was WMF opposition to the community implementing these kinds of brand and policy changes, that kind of change is organically happening now as a consequence of the impossible backlog which is being created and which volunteers are not going to be able to address without drastic changes to the review process one way or another.
  5. This is serious! It is beyond my capacity to do much here but I see what is happening and I feel that tensions are only rising. The most that I can do at this time is be aware and also review a little a week - perhaps 10 articles - because actually, 10/week is more than this community should expect out of anyone and ought to be a good standard of ideal participation. With the rising participation and current backlog, it is not feasible to expect that even high growth in participation by people who are great reviewers will keep this backlog in check.
Blue Rasberry (talk) 15:00, 23 May 2017 (UTC)
I work on the WMF's Analytics team and want to add a couple of things to this discussion as a noobie volunteer / WMF employee. First, I'm reading up on what's required to help patrol new pages. I know one more noobie isn't going to fix the problem but I hope it will help. Second, we at the WMF take this very seriously. Toby Negrin (TNegrin, head of Product at WMF) and others have already replied and showed this but I just want to echo and amplify it. For this immediate problem, we have over a dozen people working on the report TNegrin mentioned. Long term, we have at least three teams I'm aware of thinking about the various community backlogs: measuring them better, making them into first class things that we can all have a discussion about, and in general giving them prominence like having them feature on stats.wikimedia.org. From my point of view on the Analytics team we have built dashboards and datasets that help answer questions about readers and editing but have not given backlogs of tools like NPP the attention they need. This became obvious to us over the past six months and we're making plans to address it. Milimetric (talk) 10:17, 26 May 2017 (UTC)

Arbitrary break

Blueraspberry and anyone else who is interested in the meta parts of new pages patrol and moving it and AfC forward you might want to add yourself to this mailing list/project: Wikipedia:The future of NPP and AfC/Work group list. I plan on starting a conversation there next week about how to move forward with previously discussed reforms here, and being on that list will make sure you're kept aware of anything that is going on with that. TonyBallioni (talk) 15:58, 23 May 2017 (UTC)

A couple comments. 1) Am I supposed to just tell people on their talk page that they are eligible for APAT or am I supposed to go over and add their name to the request list? Currently, I have only written one editor on his TP. 2) I don't think banning new users from creating pages is a good idea. To me it sounds too much like the stamp-out-IP-to-solve-all-our-ills that surfaces from time to time. NPP exists to patrol New Pages. d.g. L3X1 (distænt write) )evidence( 00:29, 24 May 2017 (UTC)
1) Just nominate them for autopatrolled at PERM yourself. . 2) IPs currently can't create pages and haven't been able to since 2006. The question moving forward is about whether page creation should be restricted to autoconfirmed users, which there was conensus for in 2011. The other pages in the work group linked above give more background. TonyBallioni (talk) 00:51, 24 May 2017 (UTC)

How to view un-patrolled pages created 5 days ago?

Latest comment: 7 years ago14 comments4 people in discussion

Hello,
I recently got the user-right. Before getting it, I used to look at the new pages at Special:NewPagesFeed (or maybe from somewhere else, I am not sure about it). I used to select "5 days old". But after getting the right, there are only two filters regarding dates, being "oldest" and "newest". Is there any way to filter the results in such way that I would be able to review pages created 5 days ago or older? It is a wise policy to review such pages, as it gives enough time to creator and other editors enough time to work on the article. It doesnt seem right to add maintenance tags to the articles which might be still under work. 5 days seem to be appropriate time. —usernamekiran(talk) 07:14, 29 May 2017 (UTC)

Any suggestions @L3X1 and Kudpung:? —usernamekiran(talk) 07:14, 29 May 2017 (UTC)
  Comment: Found it. Special:NewPages —usernamekiran(talk) 09:20, 29 May 2017 (UTC)
@Usernamekiran: how ?? Sulaimandaud (talk) 09:08, 30 May 2017 (UTC)
@Sulaimandaud: there is an option at the top. —usernamekiran(talk) 15:53, 30 May 2017 (UTC)
Hi @Usernamekiran: can you explain how to use the option and exactly where it is at the top, because I have the same problem I only have sort by oldest and newest. Domdeparis (talk) 15:57, 30 May 2017 (UTC)
I get it! Use the special:newpages and not the feed? Domdeparis (talk) 16:01, 30 May 2017 (UTC)
@Domdeparis: Yes. The option is available in Special:NewPages. But there is one logical problem with this: newpages shows only the pages created recently, where as the feed shows pages that are not marked as patrolled. So both needs to be used alternatively. You know, like one day this, and next day that; or fee pages from this and few pages from another. —usernamekiran(talk) 16:12, 30 May 2017 (UTC)
Given that the original intention was to patrol from 5+ days, the fact that Special:NewPages caps at 30 days is largely irrelevant. An good difference to keep in mind, though. Primefac (talk) 16:16, 30 May 2017 (UTC)
That said if you use the double arrow after having tagged a page as having been reviewed it seems to go to the next non-reviewed article in the queue so the Special:NewPages can be used to start reviewing and the curation bar to move on from there. Domdeparis (talk) 16:24, 30 May 2017 (UTC)

@Domdeparis: Yes, thats a good idea. Inwill try it next time to observe which pages are brought up next. @Primefac: Yes, there is an option to view pages that were created 15 days ago. But like you said, as it is capped at 30 days, the 15 days is the last filter.

Tagging fairly new pages feels a little harsh/unjustified, and like tag-bombing even if they are just two tags. Doing it after 5 days, or later seems rational. But to chip away the backlog, both the feeds must be used. We really need experienced users to take this responsibility. I have invited two users so far, one of them was granted the user-right, the other one hasnt requested it yet. —usernamekiran(talk) 16:36, 30 May 2017 (UTC)

@Usernamekiran: I tried it with a dozen pages and each was unreviewed and created a few minutes after the one before so it seems to work. Domdeparis (talk) 16:42, 30 May 2017 (UTC)
@Domdeparis: I did not understand "each was unreviewed and created a few minutes after the one before so it seems to work." :-| —usernamekiran(talk) 17:31, 30 May 2017 (UTC)
sorry not very clear ! I meant that each page that I came across was an unreviewed page and they followed each other chronologically. So when I was on page 1 created at say 10:05 it did not have the green tick and when I clicked on the double arrow button page 2 came up which was also unreviewed and had been created the same day at 10:07 and so on and so onDomdeparis (talk) 21:16, 30 May 2017 (UTC)

New autopatrolled eligibility report

Latest comment: 7 years ago10 comments6 people in discussion

Community Tech is pleased to announce we have a new and improved report on users who are eligible to be autopatrolled. Hopefully the added information can better help you identify users who are most suitable to get the flag, and hence lessen the burden on the backlog. A new report is generated daily. Feel free to leave us feedback here or at the discussion at Wikipedia talk:Database reports#Editors eligible for Autopatrol privilege. Kind regards, MusikAnimal (WMF) (talk) 18:59, 30 May 2017 (UTC)

It looks so much better and more useful than the previous report. Hopefully it'll run more reliably too. Thanks! Cabayi (talk) 19:53, 30 May 2017 (UTC)
Oh wow. That's... actually really great. TimothyJosephWood 21:06, 30 May 2017 (UTC)
MusikAnimal (WMF), just from some random spot checking, I am seeing two accounts (out of four I checked) where it is reporting zero deleted articles despite the users having multiple articled deleted via PROD. (Read this not as a complaint, but as my making you aware) TonyBallioni (talk) 21:14, 30 May 2017 (UTC)
@TonyBallioni: Which accounts, if you don't mind sharing? Note also only deletions in the past year are counted MusikAnimal (WMF) (talk) 21:24, 30 May 2017 (UTC)
@MusikAnimal (WMF):Ah, that would be it. Out of bad habit I skipped straight to the footnote! The accounts I spot checked had no deletions in the past year. Sorry for any confusion. TonyBallioni (talk) 21:40, 30 May 2017 (UTC)
Adding my kudos as well!! What a dramatic improvement, thank you SO much! TonyBallioni, to your question I'd guess what you're seeing (Ah edit conflict, I see MA has noted this as well) is that the "Deleted" column just shows the deletes for the past year, rather than all-time deletes--I noticed this as well and want to play around with it a bit to think about whether all-time deletes would be more useful, but in any case yes in no way a complaint from me either, this is a leaps-and-bounds improvement, and I'm very grateful for the work making this tool so much more useful. Thank you! Innisfree987 (talk) 21:29, 30 May 2017 (UTC)
  • Thanks for the feedback, all! I wanted to point out that I've just added a "copyvios" column, that shows the number of possible user talk notices regarding copyright issues. This was something I did without discussion, but being an admin myself (MusikAnimal) I often find it tedious to check through the revision history for copyright warnings, which if present generally means they aren't ready to be autopatrolled. So hopefully this feature is welcomed and you also find it useful. Best, MusikAnimal (WMF) (talk) 21:39, 30 May 2017 (UTC)
Gosh, MusikAnimal (WMF), this tool really does make the process infinitely more efficient--I've done half a dozen in the last two days. It helps greatly to see who's even active, for starters, as well as block log right up front. Copyvio is a great addition. If you're still tinkering with it, I have a couple thoughts but so many thanks for this already-major upgrade!
  1. I'm wondering about having a dedicated column on revocation versus just taking those who've had the right revoked out of the list of recommended candidates altogether? Not that they could never regain the right, but in the interests of generating a list that most efficiently identifies candidates likely to pass, as well as avoiding name-and-shaming where possible.
  2. Along similar lines, I'm wondering if it's possible (or desirable?) to likewise remove candidates MusikBot would identify as having had a request declined in the last 90 days (just until the 90 days has passed)? Again not that they could never be accorded the right, but just in terms of generating a list of who'll likely pass now, and avoiding duplication of recently-performed work evaluating candidates.
  3. I'm encountering quite a few accounts where the editor has no deletions in the past year, but it's mainly because they've only made one or two this year, while a large number of their prior contributions have been deleted. Possible to add an all-time deletions column?
In any event, big thumbs up once again from me. Report's great, and it's clear how useful it'll continue to be going forward, getting new lists of recently active editors. Innisfree987 (talk) 22:34, 31 May 2017 (UTC)

Google referencing of unreviewed pages

Latest comment: 7 years ago8 comments4 people in discussion

Hi I had an interesting discussion with @Kudpung: who explained to me that unreviewed pages or rather pages that have not been ticked on the curation bar were not indexed by Google. I just started reviewing Okandé which was a translation of the French version Okandé. When created the article already had the tag {{Unreferenced|date=September 2012}} which seems to have been translated with the article. I thought I'd carry out a search and the okandé page in WP was the first result. here. What I don't understand is how that could be. I thought that maybe it was because someone had carried out edits and inadvertently unticked the box after referencing by Google. To see if this was the case I tried another unreviewed article Robert Savković that had 1 single edit on it that of the article creator who by the way does not have reviewer rights. When I carried out a search it popped up in first place here (I checked on my smartphone too to see if it wasn't the cache on my navigator that caused this and same result). Is there something I am missing? Domdeparis (talk) 13:28, 30 May 2017 (UTC)

Did you notice, Domdeparis, that both of those articles were created on 21 December ? That said, the instructions recommend that we report bugs here, but since I no longer take care of NPP/NPR, nobody appears to be answering there or coordinating with the developers. The page header there recommends reporting issues directly to Phab, but pinging Kaldari might just get you a short cut to attention. If it's something he can fix he probably will. Kudpung กุดผึ้ง (talk) 13:57, 30 May 2017 (UTC)
I did notice that as I am working through the backlog from the back forwards but to be honest I didn't know if it was a bug or something I had missed. Thanks for your help. Domdeparis (talk) 14:13, 30 May 2017 (UTC)
@Domdeparis: I also noticed something. Bheemavaram, Khammam district was created by a bot years ago as Bheemavaram. It showed up in feed. After marking it as patrolled, it did not show up in my log. There were at lest 10 articles that were created years ago, I marked them as patrolled but they never showed up in my log. A little similar issue has been discussed above. @Kudpung: You no longer take care of NPP/R?! —usernamekiran(talk) 16:05, 30 May 2017 (UTC)
I haven't been taking care of NPP/NPR or assisting with development since I resigned (very loudly)a few months ago from my self-assumed task as its coordinator over the last 6 years. However, each time I am pinged, I feel a moral obligation to respond. You won't have read about it in Signpost because that closed down too around the same time, coincidentally also the same time as the backlog began to rise dramatically again. Maybe there's a bell tolling a warning for Wikipedia - but as usual, people pretend they can't hear it. The other warning in red has been on my talk page for a year. There are a lot of cracks appearing in Wikipedia's walls. Kudpung กุดผึ้ง (talk) 16:15, 30 May 2017 (UTC)
I did not know that. And I agree with you regarding cracks appearing in the walls. Soon it might get out of control. —usernamekiran(talk) 16:28, 30 May 2017 (UTC)
@Domdeparis: NOINDEXing on English Wikipedia articles is explained at Wikipedia:Controlling search engine indexing#Indexing of articles ("mainspace"). Hope that helps clear up any confusion. Kaldari (talk) 19:22, 30 May 2017 (UTC)
@Kaldari: thanks, it is now much clearer. Domdeparis (talk) 09:00, 1 June 2017 (UTC)

Spam bot creations?

Latest comment: 7 years ago9 comments3 people in discussion

I keep finding userpages like User:Mohamed kafi that follow a very specific formula. South Asian name with username being the first and second name. Then they make an edit to the userpage of First Middle Last name. These accounts never edit anywhere else that I've dound yet. I'm going to start collecting examples here and others can too. Legacypac (talk) 12:08, 3 June 2017 (UTC)

@Legacypac: erm... I did not understand you/your words. I mean, i didnt understand anything at all. Would you please elaborate? —usernamekiran(talk) 12:16, 3 June 2017 (UTC)

User:Harish Shahu User:Karan bhati777

I suspect new user pages are being created in large numbers automatically by a bot. I've started collecting examples here. Maybe if they all come from the same location that location can be blocked. Legacypac (talk) 12:23, 3 June 2017 (UTC)

lol. I had understood that much in first go. It can be a bot. I have come across such users too. But they never edited in a few months. In some cases, never edited at all. I am not sure if it is work of a bot, or if they are genuinely lazy users. I mean, if it was a bot it would have tried at least something, right? Either, advertising, or removal of content, or some sort disruptive editing. Harish Shahu, and Karan Bhati are legit names in India (first name, surname format). But these talkpages/userpages are edited in a very similar manner. In lots of cases, I have seen "name - location" format; similar to Karan Bhati - Nagpur. —usernamekiran(talk) 12:35, 3 June 2017 (UTC)

Don't understand the benefit to someone setting a bot to make lots of userpages but then I don't understand the benefit to most vandalism. If some bot is auto creating bogus user accts that is vandelism. If it was humans we should see more random behavior. The odd test edit or whatever. Legacypac (talk) 17:28, 3 June 2017 (UTC)

When I am stamping user pages, I've noticed that some groups of users tend to think Wikipedia is some sort of Facebook or something, I never thought of it as being bot-work. d.g. L3X1 (distænt write) )evidence( 19:16, 3 June 2017 (UTC)
@L3X1: yup, I have seen a lot of inactive users who thought of wikipedia as a social networking site, when they realise other editors arent in that thing, they tend to give up.
@Legacypac: Not all the acts of vandalism are beyond understanding. —usernamekiran(talk) 19:49, 3 June 2017 (UTC)
Hey that's an accurate but misguided post you linked!
A lot of the social media youtube etc link building is misguided but not useless for the spammer. Wikipedia is a powerful provider of link juice to your YouTube video. Legacypac (talk) 20:15, 3 June 2017 (UTC)
@Legacypac: I meant, that act of vandalism can be understood. —usernamekiran(talk) 20:45, 3 June 2017 (UTC)

A question from another user

Latest comment: 7 years ago2 comments2 people in discussion

One of my article Shakeel Ahmad Khan was reviewed by you, however it is doesn't show in google search. Is there any coding or indexing problem?
This was asked to me a few hours ago on my talkpage. As I didnt want to provide inaccurate info, I havent replied him yet.

Does anybody know the answer? Thanks. —usernamekiran(talk) 07:51, 6 June 2017 (UTC)

Usernamekiran, we have no control over when or how fast Google indexes our pages. Generally new pages are indexed in a few hours, but I've seen it take almost a day. Primefac (talk) 13:02, 6 June 2017 (UTC)

Backlog 21,240

Latest comment: 7 years ago29 comments11 people in discussion
Backlog today: 21,240. Kudpung กุดผึ้ง (talk) 11:24, 5 May 2017 (UTC)
Backlog today: 21,327. Kudpung กุดผึ้ง (talk) 08:28, 6 May 2017 (UTC)
Backlog today: 21,476. Kudpung กุดผึ้ง (talk) 01:11, 9 May 2017 (UTC)
Backlog today: 21,693. Kudpung กุดผึ้ง (talk) 06:19, 10 May 2017 (UTC)
Backlog today: 22,003. Kudpung กุดผึ้ง (talk) 03:12, 16 May 2017 (UTC)
So from those numbers it looks like the backlog is growing at an average rate of 75.5 articles per day. The question is what can we do with this information? ~ ONUnicorn(Talk|Contribs)problem solving 15:32, 10 May 2017 (UTC)
Obviously marking the page as patrolled is a very important part of tackling the backlog! So what's the deal? Are people afraid? Are we so stringent about inclusion criteria that we often need second opinions? Every article cannot be perfect, but if you get it to an acceptable state, with or without maintenance tags (following WP:NPPCHK), I think you've done your job and it's OK to remove it from the backlog. I understand there is a delicate balance there, though, but I wonder why so many people are stuck on the knife edge? And what about pages that have been PROD'd or AfD'd? Doesn't that mean you "reviewed" them (Iaso Hospital, for example, where a sysop didn't mark it as patrolled)? Are there any software improvements we can make to help make it easier or remind users to mark pages as patrolled? — MusikAnimal talk 18:17, 10 May 2017 (UTC)
I'm guilty of two or three of those today. I find that I can usually identify issues with an article, but claims of significance and notability are not always clear. Perhaps I should err on the side of marking articles as patrolled unless I'm certain they should be deleted. I'm open to advice on this.
It also occurred to me that perhaps articles should automatically default to reviewed after 30(?) days. That way, at least the backlog wouldn't continue to grow to soul-crushing proportions (thank you Kudpung for constantly reminding us  ).- MrX 18:36, 10 May 2017 (UTC)
That's... kind of a terrible idea. If that were implemented a one-sentence unreferenced BLP could, in theory, be marked as patrolled even though no one has actually looked at it. Primefac (talk) 18:40, 10 May 2017 (UTC)
I'm not so sure, but I probably should have qualified my proposal: automatically review 30 day old+ pages that have been tagged, and that meet whatever other criteria are deemed necessary (number of unique editors, number of sources, number of inbound links, etc.). Failing that, we have to get more active reviewers, find a way to slow the influx of new articles, or both. I assume that those two factors (new articles and reviewers) are trending in opposite directions resulting in a queue that is roughly four times the size it was roughly a year ago. - MrX 19:05, 10 May 2017 (UTC)
Well, we are supposed to only mark pages as patrolled if they correspond to the policy. If I made changes to the page but it still does not correspond to the policy (or I am not sure and need to investigate further) I obviously do not mark a page as patrolled.--Ymblanter (talk) 19:11, 10 May 2017 (UTC)
@Ymblanter: you said "we are supposed to only mark pages as patrolled if they correspond to the policy." Which policy is that? My understanding is that if we check the page and tag or fix issues or nominate it for deletion, we should mark it patrolled. Are you saying that you feel we should not mark it patrolled if it still has issues? Please clarify. ~ ONUnicorn(Talk|Contribs)problem solving 20:31, 10 May 2017 (UTC)
I mean the tutorial. Indeed, I believe that if the issues which are compulsory for a partolled article (such as the presense of categories) are not fixed the article should not be patrolled. Or BLP violations or copyvio should not be patrolled even if nominated for deletion - patrolling them would immediately make the article visible in search engines.--Ymblanter (talk) 20:41, 10 May 2017 (UTC)
But if you use the page curation toolbar to mark it for deletion it automatically marks it as patrolled. I seem to remember some talk about submitting a bug report to phabricator about that, but I think it was decided to leave it, but I don't know why. I'll see if I can dig up a diff. ~ ONUnicorn(Talk|Contribs)problem solving 21:19, 10 May 2017 (UTC)
Ymblanter I found it. Wikipedia_talk:New_pages_patrol/Reviewers/Archive_2#Comments_and_Questions. Basically, BLP violations and Copyvio are generally speedy, and those deletion templates transclude no index, so patrolling them with a CSD tag should not make them visible to search engines. ~ ONUnicorn(Talk|Contribs)problem solving 21:35, 10 May 2017 (UTC)
And here's a list of all the templates that transclude NOINDEX. G12(copyvio) and BLPPROD are both on there. ~ ONUnicorn(Talk|Contribs)problem solving 21:41, 10 May 2017 (UTC)
And then if say the article creator removes the speedy template?--Ymblanter (talk) 05:40, 11 May 2017 (UTC)
Of course, and I didn't mean to point the finger at you and MrX regarding Salzburger Kunstverein :) I understand some people don't have time to fully evaluate an article, or think it's not up to code, but it seems that is a pattern here. I'm also seeing a lot of pages created by seasoned editors apparently fit to be autopatrolled. Perhaps we should devote some time to go through Wikipedia:Database reports/Editors eligible for Autopatrol privilege? — MusikAnimal talk 19:53, 10 May 2017 (UTC)
Actually, every time I see a mature newly created paper, I just go and look at the contribution of the editor. If they have enough experience (say several thousand edits and enough tenure), recent creations are fine, and there are no obvious red flags at the talk page, I just make them autopatrolled.--Ymblanter (talk) 20:38, 10 May 2017 (UTC)

None of the points discussed here address the odd behaviour of the chart or investigate its possible correlation or coincidences with other events.

  • in July 2016 long before serious discuss began on doing anything about it the number of unpatrolled articles Iegan to rise steeply
  • It flattened off almost immediately Nov/Dec when the new user right was rolled out, and even began to drop.
  • In February it suddenly began to rise again at is previous rate.

There are currently 401 accredited reviewers. Other, unqualified users are still allowed to tag pages (but not mark them as patrolled), (which, BTW, still does not avoid good faith creators from being bitten), despite which, 401 should be more than enough to not only keep the backlog growth down, but to significantly reduce it.

MusikAnimal, I do not understand why some people don't have time to fully evaluate an article. Either they patrol/review constructively, or they do not. It’s written big enough at the top of the feed: Rather than speed, quality and depth of patrolling and the use of correct CSD criteria are essential to good reviewing. and at the end of the day, a backlog is still better than an encyclopedia corpus full of spam and other totally unacceptable pages - most of which are reltively easily recognised for what they are, while articles made my genuine, but non autopatolled authors need only seconds to review and are not likely to be a cause of the backlog. Any articles whose deficiencies are not so easily recognised are precisely one of the main reasons why a group of sufficiently experienced reviewers was created. I suggest reviewers (and anyone else commenting here) try setting their filter on pages that 'Were created by new editors' and work on that this list for an hour and then they’ll see and fully understand the gravity of the situation. Increasing the number of editors with the autopatrolled flag, or any other suggestions being made here are only palliative, whereas restricting the creation of article to auto confirmed users would be a concrete, important step forward. Kudpung กุดผึ้ง (talk) 23:33, 10 May 2017 (UTC)

I don't think spending inordinate amounts of time evaluating individual articles is a good approach when there is such a huge and rapidly growing backlog. A triage approach seems more productive. Also, we still haven't identified any root causes for the current backlog. Is it that there are few regular reviewers; fewer reviews per reviewers; more new articles; more articles that depend on non-English sources; more articles with large lists of fake references; inconsistent feedback from CSD reviewing admins causing reviewers to second guess themselves; or any number of other reasons? I think most of us agree that ACTRIAL or something similar should be implemented, but it would also be helpful to understand how we got here.- MrX 00:01, 11 May 2017 (UTC)
I've just spent the last 1.75 hours reviewing new pages from the 'Were created by new editors' list and done 34 patrolls in that time - which included correcting a few poor patrols. I think that's about average for an experienced reviewer, and it's as much as I can stomach for a day on NPP. A triage approach might seem to be more productive (indeed, during development, the code name for Page Curation was New Page Triage), and without getting careless with my patrolling, I'm certainly no longer spending much time any more improving or rescuing any basically 'no hope' articles for authors who can't be bothered to read WP:My first article.
Triage, as you put it, MrX nevertheless needs to be done therefore with an accent on caution while probably leaning 'delete' (PROD, AfD, etc) rather than keep, which will unfortunately not find favour with the inclusionists. This won't matter quite so much at 'Were created by new editors' list where at least 80% of all the new article are created without the slightest consideration at all for what could even be broadly construed as 'Wikipedic'.
Yes, I agree it would be interesting to know how we got here, but talking about it without doing anything is not going to resolve it and like you and many others I agree that ACTRIAL or something similar should be implemented. The fears expressed by the WMF that it would lose us new editors are no worse than the new editors who are being bitten by the newbies and inexperienced patrollers who the community demanded should still be allowed to tag new pages. Fortunately, their patrolls now have to be checked by accredited reviewers, but a lot of the damage to good faith creators is already done. Kudpung กุดผึ้ง (talk) 01:40, 11 May 2017 (UTC)
  • IDK why pages aren't gettign marked as patrolled, by default, when you tag an article or CSD it, it gets accepted. I had a discussion with another patroller about this once because the logs looked liek I had accepted some pretty poor articles, but it was automatic. So either other peoples' Page Curator works different than mine, or some people are going back and un-accepting it after tagging it. I want a keyboard shortcut to advance the queue. d.g. L3X1 (distant write) 13:44, 12 May 2017 (UTC)
I think they aren't getting patrolled by default because people are either using Twinkle or manually tagging rather than using the page curation toolbar. Obviously if you manually tag it won't automatically be patrolled, and I don't think Twinkle automatically patrols, but as I don't use Twinkle I'm not sure. I'm also pretty sure some people are going back and unreviewing it after tagging it, and I have done that myself if I really wanted a second set of eyes on the article - but it's rare that I do that. ~ ONUnicorn(Talk|Contribs)problem solving 18:16, 12 May 2017 (UTC)
Strictly speaking, 'patrolled' means that an article has been reviewed by an accredited reviewer irrespective if it is to be kept or deleted. Hence articles that have been marked for deletion have been patrolled. The confusion is the belief that the word 'patrolled' means 'accepted' and the NO_INDEX tag has been removed, which in the case of articles marked for deletion is not the case, the NO_INDEX tag remains intact. Only accredited reviewers can mark pages as 'patrolled'. The quality of reviewing by non-accredited 'patrollers' still leaves a lot to be desired. Although I am trying really hard to distance myself from all things 'New Page Patrol' except for patrolling new pages myself, I have had to ask at least half a dozen new users this week to refrain from tagging new pages. Kudpung กุดผึ้ง (talk) 02:23, 13 May 2017 (UTC)

After steady increases to just over 22,000, we are now holding it at bay at just under 22,000. Do we know why it's no longer dramatically increasing? It's still not good, but it's very slowly going in the right direction and no longer increasing. Boleyn (talk) 18:23, 6 June 2017 (UTC)

We now have a core group of 50 people who between them review approximately 20,000 articles in a month. Reviewing is hard, but we're getting better at what we're doing. If the WMF sponsored a group of our most experienced reviewers to go to Wikimania and produce a set of training materials, and brought another 50 up to speed, we could double our output and crush the backlog in a month. That's probably not quite realistic. Basically, to hold the backlog at 20,000 with 1200 new articles per day you need 50 reviewers who do 24 reviews per day, or 100 who 12 per day or 400 reviewers who do 3. We currently have about 15 reviewers who can do more than twelve reviews per day. If we can bring that number up to 150, then the backlog is gone in a month. A more realistic scenario would look like this:
  • 50 reviewers who do 12 reviews per day
  • 50 reviewers who do 8 reviews per day
  • 50 reviewers who do 6 reviews per day
  • 50 reviewers who do 5 reviews per day
  • 50 reviewers who do 4 reviews per day
  • 50 reviewers who do 3 reviews per day
  • 50 reviewers who do 2 reviews per day
  • 50 reviewers who do 1 review per day
That's a total of 1300 reviews per month. Assuming 1200 new articles, the backlog can be gone by the middle of December of this year. I think that's an ambitious, but attainable goal, and knowing that we can get there is a great motivator.
A note related to my experience: Working at the back of the log, knowing that I can't even find unreviewed articles that are approching 30 days old and about to lose the NOINDEX flag, because I first have to wade through endless lists of sports stats pages that require a ridiculous amount knowledge of arcane notability guidelines and are already indexed by google makes me feel utterly hopeless. I can't wait for the day that we run a robot that sends any unsourced article (back) to Draft without human intervention. For now, if I could have better filters in the curation tool, that would help too. Mduvekot (talk) 20:15, 6 June 2017 (UTC)
Last week, the WMF published a report about the backlog: "New Pages Patrol Analysis & Proposal", and we've been talking on that talk page. There's currently about ~330 active reviewers per month, down from ~950 last year.
Unfortunately, a problem with the "x reviews per day" plan is that the pages aren't equal. Some are easy to review, some are very hard -- and people tend to work on the easy ones. Over time, that creates a core backlog of pages that are too time-consuming for an individual reviewer to bother. (For example: a page that has a mix of good sources and bad sources, but they're in Spanish, and you don't know enough about the subject.) Telling people "you need to do 12 reviews a day" is tough, if one good review could take all afternoon. DannyH (WMF) (talk) 20:35, 6 June 2017 (UTC)
@DannyH (WMF) and Mduvekot: both of you are right. "There is a lot difference in theory, and practical." Then someone said "There a lot difference in practicle between theory, and practicle than it is in theory. The backlog is like the star. Lighter elements get at the top, and fuse/burn out easily. Heavier elements gets in the core, making the core denser with time. But I think M is right, we should at least come up with a strategy, or we should at least keep trying. —usernamekiran(talk) 20:44, 6 June 2017 (UTC)

WMF: New Pages Patrol Analysis and Proposal

Latest comment: 7 years ago9 comments5 people in discussion

Hi everyone, I've posted the report that Toby mentioned last week would be coming from the WMF. It's a collaborative report by several Foundation teams, including Community Tech, Editing, Reading, Research, Analytics and Community Engagement, looking at the current New Page Review backlog, and proposing some changes that we think would help reviewers handle the backlog in a sustainable way.

It's a long piece with a lot to cover, and I'm hoping that it kicks off some new discussions between the Foundation and the New Page Reviewers. We see new page patrolling as a really important job, and we want to help make it work better. While we were working on this proposal, we fixed some bugs and created the new database report that Kaldari and MusikAnimal mentioned above. We'd like to do more, in partnership with this team. I'm looking forward to hearing what you think. -- DannyH (WMF) (talk) 22:40, 31 May 2017 (UTC)

  • Thank you. I've created a thread mentioning this report at Wikipedia talk:The future of NPP and AfC, alerting people who might watch that page but not this one of the reports existence. I also think that page might be a better place to have the followup conversation than here so as to keep this page focused on practical issues while giving the future and meta part of this project its own talk page. I wanted to raise the point to see what others think since we do have a project-space talk page devoted to the future of this endeavor. TonyBallioni (talk) 00:00, 1 June 2017 (UTC)
I have started a discussion at the talk page of the report, and I see that there are already a couple of new sections. I suppose WT:NPPAFC may be a better venue to keep the discussion centralized.- MrX 00:20, 1 June 2017 (UTC)
My thoughts were that the talk page of the report would be ideal for questions about the actual text and numbers, but that a project-space page other than the essay would be better for the "moving forward" bit. TonyBallioni (talk) 00:22, 1 June 2017 (UTC)
Great, thank you both. I'm happy to have the conversation in whatever place(s) people want to talk -- I'll keep an eye on all of those places. :) DannyH (WMF) (talk) 00:37, 1 June 2017 (UTC)
Would discussion at NPPAFC be a forum open to all comers? I had impression it was primarily intended as a narrowly-drawn working group. I think it'd be good to have this discussion broadly, and don't want to encroach on a space others set up with more specific purpose/parameters. Innisfree987 (talk) 01:00, 1 June 2017 (UTC)
I think it'd be open there (all Wikipedia pages are), but WT:NPR (this page) is more visible. My main reason for thinking it'd be better to have the conversation in a project space other than the essay is because the conversation is going to evolve over time, and holding the larger on the talk page of a report that will become dated doesn't make much sense to me. I'm also open to any approach that works, just thought I would raise the question. TonyBallioni (talk) 02:18, 1 June 2017 (UTC)
  • Innisfree987 one of the great things about Wikipedia is that it is collaborative and all its sub-projects welcome comments by other users even if the project has a small core of particularly active editors driving it forward.
Thus, here I assume this talk page to be the more general talk page for NPP/NPR, and any in-depth discussion about the development of the system would perhaps best be hosted at Wikipedia talk:The future of NPP and AfC which was created for that purpose. For one thing, it has lists of all the important previous discussions which people should visit if they want to be up to speed on the reasons why we created the NRR right in the first place, and why after years of silence, the community and the WMF are now finally taking the situation seriously. Kudpung กุดผึ้ง (talk) 05:20, 1 June 2017 (UTC)

Guidelines for granting suggestion

Latest comment: 7 years ago1 comment1 person in discussion

I suggest that the guidelines for granting be clarified for users that were mostly inactive but still with the required total number of edits, and are coming back to editing more in a recent period and requesting permission to be a new page reviewer. I noticed it with my own request and looking in the archives with that of another user in a similar situation. To be clear, I am not asking that the guidelines for granting be changed to accommodate us, but rather that they be clarified on the project page so that future users in a similar situation can know prior to requesting permission. Thanks! - SantiLak (talk) 04:42, 8 June 2017 (UTC)

Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia_talk:New_pages_patrol/Reviewers/Archive_7&oldid=1137426935"