→NPP at ANI: archived |
Shani (WMF) (talk | contribs) Tag: Reply |
||
Line 585: | Line 585: | ||
::*@{{u|Kudpung|Kudz}}, they recorded the meeting, so you can watch it as soon as they get it uploaded. As you may have anticipated, it was more of a trustee vanity meeting where they provided updates of their accomplishments and introduced some of the programs they're working on, including the new Enterprise. I wasn't even given the opportunity to speak. The panel harvested and responded to a few text questions originating from the chat (YouTube?) portion of the meeting. The meeting chair (Elena Happen) harvested the questions and the respective trustees, or project leaders responded on Zoom. I thought it odd because a few days before the meeting, WMF sent us a request via the e-list to submit our questions in advance of the meeting so they could prepare, and get them worked into the agenda. One by one the agenda items were presented and discussed only by the trustees/project leaders. An hour+ later, I wasn't getting even a hint that my reason for attending was going to be discussed so I sent Elena a private text asking when the pre-submitted questions would be addressed. A short while later she responded and publicly referenced the presubmitted questions, but started and ended with mine, saying it was not an agenda item but went ahead and asked Shani Evenstein to respond. Long story short, Shani admitted that she knew about the NPP backlog, yada yada, but it appeared to me that she neither had a clue as to what I wanted to discuss nor did she seem interested in doing so. She ended by basically saying the issues at NPP are not in WMF's purview, and that it was something the community had to figure out - discussion over. Elena moved to another topic, so I sent Shani a private text asking if I could discuss it with her via email, and she texted back with an email address. [[User:Atsme|<span style="text-shadow:#F8F8FF 0.2em 0.2em 0.2em,#F4BBFF -0.2em -0.2em 0.2em,#BFFF00 0.4em 0.4em 0.5em;color:#A2006D"><small>Atsme</small></span>]] [[User talk:Atsme|💬]] [[Special:EmailUser/Atsme|📧]] 01:28, 15 July 2022 (UTC) |
::*@{{u|Kudpung|Kudz}}, they recorded the meeting, so you can watch it as soon as they get it uploaded. As you may have anticipated, it was more of a trustee vanity meeting where they provided updates of their accomplishments and introduced some of the programs they're working on, including the new Enterprise. I wasn't even given the opportunity to speak. The panel harvested and responded to a few text questions originating from the chat (YouTube?) portion of the meeting. The meeting chair (Elena Happen) harvested the questions and the respective trustees, or project leaders responded on Zoom. I thought it odd because a few days before the meeting, WMF sent us a request via the e-list to submit our questions in advance of the meeting so they could prepare, and get them worked into the agenda. One by one the agenda items were presented and discussed only by the trustees/project leaders. An hour+ later, I wasn't getting even a hint that my reason for attending was going to be discussed so I sent Elena a private text asking when the pre-submitted questions would be addressed. A short while later she responded and publicly referenced the presubmitted questions, but started and ended with mine, saying it was not an agenda item but went ahead and asked Shani Evenstein to respond. Long story short, Shani admitted that she knew about the NPP backlog, yada yada, but it appeared to me that she neither had a clue as to what I wanted to discuss nor did she seem interested in doing so. She ended by basically saying the issues at NPP are not in WMF's purview, and that it was something the community had to figure out - discussion over. Elena moved to another topic, so I sent Shani a private text asking if I could discuss it with her via email, and she texted back with an email address. [[User:Atsme|<span style="text-shadow:#F8F8FF 0.2em 0.2em 0.2em,#F4BBFF -0.2em -0.2em 0.2em,#BFFF00 0.4em 0.4em 0.5em;color:#A2006D"><small>Atsme</small></span>]] [[User talk:Atsme|💬]] [[Special:EmailUser/Atsme|📧]] 01:28, 15 July 2022 (UTC) |
||
::::{{u|Atsme}}, Shani Evenstein is Vice-Chair of the Board of Trustees and her home Wiki is [[:en:Hebrew Wikipedia|Hebrew Wikipedia]]. Needless to say therefore that I am not surprised that she is not familiar with NPP and gave you a throwaway answer. NPP is ''very much'' in the WMF's purview - they developed it and nobody else can address the issues, nor can the local en.Wiki NPP community be expected to. The WMF is currently looking into some of the NPP issues at Phab but there are no signs of an ETA. [[User:Rosiestep]] is also a board member and it may be worth contacting her. [[User:Kudpung|Kudpung กุดผึ้ง]] ([[User talk:Kudpung|talk]]) 03:44, 15 July 2022 (UTC) |
::::{{u|Atsme}}, Shani Evenstein is Vice-Chair of the Board of Trustees and her home Wiki is [[:en:Hebrew Wikipedia|Hebrew Wikipedia]]. Needless to say therefore that I am not surprised that she is not familiar with NPP and gave you a throwaway answer. NPP is ''very much'' in the WMF's purview - they developed it and nobody else can address the issues, nor can the local en.Wiki NPP community be expected to. The WMF is currently looking into some of the NPP issues at Phab but there are no signs of an ETA. [[User:Rosiestep]] is also a board member and it may be worth contacting her. [[User:Kudpung|Kudpung กุดผึ้ง]] ([[User talk:Kudpung|talk]]) 03:44, 15 July 2022 (UTC) |
||
:::::Hello, {{ping|Atsme}} and {{ping|Kudpung}}. Thank you for all you do, for attending the last Open Conversation with the Board of Trustees (BoT), and for sharing your view in these posts. There's a lot to consider. Note, we might have different points of view, but I will do my best to answer in good faith and with the facts so people can have a full picture. A few points to consider: |
|||
:::::* The talk was live streamed automatically on YouTube. [https://www.youtube.com/watch?v=FYg9sJ4Ml3M&t=4180s&ab_channel=WikimediaFoundation This link is to the exact question mentioned] (1:09:45). Check what I said and if you share quotes, please ensure accuracy. I did not say “NPP are not in WMF's purview”, but rather “…not the ''BoT'' purview”. There is a big difference. I also stated the BoT does not deal with operations, but with strategy, and that such matters, such as NPP, are not related to Board work. I added that there is a true desire among WMF staff to help. I believe the Product department has already replied, as they are the ones with the expertise to do so, not the BoT. |
|||
:::::* In the past, the BoT was criticized by the community for being non-transparent, at times referred to as “Black Box”, the community not knowing what it does, and at times why it makes the decisions it makes. The fairly new Community Affairs Committee (CAC), which I serve as founding Chair of, has several goals (see [[foundation:Community_Affairs_Committee_Charter|Charter]]), one of which is as a platform to help the BoT be more transparent and allows dialogue between the community, WMF, staff and the BoT. Part of our efforts is conducting Open Conversations with Trustees; and opening askcac<nowiki>{{@}}</nowiki>wikimedia.org email, available for anyone. The BoT has never been as transparent and accessible as it is now. This is not a hoax, nor a publicity stunt, and certainly not coming from a place of vanity. We take our work very seriously. Various people, including Trustees (who are volunteers as well, BTW), spend many hours making these meetings as meaningful and inclusive as possible, with true commitment by the whole BoT, and our CEO and staff, to be as transparent as possible, to have honest and meaningful conversations (even if tough at times). We do our best to listen, address issues that have not been addressed (at times for years), and hopefully, also rebuild trust. |
|||
:::::* For the Open Conversations to be efficient, effective and inclusive, they have to be well organized in advance. We prepare for these meetings and send an agenda of main topics, and post it on Meta, a week in advance, and make sure to go through these agenda items. Not only Trustees and staff speak. The main agenda items are followed by a Q & A on the main topics (as opposed to general updates at the beginning). We also have 20-30 minutes in each meeting for open questions in the room (and on YouTube), or pre-submitted questions - not part of the agenda topics. We view it as the best way to run such meetings, but listen to feedback and make changes to make the format more effective almost every time. |
|||
:::::* If a community member does not like the format - they can let us know. There are proper channels to do that, like the CAC email, and the survey at the end of each meeting. We are on record for already changing multiple things to the request of the community. We’ll seriously consider a community member’s request to change the format, and answer questions in a timely manner. |
|||
:::::* Attendees were told they could write questions/comments in the chat or raise a hand and request to speak. Perhaps you hadn't arrived at the Zoom meeting when that was stated, so we'll remember to mention it more than once in future meetings. If the point is that it was important for you to ask the question live in your own words, then point taken. In future meetings an effort will be made to allow people to ask themselves, if they are in the room. |
|||
:::::* On a regular basis, the BoT receives questions from people in the movement on topics that have nothing to do with Board responsibilities. One of the goals of these Open Conversations is better informing the community what the BoT does and what it does not. Having a mechanism for pre-submitted questions allows the BoT to forward questions not within its remits to staff, to give more informed answers. Sending your question in advance, we were able to give an answer that included our Product department perspective. When any Trustee answers a question, these are official answers. They are usually checked and researched beforehand. I happen to know EnWiki quite well, including this issue and others, but even if I didn’t, I would not speak without checking first. |
|||
:::::* The BoT has limits to what it can do. “We understand, we care, but it’s not related to our work and here’s where you need to go for answers” may not have been a desired answer, but it is a legitimate answer nonetheless. It’s how organizations work - each part has a different role. This question did not touch a topic under the purview of the BoT. Does that mean the Board doesn’t care? Absolutely not. We care deeply, certainly enough to make it heard, ask staff to give more context and think strategically about longer term solutions. But the BoT does not have ‘magic’ solutions. Not always, anyway. |
|||
:::::* Trustees serve the WMF and the Movement as a whole, across all projects and all stakeholders. Any Trustee would have given the same answer, no matter what their home wiki is as the answer that was given was prepared in advance, after checking it with relevant staff. It may not have been the answer you wanted, but it was an honest answer. The BoT understands and cares about anything to do with our most important stakeholder - the volunteers who run the projects. We will continue to look for solutions and support the communities in our Movement in any way we can, even if we can’t fix everything, for everyone, right away. |
|||
:::::* Discussions do not have to be over when an Open Conversation meeting is finished. I ended my reply with an open invitation to share more details. I shared the askcac@wikimedia.org email (the proper channel to get in touch with us) multiple times. Atsme, you have also reached out privately, and I immediately responded that I am happy to continue this via email. And obviously, discussion can continue here. Hope this helps. |
|||
:::::<nowiki>Best, ~~~~</nowiki> [[User:Shani (WMF)|Shani (WMF)]] ([[User talk:Shani (WMF)|talk]]) 14:01, 16 July 2022 (UTC) |
|||
:Hey folks, a couple of things in the horizon which might be helpful here (and for others dealing with the same issues). |
:Hey folks, a couple of things in the horizon which might be helpful here (and for others dealing with the same issues). |
Revision as of 14:01, 16 July 2022
Tutorial | Discussion | New page feed | Reviewers | Curation tool Suggestions | Coordination |
NPP backlog
1, 2, 3, 4, 5, 6, 7, 8, 9, 10 |
This page has archives. Sections older than 14 days may be automatically archived by Lowercase sigmabot III. |
This page has archives. Sections older than 14 days may be automatically archived by Lowercase sigmabot III. |
Proposal: increase unreviewed new page search engine NOINDEX duration
Alexandermcnabb had a great idea above. Articles should not be searchable in 90 days automatically, but ONLY AFTER they have passed NPP.
I think this is actionable, it is probably pretty easy to file a Phabricator ticket tagged with Wikimedia-Site-requests to change $wgPageTriageMaxAge = 365;
. Should I create a WP:VPR for this? Is 365 a good number of days to bump it up to? (currently at 90 days) –Novem Linguae (talk) 18:55, 17 June 2022 (UTC)
- This is a great idea; by definition, anything still in draft is not ready to be searchable. Chris Troutman (talk) 19:08, 17 June 2022 (UTC)
- This isn't about drafts, I think: it's about mainspace articles that haven't been patrolled yet. I think drafts already aren't searchable. Extraordinary Writ (talk) 19:11, 17 June 2022 (UTC)
- I agree that this would be a good idea; hopefully it shouldn't be too controversial. Do you know where the 90-day limit was decided? All I could find was this RfC, which seems to suggest that all unpatrolled pages should be noindexed. Extraordinary Writ (talk) 19:11, 17 June 2022 (UTC)
Technical limitations?
- I think there are some technical limitations; I remember a similar discussion where it was revealed that the backlog for redirects is capped at 30 days because the page triage system cannot handle the amount of redirects that would pile up in the full 90 days. signed, Rosguill talk 19:31, 17 June 2022 (UTC)
- @DannyS712, would you be able to weigh in on any technical limitations to raising $wgPageTriageMaxAge? Is there some kind of database issue that makes this not as easy as changing a variable? –Novem Linguae (talk) 20:16, 17 June 2022 (UTC)
- @MusikAnimal, would you be able to weigh in on any technical limitations to raising $wgPageTriageMaxAge? Is there some kind of database issue that makes this not as easy as changing a variable? –Novem Linguae (talk) 23:02, 17 June 2022 (UTC)
- @Novem Linguae There's no technical concern with increasing $wgPageTriageMaxAge as far as I can tell. gerrit:356781 points to this discussion which implies it was created for this very reason. — MusikAnimal talk 05:35, 18 June 2022 (UTC)
- That was an interesting discussion. It gives the reason for 90 days - a concern that someone could "vandalize" an important article by slipping in a NOINDEX. To mitigate that, NOINDEX is ignored on "old" articles and only works on "new" articles. There was an assumption that articles would be patrolled within 90 days. MB 08:01, 18 June 2022 (UTC)
- That could be handled separately by an edit filter that detects edits introducing NOINDEX in mainspace. MarioGom (talk) 08:30, 18 June 2022 (UTC)
- Hmm. I wonder, how tightly coupled are $wgPageTriageMaxAge and __NOINDEX__? I assumed they were separate, each having their own max duration, but it's possible that their durations are controlled by the same setting. Or maybe a wgPageTriageMaxAge = 365 would place a NOINDEX but then the NOINDEX would still be set to 90 and that would defeat the purpose.
- I also wonder if it'd be technically feasible to go over 365 days. One reason I chose that number is I know that the SQL table pagetriage_log_table only keeps 365 days of data,[1] and I wonder if going over that might cause problems. –Novem Linguae (talk) 08:51, 18 June 2022 (UTC)
- Ping MusikAnimal. Was wondering if you could weigh in again, if you know offhand. Thanks! –Novem Linguae (talk) 18:05, 18 June 2022 (UTC)
- @Novem Linguae I'm not aware of a time limit on __NOINDEX__, or at least it doesn't seem to be documented. I didn't read all of the lengthy discussion but it seemed at the time, there were separate, unrelated issues where basically NOINDEX wasn't working as expected. That appeared to have been resolved as well. $wgPageTriageMaxAge acts completely independently, judging by the code. — MusikAnimal talk 21:40, 20 June 2022 (UTC)
- That could be handled separately by an edit filter that detects edits introducing NOINDEX in mainspace. MarioGom (talk) 08:30, 18 June 2022 (UTC)
- That was an interesting discussion. It gives the reason for 90 days - a concern that someone could "vandalize" an important article by slipping in a NOINDEX. To mitigate that, NOINDEX is ignored on "old" articles and only works on "new" articles. There was an assumption that articles would be patrolled within 90 days. MB 08:01, 18 June 2022 (UTC)
- @Novem Linguae There's no technical concern with increasing $wgPageTriageMaxAge as far as I can tell. gerrit:356781 points to this discussion which implies it was created for this very reason. — MusikAnimal talk 05:35, 18 June 2022 (UTC)
- I think there are some technical limitations; I remember a similar discussion where it was revealed that the backlog for redirects is capped at 30 days because the page triage system cannot handle the amount of redirects that would pile up in the full 90 days. signed, Rosguill talk 19:31, 17 June 2022 (UTC)
Sources
|
---|
Break
- This makes sense, if it's technically feasible. MarioGom (talk) 19:43, 17 June 2022 (UTC)
- Is it possible to have a max age based on categories? BLPs, for instance, shouldn't be indexed at all until reviewed. ScottishFinnishRadish (talk) 19:45, 17 June 2022 (UTC)
- The biggest issue may not be technical. This is a quote from Kudpung: "otherwise you'll just have to make an argument at the WMF to extend the un-patrolled period, and believe me, that would be no easy task, even if you have friends there and meet them personally - been there, done that" — Preceding unsigned comment added by MB (talk • contribs) 17:16, June 17, 2022 (UTC)
- I'm in favor, if it's technically feasible. Not sure what to do if WMF is problematic, but I'd hope we could work through any issues. Not sure what the issue might be - if it's "that's the way it's always been," then why hide IP addresses that have always been open, eh? Geoff | Who, me? 21:24, 17 June 2022 (UTC)
- It helps, especially as applicable to PE in that the client can't see the article so no payment will be forthcoming. Atsme 💬 📧 21:30, 17 June 2022 (UTC)
- I agree, rsjaffe, we still need to patch the hole that allows unsourced articles to slip into mainspace, and also work on a more efficient system for handling redirects. Atsme 💬 📧 00:08, 18 June 2022 (UTC)
- Now that I think about it, changing this will actually help. Right now, no one other than the reviewers has "skin in the game". No one else has an incentive to fix this. With NOINDEX, we can motivate those who care about search engines. Might provide a touch of support to meaningful changes. — rsjaffe 🗣️ 00:14, 18 June 2022 (UTC)
- I agree, rsjaffe, we still need to patch the hole that allows unsourced articles to slip into mainspace, and also work on a more efficient system for handling redirects. Atsme 💬 📧 00:08, 18 June 2022 (UTC)
- One of my favorite sayings is that deadlines spur action. So would this mean that some patrolling that happens because people are concerned about a page being indexed wouldn't happen or would it allow a more stable operation of the queue before something gets indexed? Best, Barkeep49 (talk) 00:48, 18 June 2022 (UTC)
- I've given up on deadlines. I patrol and find 1) stuff that belongs in Wikipedia, 2) bad stuff (e.g., UPE, sockpuppetry, obvious non-notable). Anything in the middle, particularly if it is sports related, I just pass on and let it age out. An interesting thing I did was look at the queue starting 31 days out. articles that are not obviously notable in 1) sports category or 2) foreign-language-reference-supported predominate. — rsjaffe 🗣️ 01:25, 18 June 2022 (UTC)
- My intention was that people who WANT their article to get any attention (ie: Search) would ensure the article is in fit condition for NPP to validate them. At that point (and not automatically after 90 days), the article would be searchable and therefore properly 'live'. This both spurs creators to ensure pages are up to par and ensures that pages that aren't are not searchable and therefore limits the need to send articles to draftification (easing the burden on AfC), limits the need to go to AfD except in the most extreme/certain to pass cases and ensures that people aren't searching WP and finding utter rubbish. Additionally, lifting the burden for performing BEFORE on NPP means that the article AS PRESENTED is patrolled - not the notional article that could have been should the creator have actually looked for the sources the article needs - reducing burden on the small number of patrollers and increasing the requirements for the large number of page creators to actually create viable pages. Thanks, Novem Linguae for picking it up. Best Alexandermcnabb (talk) 14:30, 18 June 2022 (UTC)
- I've given up on deadlines. I patrol and find 1) stuff that belongs in Wikipedia, 2) bad stuff (e.g., UPE, sockpuppetry, obvious non-notable). Anything in the middle, particularly if it is sports related, I just pass on and let it age out. An interesting thing I did was look at the queue starting 31 days out. articles that are not obviously notable in 1) sports category or 2) foreign-language-reference-supported predominate. — rsjaffe 🗣️ 01:25, 18 June 2022 (UTC)
- @Barkeep49, I'm thinking that if exposure is the main purpose of creating an article, then it holds to reason that if it isn't being seen (indexed), the creator will be incentivized to fix it. OTH, if it's a BOT creation, or one the creator doesn't give a flip about, then it's highly unlikely that (a) it's worthy of inclusion or (b) that the issues will actually be fixed; both of which warrant CSD or PROD. Any article that is notable and worthy of inclusion will be sourced, but let's not waste time on the symptoms and go straight to the root of the problem. An unsourced article never should have made it into mainspace. If we patch that hole, and prevent unsourced articles from slipping through the cracks, a lot of our problems are solved. I'm not sure where the notion of publish it anyway actually originated, but it defies our core content policies, and taxes the very core of NPP volunteerism without fair representation. As rsjaffe put it, the bulk of those who oppose deletion don't have any "skin in the game", or a "dog in the fight". This whole scenario is one of the reasons we need the stats I've asked about – we need a cost vs benefit analysis, so to speak, as it relates to the arguments for keeping unsourced articles that have been denied deletion at CSD & PROD – where did they finally end up? Did they get stuck in the NPP queue – go to AfD and were deleted – or did they actually get published with/without fixing the problems? Crafty editors will also take advantage of redirects to create articles in mainspace without drawing much attention. The redirects, reverts of redirects and those that lead to AfD are another issue – but I sincerely believe it's one we can fix at the root cause - PREVENTION, keep unsourced articles/stubs from ever seeing the light of day. Atsme 💬 📧 14:37, 18 June 2022 (UTC)
- Ultimately if more articles are being made than we have capacity to patrol it's going to be a problem at some point. Pushing out no index would give us more breathing room but not change that underlying issue. So we have to either increase capacity or decrease the number of new articles. So ultimately any increase in the NOINDEX time should be accompanied by some plan to change one of these things. Best, Barkeep49 (talk) 15:20, 18 June 2022 (UTC)
- I'm for decreasing the number of new articles, which is something we can do by heading them off at the pass; i.e. PREVENTION. A BOT could be created to assure each new article going into mainspace has at least 3 citations before it leaves Draft, or is directly created in mainspace, regardless of autopatrolled status. I don't know enough about the "theft" of redirects to address that issue - all I know is that it exists and it's a problem. Wbm1058 can probably explain it far better than I ever could. Adherence to WP:PAG and more CSD & PROD support from our admins will help to self-correct some of the other issues, as will getting more admins trained in NPP reviewing so that fewer CSD & PRODs will be rejected or sent to AfD. Redirects also need attention in an effort to make it more difficult to get a bad article back into mainspace which is another rather substantial prevailing issue. The various discussions at VPP demonstrates where a big part of the issues stem, including ambiguities and misunderstandings of PAGs topping the list. I also noticed that, for whatever reason, some editors don't have a "middle button" - they are either 100% inclusionist or 100% deletionist. What we need are more includelists. Atsme 💬 📧 15:56, 18 June 2022 (UTC)
- @Atsme: I'll give you an example case study – something I caught on my patrols and cleaned up this morning. I won't say exactly how I found it per "beans" though I do have a section on my user page where I list my work queues, and it was something on that list. Untangling the mess there was a time-consuming manual process which would be extremely difficult to automate. My focus has been on occasionally adding more patrols when I stumble across something that wasn't caught by any patrol I try to create a new patrol that finds other cases just like the one I stumbled onto. Aliana was first created 16 December 2007 as an article about an album by Aliana Lohan. This was deleted per Wikipedia:Articles for deletion/Aliana. Aliana was re-created 13 July 2017 as a redirect to Eliana via Articles for Creation. That redirect was usurped by OfTheUsername when they started a new article about Aliana, Texas (OfTheUsername later moved Aliana to Aliana, Texas: Proper naming.). That new article had been shuttled back & forth to draft space. I untangled all the crossed wires, and deemed the Texas planned community to not be a primary topic, so I made Aliana a disambiguation. I'm a "middle button" who generally leaves keep/delete decisions about articles like Aliana, Texas to others. But I did note that AIRIA Development Company might have an interest in promoting their planned community which is unlike the other communities listed on Template:Fort Bend County, Texas. I remember riding my bike through places like Crabb and Clodine on my bicycle rides west of Houston in the early to mid 1980s. OMG, Fulshear, Texas is a city with 16,000+ ppl now? I remember that place as not much more than a country corner with a BBQ joint where we ate lunch either during or after our ride (back then it had a population under 600). – wbm1058 (talk) 17:51, 18 June 2022 (UTC)
- It's a can of worms. Atsme 💬 📧 18:27, 18 June 2022 (UTC)
- bro what. I made the Aliana article independently. I don't work for/connected to AIRIA Development company or affiliates or anything. Sorry about the redirection and stuff I tried to make it a draft but then it acted weird. My bad. OfTheUsername (talk) 18:47, 18 June 2022 (UTC)
- @Atsme: I'll give you an example case study – something I caught on my patrols and cleaned up this morning. I won't say exactly how I found it per "beans" though I do have a section on my user page where I list my work queues, and it was something on that list. Untangling the mess there was a time-consuming manual process which would be extremely difficult to automate. My focus has been on occasionally adding more patrols when I stumble across something that wasn't caught by any patrol I try to create a new patrol that finds other cases just like the one I stumbled onto. Aliana was first created 16 December 2007 as an article about an album by Aliana Lohan. This was deleted per Wikipedia:Articles for deletion/Aliana. Aliana was re-created 13 July 2017 as a redirect to Eliana via Articles for Creation. That redirect was usurped by OfTheUsername when they started a new article about Aliana, Texas (OfTheUsername later moved Aliana to Aliana, Texas: Proper naming.). That new article had been shuttled back & forth to draft space. I untangled all the crossed wires, and deemed the Texas planned community to not be a primary topic, so I made Aliana a disambiguation. I'm a "middle button" who generally leaves keep/delete decisions about articles like Aliana, Texas to others. But I did note that AIRIA Development Company might have an interest in promoting their planned community which is unlike the other communities listed on Template:Fort Bend County, Texas. I remember riding my bike through places like Crabb and Clodine on my bicycle rides west of Houston in the early to mid 1980s. OMG, Fulshear, Texas is a city with 16,000+ ppl now? I remember that place as not much more than a country corner with a BBQ joint where we ate lunch either during or after our ride (back then it had a population under 600). – wbm1058 (talk) 17:51, 18 June 2022 (UTC)
- I'm for decreasing the number of new articles, which is something we can do by heading them off at the pass; i.e. PREVENTION. A BOT could be created to assure each new article going into mainspace has at least 3 citations before it leaves Draft, or is directly created in mainspace, regardless of autopatrolled status. I don't know enough about the "theft" of redirects to address that issue - all I know is that it exists and it's a problem. Wbm1058 can probably explain it far better than I ever could. Adherence to WP:PAG and more CSD & PROD support from our admins will help to self-correct some of the other issues, as will getting more admins trained in NPP reviewing so that fewer CSD & PRODs will be rejected or sent to AfD. Redirects also need attention in an effort to make it more difficult to get a bad article back into mainspace which is another rather substantial prevailing issue. The various discussions at VPP demonstrates where a big part of the issues stem, including ambiguities and misunderstandings of PAGs topping the list. I also noticed that, for whatever reason, some editors don't have a "middle button" - they are either 100% inclusionist or 100% deletionist. What we need are more includelists. Atsme 💬 📧 15:56, 18 June 2022 (UTC)
- Is there a page that shows page creation by user metrics (Daily, Weekly, Monthly)? Perhaps some users could be evaluated and "nominated" for Autopatrol by the community.Slywriter (talk) 15:29, 18 June 2022 (UTC)
- @Slywriter: Wikipedia:Database reports/Editors eligible for Autopatrol privilege maybe? – Joe (talk) 18:36, 18 June 2022 (UTC)
- I discovered that there used to exist an edit filter to flag unsourced new articles, though it was decommissioned in 2012 with the comment "disabling, no real use" (?). There's also a fairly recent unanswered query as to whether the filter should be reinstated. While this wouldn't be a universal solution, I believe (in agreement with the recent post) that warning users who are about to publish an unsourced article would either lead them to add sources or refrain from publishing. The filter also got quite a number of hits before it was deleted, but I believe it preceded ACTRIAL. Would reinstating this filter be at least a partial solution? ComplexRational (talk) 17:10, 18 June 2022 (UTC)
- Interesting discovery. I wonder if it will also catch autopatrolled in mainspace or if it only works at AfC draft to mainspace? It should be a universal catch-all filter. No article should be in mainspace without a minimum of cited RS, even if it's just 2. Could it also catch sources that are unreliable by integrating User:Headbomb/unreliable and maybe even User:SuperHamster/CiteUnseen.js would prove useful? Atsme 💬 📧 17:28, 18 June 2022 (UTC)
- From what I see, the filter's old configuration did not target any specific user groups and was set to work for all mainspace pages, with built-in exceptions for pages that aren't supposed to have sources such as redirects and disambiguation pages. I'm unsure about autopatrolled – but in that case, there could be grounds for revocation of the permission. Regarding the scripts, I doubt they could be implemented in a filter because there's enormous potential for false positives/negatives and there are simply too many cases to cover (someone with more advanced programming skills, feel free to correct me though); reactivating the old filter would simply help in catching the most egregious cases. ComplexRational (talk) 17:53, 18 June 2022 (UTC)
- Quick tech note. The edit filter is computationally expensive and it would probably not be performant to look for more than a short list of the dozen or so most egregious unreliable sources. Which we already have some filters for, e.g. filter 869. –Novem Linguae (talk) 18:12, 18 June 2022 (UTC)
- How is the filter catching no sources? Is it using citation format, or reflist, or something of that nature? Atsme 💬 📧 18:32, 18 June 2022 (UTC)
- @Atsme: It flags any new article that doesn't have any of several character strings that would suggest either the presence of sources or a page not needing sources – this means flagging new pages not having any of <ref> tags, http/https, disambiguation in the title, #REDIRECT, etc. It doesn't discriminate types of sources, only whether any could be present. ComplexRational (talk) 18:45, 18 June 2022 (UTC)
- How is the filter catching no sources? Is it using citation format, or reflist, or something of that nature? Atsme 💬 📧 18:32, 18 June 2022 (UTC)
- Interesting discovery. I wonder if it will also catch autopatrolled in mainspace or if it only works at AfC draft to mainspace? It should be a universal catch-all filter. No article should be in mainspace without a minimum of cited RS, even if it's just 2. Could it also catch sources that are unreliable by integrating User:Headbomb/unreliable and maybe even User:SuperHamster/CiteUnseen.js would prove useful? Atsme 💬 📧 17:28, 18 June 2022 (UTC)
- Ultimately if more articles are being made than we have capacity to patrol it's going to be a problem at some point. Pushing out no index would give us more breathing room but not change that underlying issue. So we have to either increase capacity or decrease the number of new articles. So ultimately any increase in the NOINDEX time should be accompanied by some plan to change one of these things. Best, Barkeep49 (talk) 15:20, 18 June 2022 (UTC)
- That 2012 RfC was a result of a couple of meetings I had with Jorm and Erik Möller over the development of New Page Triage, the working name for the new feed and curation software development. Characteristically Oliver Okeyes (WMF), jumped on the bandwaggon before any volunteers could could start the community discussion, and he later continually tried to block development right through to his leaving the community although there was a clear consensus (WereSpielChequers and Scottywong will remember all this).
- 'No Index until patrolled' was much like Jumpback which Oliver also promised and did nothing about and wasn’t resolved until TonyBallioni stepped in years later and it finally got boxed through at the WishList in 2019. Anyway, Extraordinary Writ, you can all blame me for the 90 days. When I was asked, it's what I suggested thinking it would be enough. A few years later we got ACTRIAL done which greatly reduced the flood of effluent but it wasn't long before the problems with patrollers started.
- Per Barkeep49:
Ultimately if more articles are being made than we have capacity to patrol it's going to be a problem at some point.
Yes, pushing out the 90 days would certainly just be a palliative. First off - and I know you'll all hate me for this - some stats are needed over a sample period: how many articles are kept immediately, how many are draftified, how many are PRODed, how many are CSDd and AfDd? Or did someone do that already and I missed it? Kudpung กุดผึ้ง (talk) 10:23, 19 June 2022 (UTC)
I think that extending the 90 days is beyond logical....knowing what we now know, the short 90 day number defeats the whole purpose of that (good) feature and makes it pointless. North8000 (talk) 10:33, 19 June 2022 (UTC)
- The suggestion is not to push 90 days out, Kudpung, but institute 'No review = No Search' as a flat deal and make that public, possibly with some guidance on notability and sourcing that lets new page creators know that their page will not be searchable until it reaches a minimum standard of notability/sourcing and has been patrolled. I'm not sure where stats help in this - we can all see the scale of the problem and this would apply some systemic leverage for creators to focus on sourcing. This also means sub-standard articles aren't 'rewarded' by being searchable after 90 days whether reviewed or not - when right now we have little hope of getting to articles in 90 days. The best stat I can see is the NPP queue and how little we are managing to reduce it! Best Alexandermcnabb (talk) 13:12, 19 June 2022 (UTC)
- I think it's quite clear, Alex, that at least Barkeep49 certainly concur that a palliative would not have much effect. One advantage, as the potted NPP history I wrote above is intended to demonstrate, is that the en.Wiki and a few others have since asserted their maturity vis-à-vis the WMF and can now instigate their own controls and policies.
'No review = No Search' as a flat deal
is not only an excellent suggestion, but where no amount whipping all the 750 reviewers into action is ever going to work, it's also the only solution. Just do it. It's technically a doddle so just ask at Phab for the switch to be thrown. OTOH, if the Grand Masters of Phab decide you need a community consensus first, you'll have get one. ACTRIAL is a seminal example of changing outdated 'founding principle' policy; we had some very heavy participation and extremely convincing consensus each time we ran a debate for it, but only because our mission statements were armed with a lot of significant stats and very carefully crafted proposals. Now its time for some action. Kudpung กุดผึ้ง (talk) 15:21, 19 June 2022 (UTC)- Now, I can just about get my head around NPP and AfD, and I can even create the odd page (bless him, but it was User:TonyBallioni wot granted me autopatrolled, as you mention him) but this Phab switchy stuff is, to be honest, a bit beyond me. Have you SEEN what a mess I can make with source editing? I'm not sure I'm the person to actually take this one forward, but am perfectly happy to have made the suggestion if others want to activate it - and just as happy to support. But as for anything beyond tinkering with content, I'm really sure it's not my long suit... Best Alexandermcnabb (talk) 15:56, 19 June 2022 (UTC)
- I think it's quite clear, Alex, that at least Barkeep49 certainly concur that a palliative would not have much effect. One advantage, as the potted NPP history I wrote above is intended to demonstrate, is that the en.Wiki and a few others have since asserted their maturity vis-à-vis the WMF and can now instigate their own controls and policies.
- The suggestion is not to push 90 days out, Kudpung, but institute 'No review = No Search' as a flat deal and make that public, possibly with some guidance on notability and sourcing that lets new page creators know that their page will not be searchable until it reaches a minimum standard of notability/sourcing and has been patrolled. I'm not sure where stats help in this - we can all see the scale of the problem and this would apply some systemic leverage for creators to focus on sourcing. This also means sub-standard articles aren't 'rewarded' by being searchable after 90 days whether reviewed or not - when right now we have little hope of getting to articles in 90 days. The best stat I can see is the NPP queue and how little we are managing to reduce it! Best Alexandermcnabb (talk) 13:12, 19 June 2022 (UTC)
- Yes, thank you MB for taking the initiative. Personally, I would have asked for indefinite 'NO INDEX' and see what the WMF offered. It's always better to ask for more and then negotiate down if held against the wall by the throat. That said, if you get 365 days it might just do the trick but I'm wary of being back here again in 5 years and asking for more. Anyway, by then Wikipedia might be dead, and so might I ;) Kudpung กุดผึ้ง (talk) 11:46, 20 June 2022 (UTC)
- I did ask for INDEFINITE (as you know now that you have been to the phab ticket, just clarifying here for others). There is already a suggestion there for 365 days, and the task was renamed to reflect that :( MB 13:46, 20 June 2022 (UTC)
- Yes, thank you MB for taking the initiative. Personally, I would have asked for indefinite 'NO INDEX' and see what the WMF offered. It's always better to ask for more and then negotiate down if held against the wall by the throat. That said, if you get 365 days it might just do the trick but I'm wary of being back here again in 5 years and asking for more. Anyway, by then Wikipedia might be dead, and so might I ;) Kudpung กุดผึ้ง (talk) 11:46, 20 June 2022 (UTC)
Yes, I realise that now. All the fun of the fair, just as I predicted.
The Phab request has now been marked as 'stalled'. I hope enough people are following this because the devs are going to need a lot of convincing this is necessary. Kudpung กุดผึ้ง (talk) 14:03, 20 June 2022 (UTC)
- It also says a "local discussion" is underway. If that means a private discussion among devs, that is not transparent. All our discussions on this are public. MB 14:21, 20 June 2022 (UTC)
- That is classic WMF dev-speak. There is nothing for them to discuss. It's their way of finding another ruse for stalling. They are pretending - like I said above - that the whole thing needs yet another RfC. Oh, I know Phab so well - and Bugzilla, its predecessor. So do Scottywong and The Blade of the Northern Lights. Kudpung กุดผึ้ง (talk) 14:54, 20 June 2022 (UTC)
- Well, to be fair, the phab ticket has only been open for a day or two, and it doesn't look like anyone has said anything like "we're not going to do this." So, I'd cut them some slack and give them a reasonable amount of time to figure out how to make the change without introducing any unexpected consequences. If they do eventually refuse to make the change, there are probably ways for us to make this happen on our own, without help from the devs. It wouldn't be nearly as efficient or elegant, but you could have a bot automatically add a template to all new articles that ensures they're not indexed, and then removes that template once the article is patrolled. It would be a silly way to do it, but it's possible if the devs put up a brick wall. But from what I can tell, we're not yet anywhere near the point of needing to contemplate such actions. Give them a week or so to figure out what they're gonna do. —ScottyWong— 18:04, 20 June 2022 (UTC)
- That is classic WMF dev-speak. There is nothing for them to discuss. It's their way of finding another ruse for stalling. They are pretending - like I said above - that the whole thing needs yet another RfC. Oh, I know Phab so well - and Bugzilla, its predecessor. So do Scottywong and The Blade of the Northern Lights. Kudpung กุดผึ้ง (talk) 14:54, 20 June 2022 (UTC)
Picking a specific # of days for NOINDEX proposal at Village Pump
I opened the phab and asked that the 2012 RFC, which called for unpatrolled articles to be NOINDEXED (indefinitely, until patrolled) be implemented. There was some pretty strong support for that here. In follow-up discussion at Wikipedia:Page Curation/Suggested improvements#Extend NOINDEX beyond 90 days, Novem Linguae and MusikAnimal are suggesting we not go that far and start with 6 months or a year instead of indefinite, and suggest a broader discussion since the RFC was 10 years ago. I don't think more discussion is necessary here, but how about a poll to see if we have a clear local consensus before discussing at VPP: MB 18:50, 21 June 2022 (UTC)
- 180 days
- 365 days
- indefinite
- Anything is fine with me, as long as we as NPPs coalesce around a particular number before we take this to WP:VPR. MusikAnimal recommends 180 days. I originally recommended 365 days.
There is an increase in the ease of the patch if we use an integer number, since indefinite would require custom code.–Novem Linguae (talk) 18:59, 21 June 2022 (UTC)- On Jun 23, someone submitted a patch to allow wgPageTriageMaxAge to be set to infinite easily, so there is no longer a technical barrier to implementing infinite if we desire. –Novem Linguae (talk) 07:52, 23 June 2022 (UTC)
@Chris troutman, Extraordinary Writ, MarioGom, Atsme, Barkeep49, Rsjaffe, Kudpung, North8000, Alexandermcnabb, and Scottywong: Pinging other participants in the discussion, please state your preference as to the time period. MB 19:31, 22 June 2022 (UTC)
- I was going to suggest 180 days when I saw the section title; 365 also seems fine to me and adds more of a buffer in case our backlog situation gets worse. Given that indefinite is apparently more technically difficult, I think it makes sense to pick a cutoff time, even if it's a bit arbitrary. As a sidebar: do we know how the cutoff currently affects articles created from redirects that were created more than 90 days ago? They sit in the back of the queue and it's not clear to me if they get deindexed or not. signed, Rosguill talk 19:43, 22 June 2022 (UTC)
- 365, though I prefer indefinite. I’m thinking about how that would affect the article writers, as I want to see a strong incentive to add notability sources.
- on the other hand, if we fix the workflow, the noindex duration would become less important. — rsjaffe 🗣️ 19:52, 22 June 2022 (UTC)
- 180 is enough - adding that I'm ok with 365 or even indef 03:04, 23 June 2022 (UTC) And I also have questions. Is there a way to check for dupes from time to time while an article is in draft? I've actually come across a few dupes that were in main space and also in draft. Typically, truly notable topics don't just disappear because one attempt failed. I don't see WP running out of articles because we had to draftify a few problematic or unfinished stubs/articles. I was also wondering if there's a way we can promote those types of articles to educators who want to teach the WP editing process, and can make good use of them to show students where the problems are, and how to make them main space ready? Atsme 💬 📧 20:19, 22 June 2022 (UTC)
- 365 Anything less than indefinite creates a perverse incentive for article authors to overwhelm our ability to patrol but I think a calendar year is sufficient time to patrol, especially since implementation of 365 days is apparently technically easier. Chris Troutman (talk) 20:23, 22 June 2022 (UTC)
- 90 days is a big problem.....large amounts of unreviewed articles go past that. Indefinite is best and sort of needed if it's looking at the creation date of the page because the page creation date on articles created by converting a redirect or moving from a draft space is the date of creation of the draft or redirect which can be 10 years ago for a page that just showed up as an article space yesterday. Beyond 180 days you still also have a lot of completely new articles in the NPP cue.....usually ones several NPP'ers looked at and avoided. So indefinite is best, 365 is second best, 180 days is third best. 90 days is far far too short. North8000 (talk) 20:35, 22 June 2022 (UTC)
- I'm opposed to indefinite for philosophical reasons (I think deadlines are helpful to spurring volunteer action and I have a soft spot for our "anyone can contribute" roots) and practical (I don't think the foundation would go for it). No preference between 180 or 365. Best, Barkeep49 (talk) 20:42, 22 June 2022 (UTC)
- 180 with adjustment to filter or a separate report that shows articles over 90 days (actual 90 days, not un-redirects so 2005 articles appear). Between the two, should be able to keep anything egregious out of indexing. Slywriter (talk) 20:53, 22 June 2022 (UTC)
- Slight preference for 365. If the backlog gets beyond that, then the deadline would be the least of our problems for NPP. MarioGom (talk) 21:45, 22 June 2022 (UTC)
- Indefinite or 365 as a second choice if the former is not possible. Someone, sorry I can't recall who, made the excellent point much earlier that article creators - especially UPE, who won't get paid until it goes 'live' - have a much bigger vested interest in creating better articles - I echo User:North8000's concerns about redirects - is the timer indeed article creation or is it reset by queue addition? Sorry, me no teknikal... Two 'sorries' in one sentence - spot the Brit. Best Alexandermcnabb (talk) 04:37, 23 June 2022 (UTC)
- Indefinite or 365 as a second choice if the former is not possible. We should not be offering articles to google if they have not been patrolled, since so much stuff I come across violates core content policies. (t · c) buidhe 05:22, 23 June 2022 (UTC)
- Indefinite or 365 as a second choice if the former is not possible. As there are no chances whatsoever of
'spurring volunteer action'
with or without backlog drives or canvassing for new reviewers. This is a time to be pragmatic and not wax philosophical. 'Indefinite' is not a hurdle and it would avoid having to go back and ask for more next year. As I said earlier, we should not be suggesting anything alternative to 'indefinite', but this RfC unfortunately now opens up the decision by the 'gatekeepers' at the WMF and allow them to beat us down to something that can no longer be negotiated down to when they hold the volunteer community by the throat against the wall - and they will. Kudpung กุดผึ้ง (talk) 07:34, 23 June 2022 (UTC) - Update: it looks as if the devs are going to do 'indefinite' after all. Kudpung กุดผึ้ง (talk) 07:59, 23 June 2022 (UTC)
- yay! >little indef dance< Best Alexandermcnabb (talk) 08:19, 23 June 2022 (UTC)
- Indefinite. I never understood why we had a limit. 90 days, 180 days, a year... it doesn't really make a difference. Articles should not be searchable until they've been checked by at least one other human being. – Joe (talk) 08:55, 23 June 2022 (UTC)
- Indefinite in light of the report that its implementation is technically possible. Unpatrolled pages with potential issues should not be indexed under any circumstances, though I also support a separate report (suggested by Slywriter) for pages older than 90 days so that they don't get lost in the queue and can easily draw attention from multiple reviewers if necessary. ComplexRational (talk) 12:09, 23 June 2022 (UTC)
- BTW I did some looking and also from my sometimes work at the back end of the que ....it looks like that back end of the "regular article" cue is somewhere around 9 months. There it looks like mostly articles that several NPP'ers looked at and decided not to handle. Lot's of what looks like "edge case that should probably go to AFD but I don't want to be the one to decide and do that". Back around a year (or older) it becomes more articles that have a old birth date from their birth as a redirect or draft, but where the actual article is much newer. North8000 (talk) 13:20, 23 June 2022 (UTC)
- Off topic but agree, User:North8000, I've spent the last 10 days at the back end of the horse and the resulting AfDs are at times attracting controversy and a lively debate! Best Alexandermcnabb (talk) 05:28, 24 June 2022 (UTC)
- BTW I did some looking and also from my sometimes work at the back end of the que ....it looks like that back end of the "regular article" cue is somewhere around 9 months. There it looks like mostly articles that several NPP'ers looked at and decided not to handle. Lot's of what looks like "edge case that should probably go to AFD but I don't want to be the one to decide and do that". Back around a year (or older) it becomes more articles that have a old birth date from their birth as a redirect or draft, but where the actual article is much newer. North8000 (talk) 13:20, 23 June 2022 (UTC)
- Indefinite or 365 would encapsulate the maximum size of the NPP queue that I've seen to date. Either one would suit me down to the ground. scope_creepTalk 22:53, 29 June 2022 (UTC)
- 180 days would align with the 6 month limit of draft space. The word Wiki means quick and so talk of years is absurd. And an indefinite limit would be a surrender, encouraging the idea that it's safe to let backlogs climb to infinity. The lack of any sense of urgency would tend to kill motivation so the patrol process would fail even more than it does currently. Andrew🐉(talk) 22:19, 1 July 2022 (UTC)
- 365 days . which will be long enough to deal with the material. "Infinite is more likely to lead to complications. DGG ( talk ) 23:34, 3 July 2022 (UTC)
By my count, no limit (indefinite) has clear consensus (nine including those that said any of the choices were OK, three said 180/365, and no one objected to increasing from 90). I will put a notification at WP:VPR. MB 19:47, 23 June 2022 (UTC)
- No additional comments. Will update the Phab ticket. MB 04:54, 26 June 2022 (UTC)
- I think "indefinite" is a bad idea, slight preference for 180, but OK with 365. Creating and distributing content is in our core mission - that it could be indefinitely hampered because of slow or insufficient volunteers is sort of anti-wiki. If we are going to be fine hosting bad content for an entire year, then asking some search engines to please look the other way past that time is diminishing. — xaosflux Talk 13:13, 29 June 2022 (UTC)
- Follow up. So here is a use-case scenario:
- An editor, Alice, notices we are lacking an article on something she read about Amanita muscaria var. flavivolvata
- Alice writes up a new short article on this, (assuming good faith it is a fine start-class article)
- It sits noindexed - making it harder for the general public to find
- This noindex stays forever until some other volunteer volunteers to approve the page
- Follow up. So here is a use-case scenario:
- Also, as this is a change that impacts almost all editors - as I noted at the just recently listed VP posing, this doesn't seem to be well advertised. — xaosflux Talk 13:14, 29 June 2022 (UTC)
- I thought the idea of this discussion was to workshop a proposal to take to the village pump, at which point it would be widely advertised. – Joe (talk) 13:23, 29 June 2022 (UTC)
- Oh I see, it was taken to VPP. @MB: I think Xaosflux is right, your post at VPP doesn't show sufficient consensus for this idea. It should have been formatted as a clear proposal and advertised at WP:CENT, etc. Instead it reads like you're informing VPP of a decision that's already been made (which isn't the case – the heading of this section is "Picking a specific # of days for NOINDEX proposal at Village Pump"). – Joe (talk) 13:34, 29 June 2022 (UTC)
- Xaosflux and Joe Roe, I echo what Kudpung said below. There was clear consensus at NPP here to reaffirm the prior RFC. I didn't think this had sufficient likelihood of being controversial enough to warrant forking into a separate VPP discussion and/or new RFC. I placed a neutrally worded notice there about this discussion and asked for further comments, of which there have been almost none. This change does not impact editors; people who are concerned about whether their article is indexed are usually trying to promote something which is clearly against our core policies. They should not be rewarded in getting added visibility via search engines after some arbitrary time period because NPP volunteeers are overwhelmed by the quantity of poorly sourced and difficult to review articles. NPP is trying to address the backlog and I do not believe will be any less motivated to do so when this is implemented. We are just trying to improve the encyclopedia by closing this way to get unvetteted articles fully into mainspace. I also note that any newbie who follows our suggestions and uses AFC is blocked from publishing trash, but someone a little more sophisticated bypasses AFC and has a good chance of having their trash live and indexed. This is a step towards equalizing these processes. MB 23:56, 29 June 2022 (UTC)
- I think "indefinite" is a bad idea, slight preference for 180, but OK with 365. Creating and distributing content is in our core mission - that it could be indefinitely hampered because of slow or insufficient volunteers is sort of anti-wiki. If we are going to be fine hosting bad content for an entire year, then asking some search engines to please look the other way past that time is diminishing. — xaosflux Talk 13:13, 29 June 2022 (UTC)
- @Xaosflux and Joe Roe:, IMO this is a purely local NPP issue, and an essential software request. I do not see how
this is a change that impacts almost all editors
; it is directly related to the workload of the NPP process and its inability to keep up with the stream of mostly inadmissible, or at best, articles possibly of relative unimportance, that nowadays make up the bulk of new submissions. ACPERM (also an NPP initiative) greatly reduced the tide of effluent a couple of years ago but it's already grown again to its previous proportions. Indexing for search engines may be a coveted bonus for some creators, but it is not a right. I do not understand why challenging improvements to NPP would be particularly helpful, or why the few genuinely active NPPers should be constantly be made to feel they don't do enough. Prolific reviewers are being lost already through burnout. Kudpung กุดผึ้ง (talk) 22:14, 29 June 2022 (UTC)- Am I reading that currently volunteers are encouraged to quickly patrol new pages due to the current settings, yet by extending the setting they will be less encouraged to quickly process new submissions - especially by extending the setting to forever? Reviewing new pages is indeed important and should be done regardless of the request to external indexing engines - if we are hosting bad information it should be dealt with as soon as possible. — xaosflux Talk 22:57, 29 June 2022 (UTC)
- @Xaosflux: I think you are most likely reading this wrong. Your comment:
'Creating and distributing content is in our core mission '
is correct, but hampering it due to preventing NPP from doing its work by challenging its requests for improvement to the process is, IMHO, decidedly 'anti-Wiki'. When I created the NPP user right in October 2016, it was with the express intent of introducing experienced and high quality patrolling into the system and for no other reason - it was certainly not intended to make it slower! I fail to understand why the WMF is determined to undermine its own objectives. - Hence the local Wikipedias and their task-force volunteers are obliged to take matters into their own hands like they would have done for ACTRIAL if the WMF had not acquiesced after 10 years of bitter wrangling. IMO, WMF employees should not even be participating in these community discussions. Clearly the number of articles in the corpus is a far more important boast for the WMF than the quality and reliability of the content in them. The reason that so many inappropriate or totally unsuitable new pages are submitted is directly due to the WMF's refusal to do anything about it - despite the constant begging for a proper new user welcome page. Compared to some idipendent projects using MediaWiki, the Wikipedia is totally antiquated - a Model T Ford in terms of progress in information technology - and based on ideologies that now, after 20 years have little in common with today's reality: that the English Wikipedia is no longer short of content and the mediocre stuff can either wait or be stopped before it is created. Anyone who cares for Wikipedia and recognisess that NPP is a foundering process, is invited to come up with effective solutions rather than impede them, and to get cracking on implementing them.Kudpung กุดผึ้ง (talk) 23:47, 29 June 2022 (UTC)
- @Xaosflux: I think you are most likely reading this wrong. Your comment:
- Update: I'm not sure if this was what the NPP community really wanted but the rollout that was scheduled by the WMF for next week appears has been successfully blocked. Kudpung กุดผึ้ง (talk) 22:59, 29 June 2022 (UTC)
- Atsme, I don't generally trust the Trustees much more than the WMF. I always got the impressions that the Trustees do the WMF's bidding and their main tool is a rubber stamp. I may be wrong, though, times may have changed. Question of which tail is wagging which dog. I don't follow what goes on there, and I have no time for tedious Zoom meetings. Kudpung กุดผึ้ง (talk) 00:17, 30 June 2022 (UTC)
- @Kudpung: <sigh> Can you expand a bit? What happened and where? Geoff | Who, me? 23:09, 29 June 2022 (UTC)
- @Glane23:, I can understand if not everyone knows what or where Phabricator is (it used to be called Bugzilla), so, <sigh>, here is the link to the Phab task again. Everything you need to know is there → Kudpung กุดผึ้ง (talk) 00:17, 30 June 2022 (UTC)
What's happening now?
- @Chris troutman, Extraordinary Writ, MarioGom, Atsme, Barkeep49, Rsjaffe, MB, North8000, Alexandermcnabb, Scottywong, Novem Linguae, and Buidhe: so what's happening now? Has everything stalled? With or without a backlog drive, 100s of totally unsourced articles in the feed are getting dangerously close to the current 90-day limit. Backlog drives never have a permanent impact and constantly need to be repeated. Maybe it's time to do something else - the ACTRIAL and its ACPERM were a resounding successes - perhaps now pushing 'autoconfirmed' out to XCON with a mandatory use of the Article Wizard might be an option to go for. Anyone can still edit Wikipedia, no one needs a PhD to do it, but as DGG says:
the purpose is not just getting articles; it's teaching editors. This is much more difficult and time consuming, and the existing templates do a notably poor job of it
, which clearly echoes what I was saying above at:The reason that so many inappropriate or totally unsuitable new pages are submitted is directly due to the WMF's refusal to do anything about it - despite the constant begging for a proper new user welcome page. [...] the English Wikipedia is no longer short of content and the mediocre stuff can either wait or be stopped before it is created.
Kudpung กุดผึ้ง (talk) 09:28, 3 July 2022 (UTC)- In addition to the backlog drive, the last two newsletters and Buidhe's recruitment efforts have attracted a number of a new patrollers, and I've been working through this list to find candidates for autopatrolled (I know we both have reservations about that right, but if nothing else it's effective at getting the backlog down). The backlog is going down now: by about 200 articles a day for the last ten days, which if sustained will get us under 10000 again by mid-July and theoretically to zero in a couple of months. As for permanent solutions... extended confirmed is a massive hurdle compared to autoconfirmed and I highly doubt you'd find consensus for an ACXCON. What I'm curious about is why we seemingly get cyclical backlogs like this every two years or so. We talk a lot about the number of reviewers and how many reviews they're doing, but there are other variables that could have just as much of an influence on the backlog: rate of article creation, proportion of articles autopatrolled, time taken to review each article, regularity of reviewing, etc. Maybe getting data on these would generate new ideas. – Joe (talk) 09:57, 3 July 2022 (UTC)
- One issue that confuses us is assuming that the backlog number (which is very essential and useful) indicates more than it does. At any given moment, there are only about 10-20 days worth of manual reviews sitting in backlog. The tiniest shift in our overall "equation" (of incomming articles vs. reviews getting done) causes large changes in the backlog. But it's basically all that we have. Many of the the indicators that would also be very useful are not available. North8000 (talk) 10:38, 3 July 2022 (UTC)
- The swings in the backlog over the last year or two directly correspond to some top reviewers quitting NPP or coming back. The top reviewers I am thinking of are Onel5969 and John B123. Their efforts are appreciated and they are missed. Not sure about trends over a longer time period. –Novem Linguae (talk) 11:01, 3 July 2022 (UTC)
- You are getting to the main huge topic which I avoided trying to get into here.North8000 (talk) 11:37, 3 July 2022 (UTC)
- That's certainly a recurring issue. I forget who corresponds to which spike, but the departure of SwisterTwister, for example, way back in 2016, eventually led to a backlog of more than 22,000 articles. But I think it raises more questions than it answers. Why do we end up depending so heavily on one or two reviewers, who inevitably burn out? When the backlog is low, for example, does one person taking all the low-hanging fruit lessen the engagement of other reviewers and set us up for a crisis down the line? – Joe (talk) 11:38, 3 July 2022 (UTC)
- OK, so I'll dive in. Wikipedia runs on volunteers doing what they enjoy doing, (with some "doing it for a good cause" thrown into that "enjoy" equation) For most reviewers, due to the nature of things, reviewing is an extremely painful, slow process. Then after they hit a certain threshold, where they have fluency in the zillion words of guidelines, policies, venues, have learned that it's a big fuzzy system where they have to make judgement calls, where they need to have a thick skin, and where they realize that they don't have to feel guilty about not reviewing for all of the areas where the article needs fixing / development then it becomes less slow and painful and the evolve into one of the reviewers with bigger numbers who numerically get most of this week's work done. And once in a while they evolve further in all of theses areas and also put in more time they become one of the huge-number rock stars....a handful of them could knock a 14k backlog down to 4k in a week. And then when they leave the opposite can happen. Aside from getting a few things elsewhere in Wikipedia fixed, our best approach would be to get good experienced people started, and then help them develop into the category where it becomes less painful and slow and guilt-ridden to review articles. It's not just about getting the bunch of articles that is in the backlog done, it's about attaining & maintaining the horsepower to keep it in check painlessly. North8000 (talk) 12:21, 3 July 2022 (UTC)
- It's cyclical in that BOTs are being used (they get caught & stopped, but over time, new ones are introduced and resume the spamming), plus we're getting translations from other Wikis by the 100s, and we're getting a lot of new article submissions from South Asia and the Middle East, which creates more redirects and AfDs. Add to that, seasonal sports, new movies and lists. Atsme 💬 📧 15:12, 3 July 2022 (UTC)
- OK, so I'll dive in. Wikipedia runs on volunteers doing what they enjoy doing, (with some "doing it for a good cause" thrown into that "enjoy" equation) For most reviewers, due to the nature of things, reviewing is an extremely painful, slow process. Then after they hit a certain threshold, where they have fluency in the zillion words of guidelines, policies, venues, have learned that it's a big fuzzy system where they have to make judgement calls, where they need to have a thick skin, and where they realize that they don't have to feel guilty about not reviewing for all of the areas where the article needs fixing / development then it becomes less slow and painful and the evolve into one of the reviewers with bigger numbers who numerically get most of this week's work done. And once in a while they evolve further in all of theses areas and also put in more time they become one of the huge-number rock stars....a handful of them could knock a 14k backlog down to 4k in a week. And then when they leave the opposite can happen. Aside from getting a few things elsewhere in Wikipedia fixed, our best approach would be to get good experienced people started, and then help them develop into the category where it becomes less painful and slow and guilt-ridden to review articles. It's not just about getting the bunch of articles that is in the backlog done, it's about attaining & maintaining the horsepower to keep it in check painlessly. North8000 (talk) 12:21, 3 July 2022 (UTC)
- That's certainly a recurring issue. I forget who corresponds to which spike, but the departure of SwisterTwister, for example, way back in 2016, eventually led to a backlog of more than 22,000 articles. But I think it raises more questions than it answers. Why do we end up depending so heavily on one or two reviewers, who inevitably burn out? When the backlog is low, for example, does one person taking all the low-hanging fruit lessen the engagement of other reviewers and set us up for a crisis down the line? – Joe (talk) 11:38, 3 July 2022 (UTC)
- You are getting to the main huge topic which I avoided trying to get into here.North8000 (talk) 11:37, 3 July 2022 (UTC)
- A proper new user welcome page would probably help a bit. Something that is displayed to a user when they're about to create their first new article, pointing them in all the right directions and setting their expectations. Perhaps it's worthwhile for someone to put together a page where we can brainstorm all the requirements for such a page, and maybe even put together a draft of the page itself and present it either to the wider community or WMF. Or maybe this has been done already? I think it's a bit unrealistic to just tell the WMF to "make a better welcome page" without even telling them what problems we're trying to solve or giving any suggestions for what should be on that page. —ScottyWong— 20:45, 3 July 2022 (UTC)
- I did it Scottywong, several times, that why I keep mentioning it. But it's got swept under the carpet and lost in the annals of time and on dead Mac computers, just like the Article Wizard that I painstakingly rewrote and a newbie just went and reverted it all. I have worked successfully with the WMF to get several things done for NPP, but my main complaint is that the volunteers are expected be doing the leg work (i.e. coding) on things that should be paid for, but AFAIK, no one among the 550 paid staff has skills in developing things within the scope of UX and they don't appreciate it very much when someone hands them a beautiful GUI and says, now code this in MediaWiki for us. Kudpung กุดผึ้ง (talk) 21:05, 3 July 2022 (UTC)
- Do any of these drafts still exist somewhere on WP? —ScottyWong— 10:56, 4 July 2022 (UTC)
- @Scottywong: 'swept under the carpet and lost in the annals of time', but I expect I could recreate them easily enough - reluctantly. Kudpung กุดผึ้ง Kudpung กุดผึ้ง (talk) 17:35, 4 July 2022 (UTC)
- Do any of these drafts still exist somewhere on WP? —ScottyWong— 10:56, 4 July 2022 (UTC)
- I did it Scottywong, several times, that why I keep mentioning it. But it's got swept under the carpet and lost in the annals of time and on dead Mac computers, just like the Article Wizard that I painstakingly rewrote and a newbie just went and reverted it all. I have worked successfully with the WMF to get several things done for NPP, but my main complaint is that the volunteers are expected be doing the leg work (i.e. coding) on things that should be paid for, but AFAIK, no one among the 550 paid staff has skills in developing things within the scope of UX and they don't appreciate it very much when someone hands them a beautiful GUI and says, now code this in MediaWiki for us. Kudpung กุดผึ้ง (talk) 21:05, 3 July 2022 (UTC)
- If we follow exactly the NPP instructions, we will never be able to handle things; when we have come near doing so is by using judgement for which articles are fundamentally ok and which aren't, and proceeding accordingly. Judging this takes experience, not just good understanding of the rules. Similarly with editors--it makes not minutes but hours to really teach someone, and I try it more than most, but I have never attempted more than 1 or 2 per week. (I've mostly worked at AfC not NPP the last few years, but I consider that almost identical, except that at NPP its safe to assume 90% of the material is coi.
- I have lately become much less willing to continue. In a few special areas I am the only one handling them fluently, or even handling them at all, is extremely discouraging--and I'm sure many experienced reviewers find themselves similar trying to cope with their own areas of interest almost unaided. The likely result is that many more articles will be mistakenly accepted or rejected. AfD is no substitute, for it requires the at least equal judgment. We will end by discouraging good contributors in unusual areas; we will also be letting through much more junk, but at least those can be dealt with later along with the half million similar junk accumulated over 21 years. Finding new contributors to replace those who get discouraged is much more difficult. I never expected to do this work as long as I've been doing, which is now 15 years. I had earlier expected that by now we would have new people who could take over--and there are a few, but not enough to replace those of us who are leaving--and the ones who do, mostly need further experience and background to work accurately enough. For people like me or Kudpung, WP is perhaps the most worthwhile work we've done in our lives, but we need to gradually stop before we find ourselves forced to. DGG ( talk ) 03:03, 4 July 2022 (UTC)
- Thank you for that DGG, I'm sure it will strike a chord with our long-time reviewers and NPP/AfC activists.Wikipedia has matured and most of the traditional encyclopedic areas are covered and maintained by topic experts who quietly get on with their work. As the Internet becomes more accessible in developing regions and smartphones can be bought for a few dollars, the vast majority of today's new article submissions mainly comprises football (soccer) bios, other sports people and events, Bollywood, hardly intelligible English, vanity pages, spam, and pure junk. This makes the patrolling of new pages a tedious, boring, and soul destroying task and an increasing number of articles being pushed into draftspace.
- Most new users (and some older ones too) resent being taught - believing it's their right to dump what they like in Wikipedia with the expectation that someone will clean their stuff up for them. It might be The Encyclopedia Anyone Can Edit, but editing here is still a privilege and there are rules to be followed. It's hardly a wonder that those who apply for the NPP user right give up so soon, leaving 90% of the work to 10% of the patrollers who then burn out and leave anyway. Like DGG, I'm sure that most patrollers go for the low-hanging fruit and/or topics in their own knowledge areas; I know I do.
- Unfortunately users like DGG and me are no longer spring chickens and very few of the current active reviewers have a long institutional memory or solid experience. As the curation tools get better (and they are a vast improvement on what we were using 12 or 15 years ago) the rate of detection of abuse of privileges gets better and this has led to the alarming discovery that even after the Orangemoody affair a few years ago, we are probably still only scratching the surface. . Kudpung กุดผึ้ง (talk) 05:04, 4 July 2022 (UTC)
July 5 update
It looks like there isn't clear consensus on how much notification/discussion is needed to determine the consensus on this issue. The hangup really is about indefinite, which is favored philosophically by most NPPers to ensure nothing is externally visible by without review. Since there is little to no opposition to extending from 90 days, I just asked at the phab to implement 365 days at least as an interim step. I think we should take that for now if we can get it to give some immediate relief. MB 17:54, 5 July 2022 (UTC)
- Perhaps we can get someone uninvolved to do a compromise closing - it can always be revisited. From above there seem to be 2 very-related issues that are going on: (a) Pages that aren't reviewed for a 90 days fall off the reviewing tool view. I don't think there is anyone that objects to making that longer of itself. (b) Pages that aren't reviewed for 90 days become indexible - that is what has more differing opinions. Now, these appear to be very linked in that (a) seems to just be a view of pages in (b), but if they were divorced from each other (a) being indefinite seems to be fine while values for (b) get hashed out. I don't think the system currently supports that though? — xaosflux Talk 19:04, 5 July 2022 (UTC)
- (a) shouldn't be an issue. Special:NewPagesFeed shows all unreviewed pages regardless of age. I believe this discussion has been entirely about (b). –Novem Linguae (talk) 21:22, 5 July 2022 (UTC)
- @Novem Linguae so there is nothing that will hinder the patrollers from continuing to patrol pages and have a queue of unpatrolled pages? Some of the comments above suggested that this would be hampered. — xaosflux Talk 21:32, 5 July 2022 (UTC)
- Xaosflux, as NL said above, this is entirely about (b). Articles never fall off the reviewing tool view. For example, I just looked at the NPP feed and there are articles going back to 2007 - but those are usually (or probably always) redirects newly turned into articles that still carry the original creation date. The oldest "new" article in the queue is from January 22 of this year. I assure you that in your scenario above, Alice's "fine start-class article" would be reviewed promptly. There are always some reviewers who look at the newest articles and pick off the "lowest hanging fruit". I do that myself in a few topic areas whenever I have a few spare minutes. The old articles that are still unreviewed (and those that we want to remain no-indexed) have usually been looked at by several reviewers, are often stubs, and usually have tags for notability, possible UPE, possible copyvio, etc. These get held up until someone comes along who is willing to spend the time looking for additional sources, translating foreign languages sources, etc. and determine it is fine, or become confident enough to send it to AFD. There has been a lot of VP discussion on shifting the burden for proving notability to the article creator and allowing more liberal draftification of articles that fall short based on their present state (the sources in the article), but that is unlikely to gain consensus. So for the time being, we are stuck here with a lot of difficult articles to review. As having a WP article is so valuable today, there is an endless stream of new articles on NN people and companies. These promotional type articles especially should not be indexed until they have passed NPP. Hope that helps explain things better. MB 23:29, 5 July 2022 (UTC)
- Yes, so there is nothing that is preventing patrollers from patrolling, and this is all about if we should extend the auto-indexing-of-unpatrolled value. That is what I expected was going on - but was trying to be sure to not miss anything in the comments above suggesting that this proposal only impacts reviewers, while it actually has very little impact on their own volunteer workflow - it impacts authors and readers. I contributed above and am personally fine with some of he proposed extensions. — xaosflux Talk 23:39, 5 July 2022 (UTC)
- This is a matter of perspective. This does not impact reviewers' workflow, but it diminishes their work if there is a backdoor way to get a fully-visible WP article. This is another reason (and there are others) for people to say why bother with NPP and go do something else. The only "impact" to readers and authors is that borderline articles are harder to find. IMO, people who care about seeing their article appear in a google search are probably trying to promote something - and that shouldn't be a priority to us. MB 00:23, 6 July 2022 (UTC)
- I already supported extending it above, was just making sure we didn't have something else going on that breaks the workflow unnecessarily. For example, we have a cutoff on other RCP workflows - but we have a much larger technical challenge trying to extend the entire RCP tables much further as the incoming data is much higher. — xaosflux Talk 01:09, 6 July 2022 (UTC)
- You supported extending it only to 180 or 365, and if I am not mistaken, are one of only 3-4, and probably the most vocal, opponents of going indefinite. The phab is on hold pending some indication that there is a "community consensus". Although indefinite is still a topic for discussion, I hope extending to 365 is recognized as having clear support and this, at least, is implemented ASAP. MB 01:42, 6 July 2022 (UTC)
- @MB to be clear, my main being "vocal" part is that there should be current community support for this change, whatever the value. Yes, I expressed a preference for something not-indefinite, however that's just my quiet voice :) The discussion above does seem to be trending to saying 365 is an acceptable first step, and why I suggested someone uninvolved may be able to find a compromise closure. — xaosflux Talk 10:08, 6 July 2022 (UTC)
- You supported extending it only to 180 or 365, and if I am not mistaken, are one of only 3-4, and probably the most vocal, opponents of going indefinite. The phab is on hold pending some indication that there is a "community consensus". Although indefinite is still a topic for discussion, I hope extending to 365 is recognized as having clear support and this, at least, is implemented ASAP. MB 01:42, 6 July 2022 (UTC)
- I already supported extending it above, was just making sure we didn't have something else going on that breaks the workflow unnecessarily. For example, we have a cutoff on other RCP workflows - but we have a much larger technical challenge trying to extend the entire RCP tables much further as the incoming data is much higher. — xaosflux Talk 01:09, 6 July 2022 (UTC)
- This is a matter of perspective. This does not impact reviewers' workflow, but it diminishes their work if there is a backdoor way to get a fully-visible WP article. This is another reason (and there are others) for people to say why bother with NPP and go do something else. The only "impact" to readers and authors is that borderline articles are harder to find. IMO, people who care about seeing their article appear in a google search are probably trying to promote something - and that shouldn't be a priority to us. MB 00:23, 6 July 2022 (UTC)
- Yes, so there is nothing that is preventing patrollers from patrolling, and this is all about if we should extend the auto-indexing-of-unpatrolled value. That is what I expected was going on - but was trying to be sure to not miss anything in the comments above suggesting that this proposal only impacts reviewers, while it actually has very little impact on their own volunteer workflow - it impacts authors and readers. I contributed above and am personally fine with some of he proposed extensions. — xaosflux Talk 23:39, 5 July 2022 (UTC)
- Xaosflux, as NL said above, this is entirely about (b). Articles never fall off the reviewing tool view. For example, I just looked at the NPP feed and there are articles going back to 2007 - but those are usually (or probably always) redirects newly turned into articles that still carry the original creation date. The oldest "new" article in the queue is from January 22 of this year. I assure you that in your scenario above, Alice's "fine start-class article" would be reviewed promptly. There are always some reviewers who look at the newest articles and pick off the "lowest hanging fruit". I do that myself in a few topic areas whenever I have a few spare minutes. The old articles that are still unreviewed (and those that we want to remain no-indexed) have usually been looked at by several reviewers, are often stubs, and usually have tags for notability, possible UPE, possible copyvio, etc. These get held up until someone comes along who is willing to spend the time looking for additional sources, translating foreign languages sources, etc. and determine it is fine, or become confident enough to send it to AFD. There has been a lot of VP discussion on shifting the burden for proving notability to the article creator and allowing more liberal draftification of articles that fall short based on their present state (the sources in the article), but that is unlikely to gain consensus. So for the time being, we are stuck here with a lot of difficult articles to review. As having a WP article is so valuable today, there is an endless stream of new articles on NN people and companies. These promotional type articles especially should not be indexed until they have passed NPP. Hope that helps explain things better. MB 23:29, 5 July 2022 (UTC)
- @Novem Linguae so there is nothing that will hinder the patrollers from continuing to patrol pages and have a queue of unpatrolled pages? Some of the comments above suggested that this would be hampered. — xaosflux Talk 21:32, 5 July 2022 (UTC)
- (a) shouldn't be an issue. Special:NewPagesFeed shows all unreviewed pages regardless of age. I believe this discussion has been entirely about (b). –Novem Linguae (talk) 21:22, 5 July 2022 (UTC)
- With respect, I think you're making a bit of a mountain out of a molehill MB. You asked people to "[pick] a specific # of days for NOINDEX proposal at Village Pump" and we decided on indefinite. So let's just go ahead and make the proposal at the village pump. I think it's very likely to be supported (to be safe we could propose 365 days as an option B per xaosflux) and it's much easier than trying to wikilawyer or badger the MW devs into doing it without a clear, recent consensus. I'd be happy to make the proposal myself, if it's not stepping on your toes. – Joe (talk) 11:08, 6 July 2022 (UTC)
- It may not be realised by everyone here, but 'NoIndex - Indefinite' was an original feature of the Page Curation development process. IMO this doesn't even need debating. It just needs implementing. The people working at, or in charge of Phab may not have been around at the time. Kudpung กุดผึ้ง (talk) 22:30, 7 July 2022 (UTC)
Proposal: Remove redirects to draftspace from review queue
I’ve noticed that a couple of the redirects left behind after draftification were marked as reviewed before they could be deleted. My concern is that, if the article gets recreated in mainspace or through AfC, it would not then go into the queue. Correct me if I’m wrong. Does speedy deletion of the redirect remove the “reviewed status”?
Even if reviewed status is removed by the speedy deletion, it’s still unnecessary work to have those redirects in the queue — rsjaffe 🗣️ 10:53, 25 June 2022 (UTC)
- Yes, it is unnecessary work. If they don't show up in the queue, it would be great. -MPGuy2824 (talk) 11:46, 25 June 2022 (UTC)
- A deleted article (or redirect) will be in the queue if recreated (an exception is if it is restored with UNDELETE as discussed above). I don't see a problem here; a redirect to a draft should be CSDed if found. These shouldn't be in the queue very long as CSDs are usually handled promptly. If the Draftification script is used, the redirect is deleted immediately - so I would think redirects to drafts are rare. MB 15:55, 25 June 2022 (UTC)
- I don’t have page mover rights, so a redirect is left behind with a CSD tag. And deletion takes a while, so I’ve seen the redirects in the queue and, as I stated above, some are getting patrolled. — rsjaffe 🗣️ 16:03, 25 June 2022 (UTC)
- Twinkle is your friend... Best Alexandermcnabb (talk) 16:07, 25 June 2022 (UTC)
- I don’t have page mover rights, so a redirect is left behind with a CSD tag. And deletion takes a while, so I’ve seen the redirects in the queue and, as I stated above, some are getting patrolled. — rsjaffe 🗣️ 16:03, 25 June 2022 (UTC)
- A deleted article (or redirect) will be in the queue if recreated (an exception is if it is restored with UNDELETE as discussed above). I don't see a problem here; a redirect to a draft should be CSDed if found. These shouldn't be in the queue very long as CSDs are usually handled promptly. If the Draftification script is used, the redirect is deleted immediately - so I would think redirects to drafts are rare. MB 15:55, 25 June 2022 (UTC)
- I'd prefer they DO show up in the queue. I have come across some that were not already tagged for CSD and have had to tag them. This would allow those to 'slip through the cracks' as it were. I notice them in particular since I mostly patrol redirects. Happy Editing--IAmChaos 16:30, 25 June 2022 (UTC)
- I agree, that was my point when I said "should be CSDed if found". Having them in the queue is a good way for them to be found and tagged for deletion. MB 17:05, 25 June 2022 (UTC)
- In my opinion, this sounds like reviewer error. A mainspace to draftspace redirect will always get CSDd by the draftify script. Reviewers are not supposed to mark CSDs as reviewed. Perhaps we can let the folks who are marking these as reviewed know that they are not supposed to mark these as reviewed. –Novem Linguae (talk) 18:22, 25 June 2022 (UTC)
- Umm I do this when I find them, to take them out of the queue. Should I just leave them or CSD them? Mccapra (talk) 18:52, 25 June 2022 (UTC)
- Yeah if someone draftifies and it doesnt get tagged csd for some reason, i think we are supposed to csd r2 it and leave unreviewed. –Novem Linguae (talk) 18:55, 25 June 2022 (UTC)
- P.S. On second thought, anytime a patrolled redirect is flipped to an article, it becomes unreviewed. So I dont think this is a loophole that can be exploited. –Novem Linguae (talk) 18:53, 25 June 2022 (UTC)
- By the script yes it will be tagged. IF (and only if) the mover uses the script, otherwise it leaves a redirect behind whoch should be left in the queue for someone to tag R2. Happy Editing--IAmChaos 22:26, 25 June 2022 (UTC)
- Umm I do this when I find them, to take them out of the queue. Should I just leave them or CSD them? Mccapra (talk) 18:52, 25 June 2022 (UTC)
I asked for autopatrol rights specifically to avoid creating unreviewable redirects that need patrol by another NPP'er. I didn't get autopatrol but they gave me whitelist which was supposed to solve this. But it apparently hasn't. Someday I should ask for admin to solve all of these things. North8000 (talk) 22:10, 1 July 2022 (UTC)
- @North8000: You asked for autopatrolled so that redirects you create when moving a page to correct its title wouldn't go into the queue, which is a different situation. The redirects-with-a-CSD-tag created by the draftify script are not technically redirects and therefore not affected by the redirect whitelist. But now that you're on the redirect whitelist, any regular redirects you create should be automatically patrolled by User:DannyS712 bot III. Is that not happening? – Joe (talk) 07:43, 2 July 2022 (UTC)
- @Joe Roe: Thanks. I guess I was imprecise in my (mis)understanding. I thought it was to apply to all redirects created by me. But I saw that I was on the top of the list of people with the largest amount of un-reviewed articles (all of which were those redirects) and getting notices that other NPP'ers have been manually reviewing those redirects that I left. North8000 (talk) 12:58, 2 July 2022 (UTC)
Request: Report of number of unreviewed articles, grouped by creator
I've noticed that when a creator has more than many unreviewed articles in the queue, they are mostly of the same kind (e.g. constituencies of russia OR frogs from south america OR ...) and either are are an easy pass, an easy fail, or there is something missing in all/most of them, which should be mentioned to the creator. Could someone create a report for this, at least once? If it seems useful to other NPRs, we can ask someone who runs a report bot to generate this report every 12 hours or so. -MPGuy2824 (talk) 11:53, 25 June 2022 (UTC)
- There used to be a tool for this, Rentier's NPP browser, but it hasn't worked for some time. It would be really useful, also for identifying candidates for autopatrolled. – Joe (talk) 14:54, 25 June 2022 (UTC)
- Here's a quick query showing unreviewed new pages by user. Obviously having this on a wiki page would be nicer, but it's a start. – Joe (talk) 15:33, 25 June 2022 (UTC)
- Thank you. I've tried to create this query a few weeks back, but quite clearly, my SQL skills need brushing up. -MPGuy2824 (talk) 01:51, 26 June 2022 (UTC)
- Well, using that query you can then filter the feed by username. Happy Editing--IAmChaos 16:32, 25 June 2022 (UTC)
- That's an interesting tool. ---Steve Quinn (talk) 22:53, 25 June 2022 (UTC)
- Here's a quick query showing unreviewed new pages by user. Obviously having this on a wiki page would be nicer, but it's a start. – Joe (talk) 15:33, 25 June 2022 (UTC)
I added this to Wikipedia:New pages patrol/Resources#Reports since this thread will eventually be archived. MB 05:19, 26 June 2022 (UTC)
- Here's a summary of my quick look at the top 20 from that query:
Creator | Count | Common Type |
---|---|---|
Gardenkur | 115 | Municipal corporations in India (handled by North8000 circa 7/1/22) |
Gondolabúrguer | 76 | Subdivisions in Brazil (done by North8000 circa 6/26) |
Das osmnezz | 73 | Footballers |
SonOfBasra | 73 | Iraqi football teams |
Montenois | 67 | Subdivisions of the Russian Empire (done by North8000 circa 6/26) |
Ss112 | 65 | Music albums |
CatchedY | 61 | Biographies in various fields |
Ainty Painty | 59 | Biographies and other articles in the Pakistan/Afghanistan area |
Helen Puffer Thwait | 57 | Music albums |
Immanuelle | 57 | Chinese and Japanese articles |
Thriley | 56 | Varied biographies |
Abbasulu | 55 | Bengali and Bangladeshi films and film biographies |
Rollcloud | 53 | Rare genetic disorders (done by North8000 circa 6/26) |
Pirhayati | 47 | Varied biographies |
JIP | 44 | Finnish articles |
AshMusique | 43 | Rap songs |
Pehlivanmeydani | 42 | Sport biographies and events from the 2022 Mediterranean Games |
Dwanyewest | 42 | Varied |
Haoreima | 41 | Meitei mythology from Manipur, India |
Total | 1178 |
If anyone already finds any of these topics easy/fast to review, please do so. -MPGuy2824 (talk) 06:50, 26 June 2022 (UTC)
I’d be interested in a third column showing % of articles produced by the person in the last 90 days that are unreviewed (that is, number of articles in the queue divided by number of articles written in past 90 days). I suspect some of these article types are being passed over by reviewers. For example, I’ve been bypassing all those municipal corporation articles. — rsjaffe 🗣️ 11:57, 26 June 2022 (UTC)
I also took a look at those "municipal corporations in India" articles. I also stepped back, not knowing what to do with them. IMO each would best make a paragraph or section in the article for the city that they are part of. But there is a lot of work there. Merging 100+ articles, including separating out the boilerplate material that is duplicated across them would be a huge job. As would separating out the "micro-happenings" material that makes up the bulk of the unique material and sources in each. I did leave the creator this note:
- "First, most importantly, thanks for your work! Next, during new page patrol work I came upon a larger amount of articles on departments/authorities within cities These articles seem to consist of mostly general info regarding the type of department/authority (copied between many many articles) and then a smaller amount of content that is unique to the subject. Also much of the latter is about individual related events rather than being directly about the subject as a whole. If you are still making more of these, I'd like you to consider make them instead a paragraph or section within the article of the city which they are a part of. Thanks again for your work."
Sincerely, North8000 (talk) 21:22, 26 June 2022 (UTC)
- Since the quarry results will only update if I re-run it manually (or someone else forks it and re-runs their version manually, etc.) I've asked if it can be turned into a regular database report at Wikipedia_talk:Database_reports#Request:_editors_by_number_of_unreviewed_pages. – Joe (talk) 10:05, 27 June 2022 (UTC)
- Since the backlog drive started and we don't have the automated report yet, this is the latest top 10:
User | Count | Commonality | Picked up by |
---|---|---|---|
SonOfBasra | 78 | Iraqi football teams | |
Das osmnezz | 72 | Footballers | |
Ss112 | 65 | Music albums | |
Ainty Painty | 58 | Biographies and other articles in the Pakistan/Afghanistan area | |
Abbasulu | 51 | Bengali and Bangladeshi films and film biographies | |
Dwanyewest | 43 | Varied | |
AshMusique | 42 | Rap songs | |
JIP | 39 | Finnish articles | |
OrangTangerang53 | 39 | Indonesian football clubs | |
Aidan721 | 33 | baseball under-18 teams of countries |
-MPGuy2824 (talk) 02:33, 2 July 2022 (UTC)
- NB. Ss112 has been autopatrolled since 2016, those 65 are actually redirects he created that other users have then turned into articles. – Joe (talk) 12:39, 4 July 2022 (UTC)
Unreviewed article tag
It appears that the NOINDEX until reviewed is going to happen. This is likely to cause more people to ask "why doesn't my article appear in Google?" To mitigate the extra work of answering this question, we could use a tag like this on all unreviewed articles. I'm not sure of the mechanism to add it automatically and remove it automatically, but before we worry about that, is this a good idea? 14,000 tags is a lot, but it is only 0.22% of all articles. There are that many biography notability tags alone, so it wouldn't without precedent. Thoughts? MB 06:04, 26 June 2022 (UTC)
- I like it. It forces people who are interested in getting articles indexed soon, to take notability and other tags seriously. 14,000 of these tags is a lot, but we are thinking of such ideas, with the end goal being a much smaller backlog. Mechanism of adding/removal: Some sort of hook when an article is marked as reviewed, possibly. A bot which runs hourly or so, is also good enough. -MPGuy2824 (talk) 06:18, 26 June 2022 (UTC)
- It might be useful to clarify what exactly "Wikipedia mainspace" is. I know what it is, as do a lot of other active editors, but a newbie might not. Clovermoss (talk) 12:03, 26 June 2022 (UTC)
- I like. Maybe to address User:Clovermoss' point, "While it has now been added to Wikipedia, it will not be visible to..." Best Alexandermcnabb (talk) 14:43, 26 June 2022 (UTC)
- While we're at it, it's probably useful to clarify that this process is different from AfC. Maybe in a linked FAQs section or something. Someone might try to remove the template in good faith because they think it's already been reviewed if it's an accepted AfC draft. Clovermoss (talk) 16:04, 26 June 2022 (UTC)
- Another thing to keep in mind would be about the mainspace pages that go through NPP that aren't actually articles like disambiguation pages and redirects. I'm assuming a template like this automatically being added to the page would render the redirect useless, at least temporarily. Clovermoss (talk) 16:26, 26 June 2022 (UTC)
- I like. Maybe to address User:Clovermoss' point, "While it has now been added to Wikipedia, it will not be visible to..." Best Alexandermcnabb (talk) 14:43, 26 June 2022 (UTC)
- It might be useful to clarify what exactly "Wikipedia mainspace" is. I know what it is, as do a lot of other active editors, but a newbie might not. Clovermoss (talk) 12:03, 26 June 2022 (UTC)
I like it also for a second reason which is that it's sort of invisible whether or not an article has been patrolled (or =), But we need to make sure that we don't create another 800 new jobs per day for NPP'ers which is removal of the tags when the article is reviewed. Either need a bot to do that quickly (like maybe every 15 minutes) or add it to what the curation tool automatically does. North8000 (talk) 15:18, 26 June 2022 (UTC)
This does seem like a decent idea. I second the idea of a bot, and page curation can also remove the tag as well (no reason we can't do both). One drawback I can see is that it adds clutter, but at the same time could potentially draw more attention to NPP and possibly result in more members. I would also suggest adding a link to "mainspace" as the concept of namespaces may not always be understood by new users. I can also see a lot of page creators just removing the tag (as is a typical practice for some) which might create work for editors who will then have to revert and explain to others why they shouldn't remove the tag. ASUKITE 16:06, 26 June 2022 (UTC)
Version 0.1
This version removes mention of "mainspace" as suggested above, and includes a link to a page that can elaborate such things as
- Removal is automatic after review; do not manually remove as (hopefully) a bot will just put it back
- Differences with AFC review
- Maybe a link that displays the date of the oldest unreviewed article in the "new part" of the queue to give an idea of how long it could take
- A note that we don't really care if an article is indexed and if you do, that may be a sign that you are trying to promote something (this may be going too far, some editors apparently want articles to be indexed because they prefer to search for articles with Google instead of the WP search bar).
- A link to WP:HELPDESK if there are still questions
As far as use on Redirects, it would not break a redirect if placed after the first line containing #Redirect
, or there could be special coding the way there is on a redirect listed at RFD. But it's probably best to just not use it at all on redirects. The text is geared towards articles, and I don't think the issues are the same.
DABs in the queue usually don't stay there for long, and a DAB could be considered a special kind of article. I don't think we need to do anything special. MB 21:27, 26 June 2022 (UTC)
- @MB: Is there a way for it to exclude redirects? Like if a page has an rcat or a disambiguation template? Otherwise I think version 0.1 addresses all my concerns. Clovermoss (talk) 22:41, 26 June 2022 (UTC)
- Probably should have some process to handle unwanted removal of the template. People sometimes remove maintenance templates without fixing the issue, so I could see that happening here. Perhaps a bot? — rsjaffe 🗣️ 22:51, 26 June 2022 (UTC)
- To be honest, after reading all of the discussion above I don't really see a need for this banner. "Why doesn't my article appear in Google?" is a pretty easy question to answer and I don't really expect these questions to be an exacerbating problem if the 90-day period is extended. Unlike other maintenance templates, which alert readers that someone has identified a concrete issue and invites them to fix it if they can, this one seems mostly informational and designed for the page creator, which means that it becomes the most prominent thing on the article for a whole lot of people who don't really need to see it and probably can't do anything about it. Finally, there are plenty of clean, notable, and policy-compliant articles in the NPP queue that simply haven't been patrolled because the backlog is so large and no one has gotten around to it. I don't like the idea of a bot repeatedly reverting anyone (especially an experienced editor) who tries to remove the tag on such an article. DanCherek (talk) 23:25, 26 June 2022 (UTC)
- I think the idea of the banner is excellent. Kudos to MB for thinking outside the box and coming up with it. It's one of those no-brainer things that can be done without the bureaucracy of having an RfC for it (which has sadly become a new trend even to dot an I or cross a T). It should be easy for it to appear on all new articles except redirects. Due to the vast increase in Internet access and cheap smart phones in non-English L1 regions, its language should be short, free of WP jargon, and be ENGVAR agnostic. IMO This is all it needs for the required knee-jerk:
- Of course, the best solution would be to get the WMF to finally create a proper landing page for newly registered users which they have avoided doing for 20 years. Kudpung กุดผึ้ง (talk) 13:27, 3 July 2022 (UTC)
How other Wikis do it
Just noticed that the German Wikipedia has a slick way of doing this. First, they have a page explaining it [1]. Then, unreviewed articles have a small tab at the upper right of the article, such as this one: [2]. I like the unobtrusiveness of this method and the way that it says that quality control hasn't been done on the article.— rsjaffe 🗣️ 16:08, 10 July 2022 (UTC)
Phabricator vs NPP software requests
An important request has been flatly denied. The reason given on a linked page, is that Page Curation is a 'Non maintained project':
Projects that Growth team has no capacity to maintain and are looking for new ownership or have transferred ownership to another team and are listed here for historical reference. These projects don't have any dedicated resources behind them but are available for new owners who see a future in them and would like to take up the work of developing and maintaining them. Bugs and issues coming up on this projects will ONLY receive attention if the issue is urgent, affects a large number of users and needs to be unbroken immediately.
It is my opinion, though it might not be shared by everyone, is that NPP is the only gatekeeper against possibly inapropriate new articles, and as such is a crucial process for the quality of Wikipedia. Hence the New Pages Feed and its Curation tools need funds and a developer team allocated to their maintenance and improvement. Funds are certainly not lacking. Kudpung กุดผึ้ง (talk) 01:52, 28 June 2022 (UTC)
- Agree 100%, Kudz!! It's a mess. There are bugs to be fixed, features & stats we need, & areas that need streamlining/automation, etc. How much would the average person tolerate were it a paying job, never mind doing it as a volunteer? After reading the this Signpost article, and WMF's laissez faire attitude about ill-prepared articles, PE, unsourced articles – let the volunteers fix it – and the lack of WMF support for the tools we need, it almost seems like they're purposely trying to rid the project of editors. Get back in the coal mine, and don't worry about the dead canaries![stretch] And while I don't begrudge the project earning money off the backs of our free labor, I do mind when they deny us the tools we need to provide the free work. To call it disheartening is an understatement. Atsme 💬 📧 03:27, 28 June 2022 (UTC)
- Following the message I ledt at Phab, a dev has now claimed this particular task. Anyway, it looks as if some progress is being made. What is really needed however, is a dev team being drawn up to specifically rewrite the entire NPP code, which the WMF told me a few years ago (Ithink it was Kaldari) was unfortunately not compatible with later iterations of MediaWiki and therefore not Wiki-agnostic. Now is the time to do it. There is plenty of money - the WMF is wallowing in it. Kudpung กุดผึ้ง (talk) 10:20, 28 June 2022 (UTC)
- enwiki and all WMF wikis run the latest version of mediawiki (a development version even more up-to-date than the official releases), because they use continuous deployment and release every week (WP:ITSTHURSDAY). I would hesitate to request a complete re-write of complex, working software, as the time to value ratio on this (or any rewrite of working software) would likely be low (ton of developer time, little new useful functionality). Rewrite (programming)#Risks. –Novem Linguae (talk) 20:07, 28 June 2022 (UTC)
- I'm aware of all this. I was told years ago when its original development had just been rolled out. It was explained again years later at here, along with a super log of all the requests that have been addressed, but they have discretely admitted that the whole thing should really be recoded. We worked long and hard on the design of the user interfaces in the best interests of the reviewers in order to try and make NPP a smoother and more appealing task, and later to embed some ORES functions in it but there is still some way to go. If all the features were to be included that we would still find extremely useful a complete rewrite might probably be the best way to get there. Kudpung กุดผึ้ง (talk) 20:40, 28 June 2022 (UTC)
- IMHO
has been flatly denied
is not an accurate characterization. Patches were welcome (and actually a volunteer sent a patch), and the responsible team said thatit's not something that the Growth team can work on any time soon
. Software development in big projects involves managing priorities, and managing finite time. Of course, we can argue for prioritization, and make a case for a higher priority, but I think that triggering hostility against developers just because a ticket is not initially prioritized is just wrong. MarioGom (talk) 20:41, 28 June 2022 (UTC)
- Well, @Kudpung, @Wbm1058, something's different that happened quite recently. I just tried to create a user talk subpage for a NPP trainee exercise and got the following message: The page "User talk:Atsme/NPP training/[student name]" does not exist. You can request that it be created, but consider checking the search results below to see whether the topic is already covered. Not knowing what else to do, I bypassed it via a move leaving no redirect, which may be a better option anyway, but...I'd still like to know what changed. Atsme 💬 📧 13:05, 30 June 2022 (UTC)
- @Atsme: Could be that [student name] contains something on the title blacklist? Or you were somehow logged out when you tried to create it? – Joe (talk) 13:45, 30 June 2022 (UTC)
- Good morning, Joe Roe! I only substituted [student name] for this discussion in lieu of the actual user name. I was logged in, and the move went through without issue. ¯\_(ツ)_/¯ I just now tried to create another subpage using "Tank", and got the same message: The page "User talk:Atsme/NPP training/Tank" does not exist. You can request that it be created, but consider checking the search results below to see whether the topic is already covered. ???? Something has been changed. Atsme 💬 📧 14:01, 30 June 2022 (UTC)
- Atsme, when you see this message on the "Search results" page:
- The page "User talk:Atsme/NPP training/Tank" does not exist. You can request that it be created, but consider checking the search results below to see whether the topic is already covered.
- Click on the red link User talk:Atsme/NPP training/Tank and then click the tab "Create". You don't have to "request that it be created". – wbm1058 (talk) 14:29, 30 June 2022 (UTC)
- Oh, I see it now - so they changed the appearance - wow. It's a big enough change that I didn't notice I was able to create the page after clicking the red link. The first thing I saw was "Create a redirect from this page with Sagittarius" and "Start a discussion". Apologies for the side railing. Atsme 💬 📧 14:57, 30 June 2022 (UTC)
- Atsme, when you see this message on the "Search results" page:
- Good morning, Joe Roe! I only substituted [student name] for this discussion in lieu of the actual user name. I was logged in, and the move went through without issue. ¯\_(ツ)_/¯ I just now tried to create another subpage using "Tank", and got the same message: The page "User talk:Atsme/NPP training/Tank" does not exist. You can request that it be created, but consider checking the search results below to see whether the topic is already covered. ???? Something has been changed. Atsme 💬 📧 14:01, 30 June 2022 (UTC)
- @Atsme: Could be that [student name] contains something on the title blacklist? Or you were somehow logged out when you tried to create it? – Joe (talk) 13:45, 30 June 2022 (UTC)
- If we can’t have the tools to patrol easily, I will just not bother. If I knew how to, I would make a tool myself. I’m just going to hope that the WMF develop the tools when NPP falls to its knees. | Zippybonzo | Talk | 14:07, 30 June 2022 (UTC)
- Every bit of help with patrolling would be much appreciated Zippybonzo. NPP is arguably the most important non-admin task on Wikipedia, and although the WMF has done a lot for it in very recent years at our requests, for several reasons they are reluctant to accord any priority to it. Nevertheless, the software we have today is pure luxury. You're very new here, some of us have been waiting a decade or longer for a perfect set of tools! This is what we used to have to work from, and there were no curation tools, tutorials or flowcharts. What we need right now are a lot more truly active patrolers - only about 10% of the 700+ are actually doing anything. Kudpung กุดผึ้ง (talk) 06:48, 3 July 2022 (UTC)
Is it ethical to patrol your own pages?
Obviously, I'm not going to do it, but for the sake of curiosity as a newly appointed new page reviewer: Is it ethical to patrol your own articles/redirects/other pages? It is technically possible, which could make the new page reviewer right autopatrolled in practice. I'm going to assume no, but I just wanted some clarification.
Also, I'm asking this partially because I reverted some vandalism at А and Α, (the Greek and Cyrillic letters respectively) and the redirect I restored is still unpatrolled (I won't patrol it myself).
Please let me know. Thanks to all. — 3PPYB6 — TALK — CONTRIBS — 00:27, 29 June 2022 (UTC)
- The software doesn't allow an NPP to patrol their own created pages. You would need to be autopatrolled to get around this. In the eyes of the software, editing an existing page is different than creating a page. As for the ethics of reverting redirect vandalism and then marking that patrolled, sounds ethical to me. Hope that helps. –Novem Linguae (talk) 00:41, 29 June 2022 (UTC)
- Agree with Novem when its vandalism. If it's a potential good-faith content dispute, I've left my restore un-patrolled so someone else can put eyes on it. Regardless, I have patrolled the two pages to get then out of the queue. Slywriter (talk) 00:48, 29 June 2022 (UTC)
- So it's not possible to patrol any page you have created. Chnaging from redirect to article requires a new patrol, but you are not considered the creator. I usually patrol my own when reverting vandalism to an already patrolled page, but NOT when I BLAR or remove a redirect. Happy Editing--IAmChaos 02:07, 29 June 2022 (UTC)
- Suggestion: read Wikipedia:Autopatrolled. It's a user right that is earned, and if abused, will quickly be removed, and the abusing editor will be held accountable. Atsme 💬 📧 12:17, 29 June 2022 (UTC)
- I'm not so sure about it being 'earned'. It's not an award of merit and it has absolutely no benefits for the user, unless the are an abuser. IMO the right should be scrapped. Kudpung กุดผึ้ง (talk) 15:38, 30 June 2022 (UTC)
- I'm not so sure about the second part either, sadly. We have no process for systematically reviewing autopatrolled after it's granted and it's rarely revoked. I'd also like to see it scrapped or significantly rolled back, but probably now is not the best time. – Joe (talk) 15:55, 30 June 2022 (UTC)
- I think autopatrolled is a bad idea. Everybody needs a second set of eyes to look at it, and it's also a safety net for when the wrong person gets it or when a well-intentioned person goes awry. The benefit (less for us to patrol) is numerically small.North8000 (talk) 16:02, 30 June 2022 (UTC)
- I strongly disagree. Earn means earn trust, and right means you have the right to use it but not abuse it as a trusted editor. It would not only be an open door to UPE/PE, it would eliminate doors. If we do eliminate it based on the reasons given here, we might as well eliminate administrator rights which are also based on community trust. Orange Moody, and a few other names come to mind. Give every editor the tools after they’ve had 500 edits. Atsme 💬 📧 16:28, 30 June 2022 (UTC)
- I think autopatrolled is a bad idea. Everybody needs a second set of eyes to look at it, and it's also a safety net for when the wrong person gets it or when a well-intentioned person goes awry. The benefit (less for us to patrol) is numerically small.North8000 (talk) 16:02, 30 June 2022 (UTC)
- Now is not the time to propose something that would greatly enlarge the NPP queue. We need to stay laser focused on reducing the queue. Let's stay focused, folks. –Novem Linguae (talk) 20:13, 30 June 2022 (UTC)
- Novem Linguae This is when an editor like CommanderWaterford would come in handy, huh? Atsme 💬 📧 18:09, 1 July 2022 (UTC)
- @Novem Linguae, Joe Roe, and North8000: there is no mention of anything new being 'proposed' here. On this page however (most) users are seriously concerned by 1). the continued stream of junk and other inappropriate new article submissions which has again reached pre-ACTRIAL proportions, and 2). our system (NPP) for controlling it.
- I'm not so sure about the second part either, sadly. We have no process for systematically reviewing autopatrolled after it's granted and it's rarely revoked. I'd also like to see it scrapped or significantly rolled back, but probably now is not the best time. – Joe (talk) 15:55, 30 June 2022 (UTC)
- I'm not so sure about it being 'earned'. It's not an award of merit and it has absolutely no benefits for the user, unless the are an abuser. IMO the right should be scrapped. Kudpung กุดผึ้ง (talk) 15:38, 30 June 2022 (UTC)
- The mention here is therefore an important wakeup call and no stats have been provided as to how disbanding autopatrolled would negatively impact the work at the New Pages Feed. I would assume, conservatively, that deprecating this user 'right', which has nothing whatsoever to do with being 'earned', would bring more benefits to maintaining a clean encyclopedia corpus than disadvantages to the role of the New Page Reviewers.
- The increase in the discovery of abuse by holders of the autopatrolled and NPP rights is worrying. Removal of the right is not swift and there is perhaps a sign that there still exists a huge grey area here - detection is still not as frequent as it could be despite the massive 2015 OrangeMoody affair, and noises made in The Signpost in 2018. There is nothing to stop an incentive to thoroughly investigate 'autopatroled' once and for all, and it could possibly lead to an indication that the fate of 'autopatrolled' might well be slated for debate in the not too distant future. Kudpung กุดผึ้ง (talk) 23:46, 30 June 2022 (UTC)
- We should keep our focus on reducing the NPP queue, not enlarging it. We don't need to manually patrol thousands of species articles per year, for example. Where is the hard evidence that autopatrol is currently a net negative? Has there been a large quantity of autopatrol scandals lately that cannot simply be handled by putting the discovered UPE's articles back in the queue? –Novem Linguae (talk) 00:16, 1 July 2022 (UTC)
- @Novem Linguae: I think I more than adequately explained the situation in my rather long post above. There is no hard evidence, that's why a call for stats should be made before starting any debates on it:
There is nothing to stop an incentive to thoroughly investigate 'autopatroled' once and for all
. However, IMO 'autopatrolled' is too frequently abused. When I choose to select autopatrolled articles in the feed prefs, I almost always come across some infringement of the right. I fully understand your point about species articles, but as these articles generally come from mature, serious editors and scientists, plenty of whom are not autopatrolled (so I do know what they look like). The articles are therefore generally clean and are frequently very short. They can be patrolled very quickly. But as I also said, 'autopatrolled' should ultimately be the basis of another debate. Kudpung กุดผึ้ง (talk) 01:34, 1 July 2022 (UTC)
- @Novem Linguae: I think I more than adequately explained the situation in my rather long post above. There is no hard evidence, that's why a call for stats should be made before starting any debates on it:
- We should keep our focus on reducing the NPP queue, not enlarging it. We don't need to manually patrol thousands of species articles per year, for example. Where is the hard evidence that autopatrol is currently a net negative? Has there been a large quantity of autopatrol scandals lately that cannot simply be handled by putting the discovered UPE's articles back in the queue? –Novem Linguae (talk) 00:16, 1 July 2022 (UTC)
I gave my opinion that it shouldn't exist but IMO it's not a pressing issue or high on my worry list. North8000 (talk) 02:15, 1 July 2022 (UTC)
A group of about 110 articles on entities in India by one editor that I think will need to get deleted
Further to my "100+ articles on departments/authorities within cities in India by Gardenkur awaiting review" post above and linked discussion on my talk page, I passed about 10 of the editor's open articles and I think about 110 need to get deleted. The edior has agreed to stop making these types of articles although I'm not sure that they fully understand.
If anyone is interested, I just took the first one as a test case to AFD which is at Wikipedia:Articles for deletion/Burhanpur Municipal Corporation IMO the "Municipal Corporation" bundle is the most obvious one for not existing as a seperate article. The trickier part is avoiding any outcome which would require merging material which, for the reasons I detailed there, would never get accomplished and really has no material worth merging. North8000 (talk) 17:20, 30 June 2022 (UTC)
- I worked it out with the editor and agreed to be a sort of pseudo-mentor on such things.North8000 (talk) 13:01, 2 July 2022 (UTC)
NPP July 2022 backlog drive is on!
New Page Patrol | July 2022 Backlog Drive | |
| |
You're receiving this message because you are a new page patroller. To opt-out of future mailings, please remove yourself here. |
(t · c) buidhe 20:26, 1 July 2022 (UTC)
Backlog chart
Is it just me or is the hourly chart not displaying properly, and does anyone more tech-y then I am know how to fix it? Courtesy link: [[:{{{1}}}]]sandbox Happy Editing--IAmChaos 02:42, 2 July 2022 (UTC)
- It looks fine to me. What is it you see? – Joe (talk) 06:56, 3 July 2022 (UTC)
- External link: Imgur. It is collapsed against the axis (only with the hourly parameter, the rest are fine). Happy Editing--IAmChaos 17:20, 3 July 2022 (UTC)
- It looks fine to me on my computer (Chrome) but I do see the error when I browse it in safari on my IPhone. — Insertcleverphrasehere(or here)(or here)(or here) 20:57, 5 July 2022 (UTC)
- External link: Imgur. It is collapsed against the axis (only with the hourly parameter, the rest are fine). Happy Editing--IAmChaos 17:20, 3 July 2022 (UTC)
Sigcov in songs articles
Hello, it's been a while since I was an NPP regular. (8 years, maybe?) I looked over the tutorial and am pretty confident that I still understand the basic workflow but will probably occasionally solicit feedback on edge cases.
Speaking of which, I wanted to ask about Skinny Dipping (Sabrina Carpenter song). I did a WP:BEFORE search and concluded that the song passes WP:GNG, albeit barely. However, the sources I found that I believe constitute significant coverage devote most of their attention to describing what happens in the song or music video. (The equivalent of plot summary, basically.) I'm not sure whether others would agree that is actually sigcov and thought I should ask. —Compassionate727 (T·C) 15:14, 2 July 2022 (UTC)
- I spot checked the Teen Vogue source. Seems pretty borderline. Tough call. For sigcov I usually look for 3 meaty paragraphs about the topic in question, that aren't quotations, aren't talking about a slightly different topic, and aren't based off a press release or interview. This source has about 3 paragraphs at the beginning talking about this song, and they are based off an interview. The caveat for interview-based articles (that have some prose and aren't just a giant quote) I've noticed is that if they're short, they tend not to be independent enough, but if they are long, they overcome this. Tough call. I'll let others weigh in. –Novem Linguae (talk) 18:26, 2 July 2022 (UTC)
- One could argue this is Wikipedian or not, but I try to go by what would probably happen at an AFD which had neutral participants. The fact that the artist is triply slam dunk wp:notable isn't supposed to matter but it does, if only as an indicator of likelyhood of more suitable sources. For the sources that are already in the article I'd call it an edge case. Even a quick search found other suitable-looking sources. My take would be "pass" along with the current tagging.North8000 (talk) 20:03, 2 July 2022 (UTC)
Question from new-NPP re: AfD curation tool bug
Hello! I'm a new-NPP, and have encountered a bug with the Curation Tool not completing the AfD process. I was able to resolve this today by using Twinkle (and removing the Curation Tool notice). Am I doing something wrong or is the tool buggy? I've never had problems with Twinkle in the past, so I'm fine bypassing the CT for nominations. Feedback or advise would be appreciated, esp. if I am using the tool incorrectly. Thanks in advance! Netherzone (talk) 17:57, 3 July 2022 (UTC)
- Believe consensus is Twinkle is better for everything and NPP tool is mostly for wikilove and review checkmark Slywriter (talk) 18:19, 3 July 2022 (UTC)
- +1 –Novem Linguae (talk) 20:06, 3 July 2022 (UTC)
- Yes (plus of course, also the page feed), it took me a while to learn that. We should write this up somewhere. North8000 (talk) 12:50, 4 July 2022 (UTC)
"Twinkle is better for everything and NPP tool is mostly for wikilove and review checkmark"
is IMO a rather sweeping statement. In 10 years of using it and largely coordinating NPP I have found no evidence that such a consensus exists and we have always been mindfull of its improvement and maintenance. The Curation Tool was designed and developed in full collaboration between the NPP volunteers and senior WMF developers including many face-to-face meetings and it was specifically modeled on the features of Twinkle. That said, no one disputes that the learning curve for correctly reviewing pages is quite steep and therefore a detailed tutorial exists which is one of the most comprehensive but easily digestible user rights tutorials on Wikipedia.
- In order to reduce subpar tagging and patrolling of articles through Twinkle, in 2016 a user right for NPP was created. There are nevertheless clear caveats that to be done correctly NPP requires an almost admin level knowledge of notability and deletion and that ' New Page Patrol is not a task for new or inexperienced users.' No other system provides as much information on each article both in the feed and in the tool to help reviewers decide what to do with an article; its function is fully described and illustrated here even including a video tutorial.
- Bugs: The system is more stable today than it has ever been, and new features are being regularly requested. If there are technical bugs anywhere in the system, please report them here with as much detail as possible. A coordinator will follow up. Alternatively, if you are familiar with Phabricator you are welcome to open a ticket directly there, but please notify this talk page with the Phab tracking info so that it can be followed up. @Netherzone, Slywriter, and Novem Linguae:. Kudpung กุดผึ้ง (talk) 15:55, 4 July 2022 (UTC)
- Just for the record, here's a previous discussion about the same issue with AFD in the curation toolbar: Wikipedia talk:New pages patrol/Reviewers/Archive 42 § AFD message MIA.Personally, I don't see any value in having the toolbar doing AFD nominations, since many (most?) of us are more familiar with Twinkle, and it generally works better. MarioGom (talk) 19:02, 4 July 2022 (UTC)
Many thanks for these replies, they are helpful! Netherzone (talk) 22:59, 4 July 2022 (UTC)
- The problem is not with the bugs in the tools themselves, but with the lack of reporting them properly. There's no point in complaining and claiming Twinkle is better if the developers are not told that something needs fixing. That's one of the reasons why NPP needs a coordinator. Things got done when I, then Insertcleverphrasehere, and then Barkeep49 were looking after things. The NPP community only has itself to blame. Perhaps the admins who are regular reviewers such as @Rosguill, Joe Roe, Ymblanter, and Graeme Bartlett:, might like to comment, including MusikAnimal/MusikAnimal (WMF) who has an overview from all sides. AfC should be the least of our worries, it has a backlog of less than 3,000, a highly competent and active defacto coordinator and a group of over 600 active users that gets regularly pruned for inactivity. FYI: MarioGom. Kudpung กุดผึ้ง (talk) 02:32, 5 July 2022 (UTC)
- We'd need at least one example of failed AFD creation. Netherzone, do you have any example of an AFD that failed for you? Unfortunately, I can't find the one that failed for me some time ago. MarioGom (talk) 06:14, 5 July 2022 (UTC)
- I tried to use the curation tool yesterday to nominate Swift Horses Sharp Swords: Medieval Battles which shook India for deletion (see these edits). It was able to tag the article, add the nomination to the daily log, and notify the user, but failed to actually create the nomination page (even after I had entered and saved my nomination rationale in the toolbar). I had to quickly re-type my nomination and manually create the page afterwards (Special:Diff/1096459998). It's the same behavior that was reported in tasks like phab:T238025 and phab:T239712 (h/t links taken from Joe's message below). DanCherek (talk) 14:04, 5 July 2022 (UTC)
- MarioGom, here is an example of an AfD using Page Curation: the tool did everything except create the actual AfD Discussion page.[3] So I removed the Page Curation deletion tag,[4], and AfD'd it using Twinkle instead[5] which completed each step the process in full. I then noticed that the article creator received two deletion notifications on their talk page (one from Page Curation, and one from Twinkle) so I deleted the one from Page Curation because it did not have a working link to a discussion page.)[6]. This happened to me with three AfD's the first one another editor came along and fixed things, and the other two I cleaned up myself with the process described above. I hope that helps explain the difference between how the two tools functioned differently. Netherzone (talk) 15:56, 5 July 2022 (UTC)
- We'd need at least one example of failed AFD creation. Netherzone, do you have any example of an AFD that failed for you? Unfortunately, I can't find the one that failed for me some time ago. MarioGom (talk) 06:14, 5 July 2022 (UTC)
- DanCherek, all the more reason for you to go to Phab and give them a knee-jerk. Kudpung กุดผึ้ง (talk) 15:08, 5 July 2022 (UTC)
- Also pinging @SWilson (WMF), Samwilson, and NKohli (WMF):. Kudpung กุดผึ้ง (talk) 02:50, 5 July 2022 (UTC)
- Thank you for this back-story, @Kudpung, it's always interesting to me to learn about WP history and I look forward to learning more about NPP looking forward and also understanding the past. Netherzone (talk) 02:43, 5 July 2022 (UTC)
- I never use twinkle, so I cannot compare. I find the tools are over 90% reliable. The last problem I had was that AFD was nominated twice in a second. But simple to overcome that. I appreciate the AFD feature, as it saves a complex manual process. But I only patrol one or a few per day, and mostly they just need the tick of patrolled. Graeme Bartlett (talk) 05:22, 5 July 2022 (UTC)
- I only use Page Curation for the green tick and the next button. I have done for years and I also thought it was the common practice. There are numerous outstanding phabricator tickets relating to AfD and Page Curation, the oldest of which was opened 10 years ago (phab:T239714, phab:T239712, phab:T238328, phab:T238025, phab:T44650). And bugs aside, last I tried it (which was admittedly a long time ago), the UI was far less polished than Twinkle. I think the initial design of Page Curation was good but it really should have been implemented as local Javascript gadget for enwiki, not a Mediawiki extension. Maybe another reason that AfC is doing comparatively well is that their tool is a gadget and it's actively maintained by local volunteers, not reliant on the WMF/MW developers for every fix or update. – Joe (talk) 08:49, 5 July 2022 (UTC)
- Joe, back in 2010-2012 NPP was already in crisis, no one here was going to offer to build a new system to replace the Special:NewPages to allow faster and smoother patrolling, and despite the huge turnout and massive consensus for WP:ACTRIAL, the WMF flat out (and with extreme incivility) refused to allow it to be implemented. As a consolation, they offered to build a new feed and the Curation Tool was part and parcel of it. Who were we to refuse such a 'magnanimous' offer? Years later after a total turnover of WMF staff, new devs admitted that building it as a MediaWiki extension was an error. Nevertheless, I had collaborated on the design of the GUI and the choice of functions, and although I say it myself, I find the system we have today a vast improvement on the old one - just like anyone who uses a Mac would never want to go back to Windoze.
- I don't deny that you might be right about Javascript and Github, but I firmly believe that the way to go is to address the bugs and improve it. If enough of the right kind of pressure were to be exerted on the right WMF department heads (are there any?), I'm sure it would get done. After all, the en.Wiki is their flagship project and it would be in their interest to have a clean encyclopedia corpus. The old chestnut 'not enough funds' is just plain ludicrous as we all know.
- AfC does well because they have over 600 active reviewers, a tiny backlog, no deadlines, and a coord who is highly experienced and constantly on the ball. The work is also less soul destroying and mind numbing soul destroying. The daily flood content of the New Pages Feed is basically a torrent of effluent. Kudpung กุดผึ้ง (talk) 13:49, 5 July 2022 (UTC)
Wikilove
If I were to criticise myself for my NPP performance (others will no doubt add more if I'm going too easy on myself!), it would be that I don't do enough 'Wikilove'. But the toolbar only has an option to send a message to the creator if you tag a page (which is generally not the best page to send love for!) rather than have a drop down under 'Reviewed' that allows you to direct a message to the talk page, the originator or another contributing editor (for instance, where a redirect has been later converted into an article). Noting User:Kudpung's historical involvement in the toolbar from the above discussion and their familiarity with this arcane Phabricator tickety thing, would others agree that might be a useful feature or am I, as usual, just missing something huge and basic staring me in the face? Best Alexandermcnabb (talk) 06:41, 5 July 2022 (UTC)
- Alexandermcnabb, when you click the "mark as reviewed" button on the toolbar, don't you see a box that says "Add a message for the creator. This message will also be posted to the article talk page." (similar to the box in this picture)? Extraordinary Writ (talk) 06:49, 5 July 2022 (UTC)
- Extraordinary Writ Yes, but that's not sending a message to the editor's talk page. It only allows that if you tag the article. And that's my point - a message, a chococlate biscuit even, to the user is more direct, personal and - arguably - motivating??? Best Alexandermcnabb (talk) 06:55, 5 July 2022 (UTC)
- It does send a message to the editor's talk page (example). The message is "also" posted on the article talk page, but if you test it out you'll see that the message is indeed delivered to the editor's talk page. Cheers, Extraordinary Writ (talk) 06:59, 5 July 2022 (UTC)
- See? "something huge and basic staring me in the face" as usual... Thanks! :) Best Alexandermcnabb (talk) 07:03, 5 July 2022 (UTC)
- It does send a message to the editor's talk page (example). The message is "also" posted on the article talk page, but if you test it out you'll see that the message is indeed delivered to the editor's talk page. Cheers, Extraordinary Writ (talk) 06:59, 5 July 2022 (UTC)
- Extraordinary Writ Yes, but that's not sending a message to the editor's talk page. It only allows that if you tag the article. And that's my point - a message, a chococlate biscuit even, to the user is more direct, personal and - arguably - motivating??? Best Alexandermcnabb (talk) 06:55, 5 July 2022 (UTC)
- Wikilove and messages are too loud for my taste and so I save them for cases where I have something specific to say. As a simple courtesy, I will usually thank a creator for starting an article. This seems to require inspection of the article history but that can be useful in confirming who has contributed and how. After I repeatedly thanked such a creator, they recently replied, "I thank you sincerely for your kind reviews on all of my article creations. Those are quite rare on this project. I thank you very much for showing me that kindness!"
- Andrew🐉(talk) 10:07, 5 July 2022 (UTC)
- Now that is a nice idea! Thanks! Best Alexandermcnabb (talk) 10:13, 5 July 2022 (UTC)
- At the time, I didn't believe incorporating WikiLove into the Curation Tool was a particularly good idea. And I still don't. However, from the comments I've seen over the years, I'm sure the feed, its filter options, and the Curation Tool are underused. I'm convinced that newer reviewers still haven't fully explored and exploited all the features of the Curation Tool. If they did, and read the tutorial properly, there would be far less criticism of it and fewer claims that Twinkle is so much better. Those who requested the user right would then be more willing to use it rather than be another of the many inactive users among the 700+ reviewers. Kudpung กุดผึ้ง (talk) 14:53, 5 July 2022 (UTC)
- Ah, come on, now. Twinkle is pretty sexy... not that I would want to argue with Your Grace... Best Alexandermcnabb (talk) 15:04, 5 July 2022 (UTC)
Sidebar AfD issue
I've been having issues nominating articles for AfD using the NPP sidebar tool the past couple of days. This type of malformed nomination has happened twice to me (PROD and CSD work fine, though; it is just AfD that has the issue). Is this a problem for anyone else, or am I just peanut-brained? Curbon7 (talk) 21:54, 5 July 2022 (UTC)
- Nevermind I just read the thread above, looks like I'm not the only one with this issue. Curbon7 (talk) 21:55, 5 July 2022 (UTC)
- If I'm reading your contribs right, the error is that it did everything except make the AFD page, right? Just gathering info so I can add it to a Phabricator ticket. –Novem Linguae (talk) 22:28, 5 July 2022 (UTC)
- @Novem Linguae, I had the exact same problem as Curbon7 as detailed above. The Curation Tool did everything except making an AfD page. Thanks for looking into this! Netherzone (talk) 23:19, 5 July 2022 (UTC)
- If I'm reading your contribs right, the error is that it did everything except make the AFD page, right? Just gathering info so I can add it to a Phabricator ticket. –Novem Linguae (talk) 22:28, 5 July 2022 (UTC)
Appropriate draft?
I am a relatively new reviewer, but I want to make sure that I am DRAFTIFYing articles appropriately. I drafted article Draft:William Taylour because…
1. The topic has some potential merit
- He was a son of Geoffrey Taylour, 4th Marquess of Headfort.
2. the article does not meet the required standard
- No in-line citations and only passing mentions in the linked source.
3. There is no evidence of active improvement
- The article was not edited to correct these issues for several hours
4. the article does not contain copyright violations.
- None.
Does anyone disagree with my assessment? I just want to know to assure that I do not inappropriately DRAFTIFY something in the future. CollectiveSolidarity (talk) 02:06, 6 July 2022 (UTC)
- As he has an entry in the Oxford Dictionary of National Biography, which is cited in the article, the subject clearly passes WP:ANYBIO criterion 3. The use of general references, as opposed to inline references, is not preferred, but it is permitted. I don't think this was an appropriate draftification. Spicy (talk) 02:14, 6 July 2022 (UTC)
- @Spicy Since you have page mover, can you move it back? Also for future reference. Can you give me an example of something that I should DRAFTIFY? CollectiveSolidarity (talk) 02:17, 6 July 2022 (UTC)
- @CollectiveSolidarity PMR doesn't help if there are multiple edits in the redirect's history. I tagged it for G6 (or R2 who knows what the deleting admin will click) so as soon as an admin finds it and deletes, you (or anyone) can move the draft back. Happy Editing--IAmChaos 02:35, 6 July 2022 (UTC)
- Moved, thank you. CollectiveSolidarity (talk) 02:43, 6 July 2022 (UTC)
- @CollectiveSolidarity PMR doesn't help if there are multiple edits in the redirect's history. I tagged it for G6 (or R2 who knows what the deleting admin will click) so as soon as an admin finds it and deletes, you (or anyone) can move the draft back. Happy Editing--IAmChaos 02:35, 6 July 2022 (UTC)
- @Spicy Since you have page mover, can you move it back? Also for future reference. Can you give me an example of something that I should DRAFTIFY? CollectiveSolidarity (talk) 02:17, 6 July 2022 (UTC)
Concerns over AI are warranted
I cannot overemphasize the AI/Bot threat to WP. Perhaps we need WMF to create a tool that can detect such submissions? See the following articles: New AI Generates Horrifyingly Plausible Fake News, This Site Uses AI to Generate Fake News Articles, but there may be hope THIS AI CAN HELP HUMANS SPOT AI-GENERATED FAKE NEWS. An acquaintance emailed an article titled SOMEONE SLIPPED HUNDREDS OF FAKE ARTICLES ABOUT THINGS THAT NEVER HAPPENED ONTO WIKIPEDIA published in The Byte. Atsme 💬 📧 13:30, 6 July 2022 (UTC)
- Regarding the Wikipedia articles (which require a lot more than just text. e.g. references) that was a human that did that. But the threat of fake articles is real. I once uncovered a set of dozens of them. Very technical looking, heavily referenced and very wikipedian looking, but which were actually gibberish. North8000 (talk) 14:09, 6 July 2022 (UTC)
- Oops - should've made it a separate ¶. I miscalculated thinking "someone" in the headline would be the clue along with the username in the article. Regardless, my point was that if a human can slip-in hundreds, imagine what a Bot can do. In the interim, we're here and at a few other venues trying to determine whether to include/exclude (a) imperfect articles that have -0- RS or inadequately sourced, (b) moving them to Draft, or nominating them for deletion, and (c) what is or isn't notable. It appears minor when one considers the hazards of being swamped beyond human capability and the damage fraudulent (and flawed, unsourced, or poorly sourced) articles could do to the project's credibility if we don't work on prevention. j/s ¯\_(ツ)_/¯ Atsme 💬 📧 15:01, 6 July 2022 (UTC)
- We're probably particularly vulnerable on imported articles where there is less scrutiny in the wiki where it was developed and where all of the sources are non-english languages and non-arabic character sets. North8000 (talk) 17:16, 6 July 2022 (UTC)
- North, some of our vulnerabilities are insurmountable without the proper automation/tools. If it takes a decade to discover a single troll editor, how many more have gone undetected? Mark my words - the bots are coming - in fact, they've already been here, and it's just beginning. (Sounds more like the title for a horror movie.) Atsme 💬 📧 17:28, 6 July 2022 (UTC)
- We're probably particularly vulnerable on imported articles where there is less scrutiny in the wiki where it was developed and where all of the sources are non-english languages and non-arabic character sets. North8000 (talk) 17:16, 6 July 2022 (UTC)
- Oops - should've made it a separate ¶. I miscalculated thinking "someone" in the headline would be the clue along with the username in the article. Regardless, my point was that if a human can slip-in hundreds, imagine what a Bot can do. In the interim, we're here and at a few other venues trying to determine whether to include/exclude (a) imperfect articles that have -0- RS or inadequately sourced, (b) moving them to Draft, or nominating them for deletion, and (c) what is or isn't notable. It appears minor when one considers the hazards of being swamped beyond human capability and the damage fraudulent (and flawed, unsourced, or poorly sourced) articles could do to the project's credibility if we don't work on prevention. j/s ¯\_(ツ)_/¯ Atsme 💬 📧 15:01, 6 July 2022 (UTC)
- True that this is very concerning, but I don't see it as a separate problem as much as an extension of the current problem. We're always gonna be on the backfoot in the war against vandalism, whether human or AI; the determinant for us is just our capability to react. It's pretty easy to figure out historical or biographical hoaxes, but extremely difficult with regards to advanced sciences and maths. As North said, it's very easy for a vandal to portray themselves as a someone knowledgeable in a technical field (like a mathematician) and then create nonsense. Some reviewers might assume they're credible because they seem to know what they're talking about, and the topics involved are more complex than 95% of us can understand. I can only imagine how many of our articles on mathematics are complete fiction. Curbon7 (talk) 17:41, 6 July 2022 (UTC)
The big batch that I caught consisted of random pieces of copied text from random sources. So no one item was identifiable as bogus, but the whole article was random gibberish. Most readers would just figure that it was their own weakness that they couldn't understand what was being said. In technical areas outside of my fields I have some Platte river type expertise in many, and my "inch deep" knowledge was enough to spot it. North8000 (talk) 17:58, 6 July 2022 (UTC)
- Interesting topic! Wikipedia's defenses against vandalism and hoaxes (and pretty much any bad edits) are multi-layered, like an onion. The layers from outer to inner are something like ClueBot -> Huggle -> recent changes patrollers -> new page patrollers -> watchlisters -> subject matter experts. The good news is that very few bad edits are able to get through all the outer layers. So the amount of bad edits that need a subject expert to remove them is (in my opinion) pretty low. Considering that we work with millions of articles yet Wikipedia:List of hoaxes on Wikipedia only has around 450 entries, and also that the scandal mentioned in the above newspaper article is from the Chinese Wikipedia, we may be doing OK. –Novem Linguae (talk) 21:42, 6 July 2022 (UTC)
- It took 10 years to discover the 100+/- articles by a single editor. We've had the wool pulled over our eyes more than a few times. I'm not quite as confident that (1) it isn't serious enough to be too concerned over, (2) that it isn't happening as I type this response, or (3) Bots aren't busy creating hundreds of articles in main space that we may not be aware of now or for quite a while into the future, if ever – we need certain tools to do the job. I realize that it's hard to be concerned over something that isn't right there staring back at us, or taking place in some other part of the project without our knowledge. Granted, luck and good editing skills have been our allies, but why discount the fact that we have serious problems that need automated tools to fix? The WMF is quite capable of getting techs together to write scripts for a crawler or other type of hound dog algorithm that can sniff out some of the problems. It's better for us to have 5,500,000 quality articles than 6,500,000 that include a million garbage articles. Funders donate because they are expecting some level of quality in what we publish. Either way, it's certainly worth pursuing, keeping in mind that an ounce of prevention is worth a pound of cure. Atsme 💬 📧 00:22, 7 July 2022 (UTC)
- We could have a CAPTCHA as a solution to the AI problem when a user creates a new article. The counter-vandalism effort will probably go on just as long as this conflict, but there are plenty of bots and users willing to help out. In regards to hoaxes, they will eventually be discovered. Sure, there are probably some 20-year-old hoaxes that still linger out there, but nothing that gets created will go unnoticed forever. CollectiveSolidarity (talk) 21:15, 7 July 2022 (UTC)
- It took 10 years to discover the 100+/- articles by a single editor. We've had the wool pulled over our eyes more than a few times. I'm not quite as confident that (1) it isn't serious enough to be too concerned over, (2) that it isn't happening as I type this response, or (3) Bots aren't busy creating hundreds of articles in main space that we may not be aware of now or for quite a while into the future, if ever – we need certain tools to do the job. I realize that it's hard to be concerned over something that isn't right there staring back at us, or taking place in some other part of the project without our knowledge. Granted, luck and good editing skills have been our allies, but why discount the fact that we have serious problems that need automated tools to fix? The WMF is quite capable of getting techs together to write scripts for a crawler or other type of hound dog algorithm that can sniff out some of the problems. It's better for us to have 5,500,000 quality articles than 6,500,000 that include a million garbage articles. Funders donate because they are expecting some level of quality in what we publish. Either way, it's certainly worth pursuing, keeping in mind that an ounce of prevention is worth a pound of cure. Atsme 💬 📧 00:22, 7 July 2022 (UTC)
- While patrolling, I caught and CSD flagged an article yesterday, Atsme, that I immediately recognised as a hoax (if I had admin tools I could have deleted it on the spot because it was potentially also a negative BLP). I am sure that it would not have been detected by a new(ish) reviewer. This intuition only comes from years of solid experience just like an older soldier's instinct for self-preservation on active service in an urban or jungle war zone, or and older doctor's skill for diagnosis rather than relying on modern med-tech. The problems as far as begging for AI solutions are concerned, is that the WMF (or more exactly the boss of Phab) is still claiming only yesterday, lack of resources for addressing even simple but urgent software bugs. I contend that as the WMF is awash with money, there is no excuse for not hiring more software engineers. The dilemma is knowing whom to escalate to in the vague hierarchy of the WMF when requests are denied or shunted into a holding pen by Phab. Kudpung กุดผึ้ง (talk) 22:49, 7 July 2022 (UTC)
- Kudz, I plan to attend the Trustee meeting next week with bells on, and a positive attitude. Rest assured that I am cautiously optimistic about making our concerns known and getting them escalated to priority status. I am also waiting for the next in-person WikiCon and Wikimania event where I can meet the people I need to meet to get things done. I recently shared some ideas with Bluerasberry re: a special project I was thinking about submitting to WMF, and he was receptive to the idea. I just need to put a team together to help me get it implemented. Oh, and I already submitted some of my written concerns with diffs for the WMF Trustees to review ahead of next week's meeting so they will be prepared for my questions, and if I'm lucky, they'll add it to the agenda. It certainly can't hurt.
- A summary thought in response to what you described above: I do try to teach my trainees what to look for when reviewing articles, especially WP:NCORP and related BLPs, and to use their critical thinking skills when reviewing rather than simply approaching articles from a linear PAG approach. I am very pleased with the last few rounds of trainees I've had the absolute pleasure of working with - they are sharp, know what to look for, know PAGs, and while I haven't kept them under a microscope, I do try to read some of their discussions as reviewers from time to time. But like you, I'm concerned they will get burned out if we don't get the tools we all need to automate some of the mundane tasks. Atsme 💬 📧 23:48, 7 July 2022 (UTC)
- @Atsme, CollectiveSolidarity, Novem Linguae, North8000, and Curbon7: Maybe the NPP School I created 10 years ago is being underused. Perhaps most reviewers don't think they need further training but I remain convinced that many of them never bothered to read the comprehensive tutorial before asking for the right. Maybe others seriously need a refresher course. Around half the 750+ reviewers have never made a patrol while less than 100 are truly active. The tools we most urgently need are listed in the table on the Suggestions page and reviewers are being asked what they think the priorities are (see new tread below). But what we need now are reviewers who are genuinely interested in patrolling new pages, and among them, many more who have a thorough knowledge of Asian media. We are all very interested to hear the results of the Zoom meeting. Kudpung กุดผึ้ง (talk) 22:21, 14 July 2022 (UTC)
- @Kudz, they recorded the meeting, so you can watch it as soon as they get it uploaded. As you may have anticipated, it was more of a trustee vanity meeting where they provided updates of their accomplishments and introduced some of the programs they're working on, including the new Enterprise. I wasn't even given the opportunity to speak. The panel harvested and responded to a few text questions originating from the chat (YouTube?) portion of the meeting. The meeting chair (Elena Happen) harvested the questions and the respective trustees, or project leaders responded on Zoom. I thought it odd because a few days before the meeting, WMF sent us a request via the e-list to submit our questions in advance of the meeting so they could prepare, and get them worked into the agenda. One by one the agenda items were presented and discussed only by the trustees/project leaders. An hour+ later, I wasn't getting even a hint that my reason for attending was going to be discussed so I sent Elena a private text asking when the pre-submitted questions would be addressed. A short while later she responded and publicly referenced the presubmitted questions, but started and ended with mine, saying it was not an agenda item but went ahead and asked Shani Evenstein to respond. Long story short, Shani admitted that she knew about the NPP backlog, yada yada, but it appeared to me that she neither had a clue as to what I wanted to discuss nor did she seem interested in doing so. She ended by basically saying the issues at NPP are not in WMF's purview, and that it was something the community had to figure out - discussion over. Elena moved to another topic, so I sent Shani a private text asking if I could discuss it with her via email, and she texted back with an email address. Atsme 💬 📧 01:28, 15 July 2022 (UTC)
- Atsme, Shani Evenstein is Vice-Chair of the Board of Trustees and her home Wiki is Hebrew Wikipedia. Needless to say therefore that I am not surprised that she is not familiar with NPP and gave you a throwaway answer. NPP is very much in the WMF's purview - they developed it and nobody else can address the issues, nor can the local en.Wiki NPP community be expected to. The WMF is currently looking into some of the NPP issues at Phab but there are no signs of an ETA. User:Rosiestep is also a board member and it may be worth contacting her. Kudpung กุดผึ้ง (talk) 03:44, 15 July 2022 (UTC)
- Hello, @Atsme: and @Kudpung:. Thank you for all you do, for attending the last Open Conversation with the Board of Trustees (BoT), and for sharing your view in these posts. There's a lot to consider. Note, we might have different points of view, but I will do my best to answer in good faith and with the facts so people can have a full picture. A few points to consider:
- The talk was live streamed automatically on YouTube. This link is to the exact question mentioned (1:09:45). Check what I said and if you share quotes, please ensure accuracy. I did not say “NPP are not in WMF's purview”, but rather “…not the BoT purview”. There is a big difference. I also stated the BoT does not deal with operations, but with strategy, and that such matters, such as NPP, are not related to Board work. I added that there is a true desire among WMF staff to help. I believe the Product department has already replied, as they are the ones with the expertise to do so, not the BoT.
- In the past, the BoT was criticized by the community for being non-transparent, at times referred to as “Black Box”, the community not knowing what it does, and at times why it makes the decisions it makes. The fairly new Community Affairs Committee (CAC), which I serve as founding Chair of, has several goals (see Charter), one of which is as a platform to help the BoT be more transparent and allows dialogue between the community, WMF, staff and the BoT. Part of our efforts is conducting Open Conversations with Trustees; and opening askcac{{@}}wikimedia.org email, available for anyone. The BoT has never been as transparent and accessible as it is now. This is not a hoax, nor a publicity stunt, and certainly not coming from a place of vanity. We take our work very seriously. Various people, including Trustees (who are volunteers as well, BTW), spend many hours making these meetings as meaningful and inclusive as possible, with true commitment by the whole BoT, and our CEO and staff, to be as transparent as possible, to have honest and meaningful conversations (even if tough at times). We do our best to listen, address issues that have not been addressed (at times for years), and hopefully, also rebuild trust.
- For the Open Conversations to be efficient, effective and inclusive, they have to be well organized in advance. We prepare for these meetings and send an agenda of main topics, and post it on Meta, a week in advance, and make sure to go through these agenda items. Not only Trustees and staff speak. The main agenda items are followed by a Q & A on the main topics (as opposed to general updates at the beginning). We also have 20-30 minutes in each meeting for open questions in the room (and on YouTube), or pre-submitted questions - not part of the agenda topics. We view it as the best way to run such meetings, but listen to feedback and make changes to make the format more effective almost every time.
- If a community member does not like the format - they can let us know. There are proper channels to do that, like the CAC email, and the survey at the end of each meeting. We are on record for already changing multiple things to the request of the community. We’ll seriously consider a community member’s request to change the format, and answer questions in a timely manner.
- Attendees were told they could write questions/comments in the chat or raise a hand and request to speak. Perhaps you hadn't arrived at the Zoom meeting when that was stated, so we'll remember to mention it more than once in future meetings. If the point is that it was important for you to ask the question live in your own words, then point taken. In future meetings an effort will be made to allow people to ask themselves, if they are in the room.
- On a regular basis, the BoT receives questions from people in the movement on topics that have nothing to do with Board responsibilities. One of the goals of these Open Conversations is better informing the community what the BoT does and what it does not. Having a mechanism for pre-submitted questions allows the BoT to forward questions not within its remits to staff, to give more informed answers. Sending your question in advance, we were able to give an answer that included our Product department perspective. When any Trustee answers a question, these are official answers. They are usually checked and researched beforehand. I happen to know EnWiki quite well, including this issue and others, but even if I didn’t, I would not speak without checking first.
- The BoT has limits to what it can do. “We understand, we care, but it’s not related to our work and here’s where you need to go for answers” may not have been a desired answer, but it is a legitimate answer nonetheless. It’s how organizations work - each part has a different role. This question did not touch a topic under the purview of the BoT. Does that mean the Board doesn’t care? Absolutely not. We care deeply, certainly enough to make it heard, ask staff to give more context and think strategically about longer term solutions. But the BoT does not have ‘magic’ solutions. Not always, anyway.
- Trustees serve the WMF and the Movement as a whole, across all projects and all stakeholders. Any Trustee would have given the same answer, no matter what their home wiki is as the answer that was given was prepared in advance, after checking it with relevant staff. It may not have been the answer you wanted, but it was an honest answer. The BoT understands and cares about anything to do with our most important stakeholder - the volunteers who run the projects. We will continue to look for solutions and support the communities in our Movement in any way we can, even if we can’t fix everything, for everyone, right away.
- Discussions do not have to be over when an Open Conversation meeting is finished. I ended my reply with an open invitation to share more details. I shared the askcac@wikimedia.org email (the proper channel to get in touch with us) multiple times. Atsme, you have also reached out privately, and I immediately responded that I am happy to continue this via email. And obviously, discussion can continue here. Hope this helps.
- Best, ~~~~ Shani (WMF) (talk) 14:01, 16 July 2022 (UTC)
- Hello, @Atsme: and @Kudpung:. Thank you for all you do, for attending the last Open Conversation with the Board of Trustees (BoT), and for sharing your view in these posts. There's a lot to consider. Note, we might have different points of view, but I will do my best to answer in good faith and with the facts so people can have a full picture. A few points to consider:
- Atsme, Shani Evenstein is Vice-Chair of the Board of Trustees and her home Wiki is Hebrew Wikipedia. Needless to say therefore that I am not surprised that she is not familiar with NPP and gave you a throwaway answer. NPP is very much in the WMF's purview - they developed it and nobody else can address the issues, nor can the local en.Wiki NPP community be expected to. The WMF is currently looking into some of the NPP issues at Phab but there are no signs of an ETA. User:Rosiestep is also a board member and it may be worth contacting her. Kudpung กุดผึ้ง (talk) 03:44, 15 July 2022 (UTC)
- Hey folks, a couple of things in the horizon which might be helpful here (and for others dealing with the same issues).
- We hope that phab:T273220, StopForumSpam, should be helpful for combatting automated insertion of content, and might be relevant here. For the next step, at some point this quarter, it will be tested on a few wikis in report mode, so we can see that it works well in real-world setting.
- Atsme is already involved in this ticket, but for the rest of you, one interesting way of detecting hoaxes might be phab:T312841#8075735 – see the conversation there. That's promising, but still fairly theoretical – difficult to say what will work and how yet.
- Technical solutions to specific workflows is pretty far from what the Board of Trustees work on. Community Tech did a lot of New Page Patrol work in 2019. For specific tools beyond what's linked above, I think the Community Wishlist is probably the best realistic venue. /Johan (WMF) (talk) 17:38, 15 July 2022 (UTC)
- Thank you, Johan (WMF). Your reply is much appreciated...and encouraging. Atsme 💬 📧 18:03, 15 July 2022 (UTC)
Asian media
- Perhaps a NPP School template should be advertised on the main page for the project (WP:NPP). It could help alert users that they should come by if they want to join the patrol, or advise newer patrollers who need a refresher on the process. In regards to Asian media, maybe an essay can be written on foreign media to help native English speakers correctly patrol these pages. Not necessarily a policy, but some advice from an experienced patroller that has reviewed a lot of Asian media. CollectiveSolidarity (talk) 22:48, 14 July 2022 (UTC)
- CollectiveSolidarity, essays get read far less often than you would imagine. The problem here is that the Southasian continent has a greater population than all the traditional English L1 speaking regions together. To write such an essay that would be helpful to the majority of NPPers, every one of the thousands of possibly reliable Asian sources would need to be tabulated. At least in the native English speaking world we are fairly clear about what constitutes a reliable source and even one of the best known mainstream national daily newspapers is banned as a source. Kudpung กุดผึ้ง (talk) 02:12, 15 July 2022 (UTC)
- @Kudpung. So that’s it then? More work expanding this massive list? It would undoubtedly just create an issue at WP:RSN. I wish that considering all the L2 speakers that have joined the project in recent years, we could have at least had a multilingual task force assisting with source assessments. I would’ve suggested translating source assessments from the Southasian language wikis, but they may have different processes for determining reliable sources. CollectiveSolidarity (talk) 03:58, 15 July 2022 (UTC)
- CollectiveSolidarity, essays get read far less often than you would imagine. The problem here is that the Southasian continent has a greater population than all the traditional English L1 speaking regions together. To write such an essay that would be helpful to the majority of NPPers, every one of the thousands of possibly reliable Asian sources would need to be tabulated. At least in the native English speaking world we are fairly clear about what constitutes a reliable source and even one of the best known mainstream national daily newspapers is banned as a source. Kudpung กุดผึ้ง (talk) 02:12, 15 July 2022 (UTC)
- I'm beginning to wonder if we're slowing the growth of other language encyclopedias by accepting so much foreign language material and sources into this English encyclopedia? It's becomes even more concerning when it's UPE and they're marketing a person, place or thing citing foreign language sources that we have no access to, much less are able to read. I realize there's a push for globalism, but we simply don't have enough resources to cover different languages coming at us from countries with populations in the billions. I've got a headache just thinking about it. Atsme 💬 📧 04:10, 15 July 2022 (UTC)
- CollectiveSolidarity, you may find some of these resources helpful: WP:NPPSG, Wikipedia:WikiProject Korea/Reliable sources, User:Novem Linguae/Scripts/CiteHighlighter –Novem Linguae (talk) 04:25, 15 July 2022 (UTC)
- @CollectiveSolidarity and Atsme: That's it, folks, we just don't have enough resources. All the Wikipedias are owned by the WMF but being responsible for their own content they operate very different criteria for notability and sourcing. The French Wiki, for example, is notoriously lax. The en.Wiki is extremely strict. I have often said (and I hope it won't get me into trouble) that en.Wiki is fast becoming the South Asian Wikipedia in English... Kudpung กุดผึ้ง (talk)
- I doubt many have given much thought to what a global democracy would look like, much less a global WP in English. It will become so large and unwieldy readers will lose interest. Right now, globalism seems like a great deal because most are focused on the low hanging fruit of a global economy and the $$ it will bring. What they probably haven't thought about is how things will operate once states and then entire nations lose their autonomy, and are forced into a global currency, global laws, and the mass majority will be governed by some form of global democracy wherein the majority wins. The important question then becomes: who will comprise the majority? It's hard enough to get consensus on WP, but just thinking about what could go wrong in a globalist democracy gives me pause. It appears to me that foresight is seriously lacking on all fronts. Atsme 💬 📧 18:39, 15 July 2022 (UTC)
- @CollectiveSolidarity and Atsme: That's it, folks, we just don't have enough resources. All the Wikipedias are owned by the WMF but being responsible for their own content they operate very different criteria for notability and sourcing. The French Wiki, for example, is notoriously lax. The en.Wiki is extremely strict. I have often said (and I hope it won't get me into trouble) that en.Wiki is fast becoming the South Asian Wikipedia in English... Kudpung กุดผึ้ง (talk)
NPP at ANI
There is a discussion at ANI that is relevant to NPP. Wikipedia:Administrators noticeboard/Incidents#NPP Polyamorph (talk) 16:16, 9 July 2022 (UTC)
I do recommend that all reviewers read that thread, especially newer ones and ones with high patrolling counts, because there are lessons in it for all of us. Kudpung กุดผึ้ง (talk) 02:21, 10 July 2022 (UTC)
Now archived at Wikipedia:Administrators' noticeboard/IncidentArchive1103 § NPP. MarioGom (talk) 14:00, 16 July 2022 (UTC)
AP editor at RfD
An editor has been taken to RfD for very many redirects yesterday (and today's log too). I started to look at their creations, as they are on WP:RAL, so not previously patrolled by us. I found a few more to nominate and another few that I bundled to a previous nom - I'm going to bed but if anyone wants to take a quick look to catch what I missed, I'd appreciate the second set of eyes, as mine are quite tired. Happy Editing--IAmChaos 05:31, 11 July 2022 (UTC)
- Looking at your contribs, I assume you're referring to articles such as Filipinos in c –Novem Linguae (talk) 14:11, 11 July 2022 (UTC)
- The editor who created Filipinos in c (and the large number of similar redirects) appears to have begun realising that these redirects aren't deemed helpful by the rest of the community, but their future creations will definitely need to start getting patrolled again. – Uanfala (talk) 10:40, 15 July 2022 (UTC)
- An WP:RAL (redirect whitelist) user with a large number of RFDs is a bit concerning. I suppose we should ping the editor Castncoot so they have a chance to see this discussion and chime in if they wish. Also pinging admin Rosguill. –Novem Linguae (talk) 11:03, 15 July 2022 (UTC)
- I've removed them from the autolist. signed, Rosguill talk 14:45, 15 July 2022 (UTC)
- Thanks for the ping User:Novem Linguae and User:Rosguill, I just saw this now. I had misunderstood the intention of redirects as being time-saving shortcuts, as with Google-search autopopulating remaining characters or even phrasing, but now I understand that in fact they are meant to hold more weight as a standalone moniker. It would have been nice if someone had mentioned this to me all these years instead of pointing out years of apparent misdirects years later all at once in a tsunami. But now I stand corrected and will be careful to make these fit this tighter criterion. Best, Castncoot (talk) 00:05, 16 July 2022 (UTC)
- I've removed them from the autolist. signed, Rosguill talk 14:45, 15 July 2022 (UTC)
Reopening unreviewed tag discussion
Just noticed that the German Wikipedia has a slick way of doing this. First, they have a page explaining it [7]. Then, unreviewed articles have a small tab at the upper right of the article, such as this one: [8]. I like the unobtrusiveness of this method and the way that it says that quality control hasn't been done on the article. — rsjaffe 🗣️ 10:54, 11 July 2022 (UTC)
- yes, the German Wikipedia manner does seem interesting, dont know if it can be done here...--Ozzie10aaaa (talk) 14:52, 11 July 2022 (UTC)
- I agree that the German method of tagging this is good for clarity and transparency so that readers and editors understand what's happening. But it seems that this is associated with their use of Flagged Revisions and that's something that we don't have. It's not clear to me how they explain the difference between an initial review and a version review. Andrew🐉(talk) 17:50, 11 July 2022 (UTC)
- I also like the look of that button a lot better than the pseudo-maintenance template. DanCherek (talk) 18:21, 11 July 2022 (UTC)
I also like that button. On the other hand (I lived in Germany for nearly 20 years) the culture in the German language region is very pragmatic and they design good products and systems that are sometimes rather complicated. They are used to it. As an extremely multicultural Wikipedia sharing the English language, we're not. Kudpung กุดผึ้ง (talk) 02:20, 12 July 2022 (UTC)
What PageTriage bugs and features would you like us to prioritize?
Howdy folks. I was recently asked to be the NPP technical issues coordinator and I happily accepted. MB and I are going through the old PageTriage bug reports and feature requests. PageTriage is the software that pops up the NPP toolbar on the right of your screen. There's a ton of old bug reports and feature requests, and some are outdated. Anyway, are there any bug reports and feature requests that jump to mind that you think we should prioritize? Feel free to name one or two issues that are important to you, to get them on our radar. Thanks. –Novem Linguae (talk) 08:42, 14 July 2022 (UTC)
- As a relatively new NPP reviewer, I have fallen foul of a bug in the AfD option three times. The AfD process is started but the tool doesn't complete it properly. I have had to re-edit the article to manually add the deletion tag on the article (plus edit summary) and edit the article's deletion discussion page to add the recommended code ("{{subst:afd2 ..." etc). This occurred most recently (yesterday: 13 July 2022) in relation to Susan Ruth Kamunya. Thanks Paul W (talk) 09:33, 14 July 2022 (UTC)
- Do you get any popups when you experience the AFD bug? If so what do they say? –Novem Linguae (talk) 21:03, 14 July 2022 (UTC)
- Sadly, no. The initial outcome is a deletion tag that includes a lot of italicised text. When it first happened to me, I didn't know it was faulty (but wondered why it didn't immediately appear in that days's AfD listing) - another editor (User:JPxG) kindly fixed the problem and alerted me to it. Paul W (talk) 09:26, 15 July 2022 (UTC)
- Do you get any popups when you experience the AFD bug? If so what do they say? –Novem Linguae (talk) 21:03, 14 July 2022 (UTC)
- Why can’t the NPP tool use Twinkle (or the Twinkle code) for everything Twinkle does. That way, we’re reducing the maintenance effort for NPP and using a code base that’s heavily used and well-maintained. Then place the NPP effort on all the other components’ bugs. — rsjaffe 🗣️ 11:16, 14 July 2022 (UTC)
- Unfortunately this one is harder than it looks and would likely require dozens of hours of dev time on both codebases. The problem has to do with tightly coupled code. –Novem Linguae (talk) 20:23, 14 July 2022 (UTC)
- I guess it’s too late to state that tight coupling is considered a “code smell” in many programming environments. — rsjaffe 🗣️ 20:29, 14 July 2022 (UTC)
- Hmm. One simple implementation of this would be to just have clicking certain PageTriage buttons (such as the "tag" button) pop up the Twinkle tag window. Now that I'm talking this out, that might be an idea worth exploring. Eliminating code duplication would be a win for maintainability, and this would also fix the AFD bug. One complication would be that the user would have to have the Twinkle gadget turned on and fully loaded. Another complication would be that we'd lose Special:Log page curation logging of deletion and tagging actions, and notification tray notifications to the article creator. –Novem Linguae (talk) 20:43, 14 July 2022 (UTC)
- Twinkle Is already aware of NPP and has options pertaining to marking pages reviewed. Perhaps that other functionality could be added to it. — rsjaffe 🗣️ 05:33, 15 July 2022 (UTC)
- Removing tagging and deletion icons from the tool and just having reviewers use Twinkle could also help with the icon spacing issue. It would also remove those “dangerous” buttons from the interface which could improve the Ux (I.e., if I needed to do something about the article other than approval or wiki love, I’d move to Twinkle—conceptually, that seems clearer to me, and would be great for people who already use Twinkle for non NPP purposes). — rsjaffe 🗣️ 07:49, 15 July 2022 (UTC)
- Twinkle Is already aware of NPP and has options pertaining to marking pages reviewed. Perhaps that other functionality could be added to it. — rsjaffe 🗣️ 05:33, 15 July 2022 (UTC)
- Hmm. One simple implementation of this would be to just have clicking certain PageTriage buttons (such as the "tag" button) pop up the Twinkle tag window. Now that I'm talking this out, that might be an idea worth exploring. Eliminating code duplication would be a win for maintainability, and this would also fix the AFD bug. One complication would be that the user would have to have the Twinkle gadget turned on and fully loaded. Another complication would be that we'd lose Special:Log page curation logging of deletion and tagging actions, and notification tray notifications to the article creator. –Novem Linguae (talk) 20:43, 14 July 2022 (UTC)
- I guess it’s too late to state that tight coupling is considered a “code smell” in many programming environments. — rsjaffe 🗣️ 20:29, 14 July 2022 (UTC)
- Unfortunately this one is harder than it looks and would likely require dozens of hours of dev time on both codebases. The problem has to do with tightly coupled code. –Novem Linguae (talk) 20:23, 14 July 2022 (UTC)
- The message box under "Mark as Reviewed" should have the language adjusted to say " Add a message to Talk page for the creator and/or other Reviewers. This message will also be posted to the creator's talk page" This would better match the message left on Article Talk page when using and also hopefully encourage reviewers to use the feature on difficult articles, you know the ones we all skip over repeatedly. Slywriter (talk) 11:58, 14 July 2022 (UTC)
- I have two requests. 1. Move the “mark as reviewed” tickbox from just above the “add tags” button to the right of it. It is easy to hit both together on a touch screen interface and separating them would prevent this. 2. Add “orphaned” to the common tags list. I’ve never got the curation toolbar to do AfDs properly but I think we should either fix it or strip that functionality out. Mccapra (talk) 12:43, 14 July 2022 (UTC)
- #2 would be easy to implement by putting in an edit request at MediaWiki:PageTriageExternalTagsOptions.js. Making this comment as a note to myself to do this, after testing the adjusted code. –Novem Linguae (talk) 20:25, 14 July 2022 (UTC)
- I have problems with the touch interface too. Consider increasing the vertical spacing. — rsjaffe 🗣️ 20:33, 14 July 2022 (UTC)
- The list of unfulfilled bugs and features is at Open Phab requests replete with a handy table made by MB and Novem Linguae that includes the Phab tracks. It might be best to start by listing the items in the table in order of reviewers' priority rather than start requesting new features. That's why this thread has been opened in order to obtain feedback. New patrollers might not be familiar with the history of NPP and its long road to development and further upgrades. No one here would want to go back to the ancient page patrolling system which some of us have been around long enough to remember. On a more personal note, I would be against reverting to Twinkle for all or any of the features of the Curation tool. In my opinion a compact system for patrolling and controlling the quality of new articles is paramount to the ultimate quality of the encyclopedia. Therefore, as the post-2012 system was a salaried WF development (albeit with a great deal of direct collaboration with us), the effort should be to persuade them to continue to maintain and improve it rather than burden the volunteers with it, even if it means rewriting the MediaWiki extension from the ground up. One of the days the system will want to be adopted by other language Wikipedias that don't use Twinkle. Kudpung กุดผึ้ง (talk) 21:36, 14 July 2022 (UTC)
My own experience is having AFD on it malfunction many times. Usually the AFD page not getting created, or getting created without the contents. I started to try to ask about it here and the experts just said use Twinkle. North8000 (talk) 22:56, 14 July 2022 (UTC)
This is a long list of things that are quirks/hurdles for newcomers to overcome but then are fine after one gains experience. Let me know if you want me to list those. North8000 (talk) 23:01, 14 July 2022 (UTC)
- North8000. Sure. Feel free to list them if they are software issues you'd like us to prioritize. –Novem Linguae (talk) 21:32, 15 July 2022 (UTC)
- OK here goes....."list of things that are quirks/hurdles for newcomers to overcome but then are fine after one gains experience":
- Lack of nuts-and-bolts documentation on how it works / what it does. I don't mean coding-level stuff, I mean user level stuff...like "this is specifically what happens when you click this". In other words, the missing real manual at the curation tool info.
- Some way to turn it on. Right now the only way I know is to go to the page feed via page curation and look at articles that it lists. For example, to use it to un-mark a reviewed article.
- Sometimes the "message bar" default is at "message the reviewer" and if you don't watch for that you end up messaging yourself instead of the editor. I don't know why that even exists, but it should never switch to that automati9cally.
- Switch off or bring a warning screen up for "message the creator" when the "creation" was more than a year ago. Inevitably on those the "creator" is like somebody who created a redirect 10 years ago and not the actual creator which converted the redirect into an article 3 months ago. Or have it go to the person whose action turned the "reviewed" flag off.
- Have the creator in the page feed be the person whose action turned the reviewed flag off, not the one who originally created the article. So if "person A" created it as a redirect 10 years ago, and "Person B" changed the redirect into an article 3 months ago, we see the info on "Person B" not the info on "Person A"
- Sincerely, North8000 (talk) 13:13, 16 July 2022 (UTC)
- OK here goes....."list of things that are quirks/hurdles for newcomers to overcome but then are fine after one gains experience":
- The list is not really long but it obviously contains some requests that are more important and/or urgent than others. The ultimate aim is to convince the WMF rather than the self-appointed 'policy controllers' at Phab that they really should be completed as soon as possible, and that the work should be done by paid devs and not expected to be carried out by community volunteers who are already at breaking point. Theoretically the WMF works for us and not vice versa. Some of the requests go back 10 years and are not complex to implement. It should not be too difficult for reviewers to make a top 10 list of priorities, based not only on their personal preferences, but on what they believe is most important for NPP. Twinkle is not an option, Triage/Curation was based explicitly on the premise that it should be compact and not force patrollers to dash around all over the site to find a function they need, and we should not be expected to go back to the pre-2012 NPP system..Kudpung กุดผึ้ง (talk) 23:48, 14 July 2022 (UTC)
- As noted by Slywriter above, and previously by li'l ole me, it would be nice to expand the 'add message' functionality to include another contributor (maybe a drop down of contrbibutors? Shouldn't be too long with newish articles - and it's already there in the 'Wikilove' option...) other than the creator, for instance where redirects have been turned into articles. I'd also like that functionality to be open if you aren't tagging the article - so that a note can be shared with the creator/major contributor to improve the page without dropping it on the article talk and where a tag is not necessary. For instance, I find many editors neglect to put a project tag on the article talk page... Best Alexandermcnabb (talk) 05:10, 15 July 2022 (UTC)
PROD bug
I see there are plenty of reports of the AfD bug and tracking in Phabricator, though lately I've also encountered issues with PROD (in the page curation tool). When tagging a page for PROD, the tag (subst:prod) is placed and the creator is notified, though neither message includes the PROD reason – [9] [10] and [11] (page was later CSD'd) [12] – and so I need to copy and paste manually. (I could also PROD with Twinkle, but such nominations won't be recorded for the backlog drive AFAIK.) Has anyone else experienced this issue, and if so, is there already a Phabricator task for it? ComplexRational (talk) 21:18, 14 July 2022 (UTC)
- ComplexRational, Perhaps you could also list this on the request and suggestions spage. Now that we finally have some coords again, that page is being closely monitored. Kudpung กุดผึ้ง (talk) 21:37, 14 July 2022 (UTC)
- I created a bug report just now. Unlike the AFD bug, I was able to reproduce this one, which should help it get fixed quicker. Thanks for reporting. –Novem Linguae (talk) 01:55, 15 July 2022 (UTC)
Bangla, Thai, Indian and similar television
I notice a significant part of backlog is Asian television shows. Does anyone have a good indicator of reliable sources? or a general rule to follow - Channels that imply notability? I've approved some that at least listed ratings (though also discovered at least one "news site" regularly inflates, so even those require a deeper dive), prodded a few for only sourcing from promotional materials but been skipping the same ones for weeks now and wondering how we can cut down that part of the queue. Slywriter (talk) 21:27, 14 July 2022 (UTC)
- I have noticed that the majority of the articles in the New Pages Feed are nowadays on Asian topics and that a great many of them either fail GNG or are in a state where they can't sensibly be put in mainspace. It would cramp the style of the native English speaking and/or non Asian reviewers to have to decide on the reliability of Asian sources of which there is often a plethora in some even short articles. Article creators who scrape the Internet's barrel for every website that has a fleeting mention are not being helpful, especially where around half the 750 holders of the New Page Reviewer right have never patrolled a page, and fewer than 100 reviewers are truly active. Kudpung กุดผึ้ง (talk) 21:49, 14 July 2022 (UTC)
- This is one of the reasons we need to stick closer to WP:NOT - WP should not be turned into the TV Guide, a tourist brochure, a schedule for sports broadcasting, a list for game scores, a venue for promoting new movies, etc. We have become the antithesis of NOT, and now that we've got new articles coming at us in uncontrollable numbers as translations from different languages and what not, we will either have to (a) develop AI to control it, (b) put a monthly limit on translated submissions or better yet, (c) have each language encyclopedia provide for their own English translations in their respective encyclopedias, most of which are governed differently from en.WP. For example, their articles can having various language flags at the top of their original article's translation, and keep those translated articles in their project pages. We either organize properly or who knows what will happen, but it doesn't appear to me that we're headed in a good direction if we don't make some preventative changes quickly. We simply don't have the workforce that can handle billions. ~ Eternal Optimist but fading rapidly 12:31, 16 July 2022 (UTC)
Is this spam?
Lycanark. What, in blazes, is a pokemon singer? -MPGuy2824 (talk) 08:25, 15 July 2022 (UTC)
- Looks like an elaborate hoax for me. Exhibit A: Lycanark is a pokemon character, but no mention of it being a singer. Exhibit B: Google searches for those album names turn up zero results. I went ahead and tagged it {{Db-hoax}}. –Novem Linguae (talk) 09:16, 15 July 2022 (UTC)
- The infobox mentions Lycanroc - mentioned in a Wikipedia page about Pokemon: List of generation VII Pokémon. Paul W (talk) 09:28, 15 July 2022 (UTC)
- Yeah. It took me a while, but I finally figured out that these were hoaxes. See User talk:LycanrocDusk134. Since then, all the other pages they developed have been deleted . A waste of time, but sort of an interesting puzzle for the pre-Pokémon generation. — rsjaffe 🗣️ 19:40, 15 July 2022 (UTC)
- The infobox mentions Lycanroc - mentioned in a Wikipedia page about Pokemon: List of generation VII Pokémon. Paul W (talk) 09:28, 15 July 2022 (UTC)