Wikipedia talk:The future of NPP and AfC/Archive 3

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1 Archive 2 Archive 3 Archive 4 Archive 5 → Archive 7

Deletion statistics

New User deletions
Autoconfirmed User deletions
Comparison of article status

For those who have not yet seen it, we have some numbers based on deletions from the WMF. I've produced the charts based on them. This is based on taking the articles created the first week of November 2016 and checking their status as of 14 June 2017. You can see the data at User:MusikAnimal (WMF)/NPP analysis. WMF is working on getting us more numbers on this, but these also give us a snapshot of what has happened to articles that have all been reviewed. TonyBallioni (talk) 18:06, 15 June 2017 (UTC)

Wow- it's all starting to happen now. I think that those graphs very clearly show ACTRIAL is still a good idea. One thing I absolutely think we need to do is clarify if we're going for a second RfC or not. Having read Kudpung's comment here I am somewhat afraid if we try and push it on using the now over 5 year old consensus we may blow our chances and allow others to set the narrative. jcc (tea and biscuits) 19:08, 15 June 2017 (UTC)
It's the 80/20 rule at work. I'll bet a good portion of the deleted pages by autoconfirmed users were started by inexperienced editors too. These stats would also be mirrored in AfC submissions. Legacypac (talk) 19:15, 15 June 2017 (UTC)
(edit conflict) Jcc, I cross posted this here along with at the WMF report because I know some people have only one page on their watchlist, and I think the data is important to see. FWIW, I think this is the appropriate page for discussion of implementation: i.e. to launch another RfC or not. MusikAnimal has said more deletion data is coming.

I'm of the view that an RfC is needed to two reasons. First, when someone floated extended confirmed trial at the village pump in Feb (and tried to launch a hastily done RfC to reimplement ACTRIAL without any planning), the discussion there seemed to be that people expected another go round of consensus gathering, etc. Second: like you said, narrative. If this is implemented with firm consensus the fallout would be a lot more manageable in my opinion. I had been planning on launching a conversation about drafting an RfC literally the day after the WMF came out with the report, and out of courtesy to them didn't. If others are for beginning the drafting process, I think now is the time to start since we are finally getting the data that matters. TonyBallioni (talk) 19:21, 15 June 2017 (UTC)

User:TonyBallioni Yes, I think there's no harm in starting to draft. As I see it, an equally important thing that would help: getting the AfC backlog down. As I understand it, ACTRIAL proposed sending new users to wanted to start articles to AfC. Right now the backlog is three weeks long and I can imagine many oppose comments being based around the fact that we would be giving people an potentially (given more submissions would result) two month wait for their article to be reviewed which would undoubtedly put off new editors. If we were to make a concerted effort as a group I think it is not unreasonable that we could get the backlog down to reasonable level- I think in the past year we managed to. jcc (tea and biscuits) 19:44, 15 June 2017 (UTC)

There are over 9000 pages counted which is a statisticly significant sample. There is no reason to believe any other time period sampled would give a very different result. Legacypac (talk) 19:27, 15 June 2017 (UTC)

I'll bet a good portion of the deleted pages by autoconfirmed users were started by inexperienced editors too. That is almost certainly true. 4 days/10 edits is a very low threshold. If some stats could be produced on say, 10 days/25 edits, it would also be very interesting.
ACTRIAL would cut out 100% of the vandal pages, 100% of the attack pages, 100% of the nonsense and other junk pages and probably 80% of the spampages and paid advertorial and politcial candidate bios. The genuine pages left that would go through AfC would only minimally increase the load on that project, but if the suggestion to merge the AfC and NPP into one GUI were accomplished, the pressure would be reduced for the NPPers and the load could be easily shared. Also, with only 4 days and 10 edits to wait, I'm confident that those who have serious articles to post won't mind waiting till their accounts are confirmed, and they will not need to go through AfC.
If it is felt that a new RfC for ACTRIAL is necessary, there are a few things that would need to be discussed off-Wiki in great detail by a small work group first. And I do not mean IRC. Kudpung กุดผึ้ง (talk) 20:27, 15 June 2017 (UTC)
I'm game to eliminate the AfC backlog. I've been chipping away at declined pages to reduce the overall collection of junk in Draft space. I've noticed that often the creators of the better pages just put them in mainspace after submitting to AfC. I also find the real que is only a few hours when I submit pages for review I find that I'd like a second opinion on. Legacypac (talk) 22:49, 15 June 2017 (UTC)

Long-term numbers

MusikAnimal has been kind enough to provide analysis based on the time period of February 15-March 13, and the picture isn't that different. You can see all the data (and it is quite a lot and well done) at User:MusikAnimal (WMF)/NPP analysis, but I'm going to post the pie charts here, because they best compare to the ones I made above. Not much difference at all.

Thanks to everyone who has followed this conversation. TonyBallioni (talk) 14:39, 16 June 2017 (UTC)

I've just spent 20 minutes reading this whole talk page again, and it leaves me totally stunned that nearly 9 months later the WMF has not made the slightest effort to be of help. All we've got is the set of new data from MusikAnimal which proves once more that his bosses are wrong. See : Is Community Tech willing or planning to accept direct proposals? Voting seems to prioritize what editors want as a very late Christmas present, such as new nifty gadgets or convenience features (such as dead link fixing), while other important software get ignored by the vote... — Esquivalience 4 October 2016.
I think this all proves that any bending or stretching of the new data will still only conclude that Scottywong's figures were right in 2011, and that ACTRIAL is still the only realistic solution, and that the WMF hasn't really got any intention of providing any realistic, practical help. The Wikimedia Foundation is fast developing the inflexible behaviour patterns of a cult or a sect. Its dogmatic social ideology reminds me of those who believe in Martians or that the Apollo Moon landing were a massive hoax. I think we are sitting near the horns of a dilemma. Someone needs to remind the WMF that out here were trying to write an encyclopedia. Kudpung กุดผึ้ง (talk) 12:55, 17 June 2017 (UTC)
These pizza slices conclusively prove that the GFOO really exists and that the WMF's opposition to ACTRIAL is completely irrational. IMHO we would be fully justified to unilaterally impose ACTRIAL without the WMF's agreement. These stats prove that they are part of the problem, not the solution. Roger (Dodger67) (talk) 15:44, 17 June 2017 (UTC)
The WMF is running a colosseum, not an encyclopedia. They've done a good job though at pacifying the editors with all of their prolefeed they write. Esquivalience (talk) 02:51, 18 June 2017 (UTC)
Esquivalience I wonder if the prolefeed is more soylent green than panem et circenses... Roger (Dodger67) (talk) 15:59, 18 June 2017 (UTC)


Having spent time in AfC, there is zero benefit to letting new editors create pages. The pages that survive include ones where the new user managed by luck to pick a title that survived, while we don't see how much of their content survived. Legacypac (talk) 03:12, 18 June 2017 (UTC)

  • No net benefit for sure, in draftspace. AfC attracts to no-idea newcomers who at least feel in their hearts that their topic doesn't belong in mainspace. By chance, some few are suitable, I think Legacypac has a grasp on the ratio. Newcomer pages in mainspace I feel are much better on average than newcomer pages in draftspace. --SmokeyJoe (talk) 04:46, 18 June 2017 (UTC)
I wouldn't agree with "AfC attracts to no-idea newcomers who at least feel in their hearts that their topic doesn't belong in mainspace". In fact, a lot of AfC users are determined that their draft is notable and part of the problem of the backlog is that users are not being told that their draft subject isn't notable, end of story (although recently that problem is being tackled). jcc (tea and biscuits) 13:43, 18 June 2017 (UTC)

Arbtrary break 1

I can think of better things to do with experienced editor time than delete 7741 newbie articles per month. Can we please have some productive suggestions for improving the quality of content? The WMF report, in this light, is far too focused on addressing the backlog and doesn't step back and think about how to best improve the encyclopedia. As this project ages, it should be obvious to anyone who understand that there are finite number of notable topics in the universe, that we're going to want/need to shift from expanding the encyclopedia to improving existing content. ~Kvng (talk) 12:54, 18 June 2017 (UTC)

As I have pointed out every chance I get, being the most trusted source of knowledge by 2030 is one of WMF's strategic goals: keeping crap out and improving the things that should be in the encyclopedia are both key to this. You can't achieve that goal without both, and this project is half of that equation. TonyBallioni (talk) 16:03, 18 June 2017 (UTC)
One of the other strategic goals for the WMF is "Healthy, inclusive communities", and that's where the current concern is coming from. I think there are a lot more notable topics in the world than are currently in the encyclopedia, but they may be on topics that the current body of editors aren't interested in. Ideally, we'd like to see more people join the project, to bring more diverse perspectives, and it's concerning when folks say that the encyclopedia is basically done and now we just need to polish it up. That being said, you all have been working really hard on this process for a long time, without support from the WMF, and it's tough for us to come in x years late to the situation and start telling you what to do. That's why over the last few weeks we've been offering ideas, stats and analysis, so that we can earn a place in the discussion.
One thing that I'd like to ask about is whether ACTRIAL is still meant as a short-term experiment, and if so, what the definition of success would be. As we've seen lately, getting stats is a pain, and I want to make sure that we've got baseline data that helps everyone make informed decisions. If ACTRIAL is going to happen, then I'd rather be prepared with those baseline stats, so that we can see how it works. The ACTRIAL page has a list of possible statistics to monitor, but it seems like more of a brainstorming list than a complete plan.
Because these conversations have been happening specifically around the NPP process, that's been the focus -- that ACTRIAL is necessary because of the impact on reviewers and the NPP backlog. So what's the optimal end result for NPP? Obviously, there's fewer new pages to review, which means the backlog number will go down. The chart showing the makeup of the current backlog indicates that 15% of the backlog is made of pages created by non-autoconfirmed users. Some of the users who are motivated to create new pages may go and make 10 edits before creating their page, so the backlog probably wouldn't go down by 15% automatically, but let's say it's in that area.
That frees up the reviewers to spend their time reviewing more pages created by autoconfirmed users. What's the result that we're looking for there? If the problem is that there are delete-worthy pages created by autoconfirmed users, then having more time would mean identifying more pages to delete. Does a successful ACTRIAL run mean that more new pages are deleted, or is there a better metric than that? DannyH (WMF) (talk) 22:34, 19 June 2017 (UTC)
The purpose of Wikipedia is not to engineer a "healthy, inclusive community", it is to write and maintain an encyclopedia. Until a user proves that they can write coherently, they should not be allowed to create a new article. Other encyclopedias, all less relied on then Wikipedia, require strict fact-checking before anything is published in their next edition. If they even published one article with the garbage that comes into the new pages feed daily, then their reputation would collapse faster than light. But Wikipedia abuses its reputation for unreliable content by allowing this content, and nobody cares because they've done all the thrashing they needed to do and a garbage article is just another normal. Esquivalience (talk) 23:01, 19 June 2017 (UTC)
I like pie charts better than bars or lines, so here is the one MusikAnimal did that breaks down the backlog
That's a lot more than the 15% that enters the daily backlog and the time saving argument makes a whole lot more sense when presented this way than when looking at it on a day-by-day basis. Additionally, as I have pointed out before directing someone to draft space is a whole lot more healthy and inclusive than seeing a speedy deletion template within 24 hours of joining Wikipedia. Since we know that ~80% of articles created by new users end up getting deleted, it is arguably much more in line with the WMF's vision for healthy communities to implement on a trial basis restrictions on page creation. TonyBallioni (talk) 00:01, 20 June 2017 (UTC)
@DannyH (WMF): At this point I personally see ACTRIAL as a permanent improvement. I think with the data we're looking at here, it is clear that it will reduce the number of new submissions deleted and I think reducing the amount of deletion is good for the quality of the encyclopedia and it is good for retaining new editors. I would judge it a success if it reduced deletions which, in my opinion, run counter to the "healthy, inclusive community" goal.
I know you and the WMF are concerned about the "anyone can edit" mission and what contributions we'd be losing by not allowing editors to start creating articles immediately. I think the data clearly shows that the quality of these contributions is not particularly high. I am concerned that we're unable to retain editors if their initial experiences are negative. The best way to have a positive experience in the current Wikipedia environment is to start off small and observe. Creating a new article is not starting off small and the only chance to observe is having your substantial hard work deleted. I appreciate that if we restrict new editors from immediately doing what they want to do (e.g. create an autobiography), they may not sign up or edit in the first place. So I'd be willing to consider ACTRIAL a failure if it accelerated the gradual falloff we've been seeing of the number of contributions (of any kind) from new editors. I'd also be willing to call it a failure if it did not have a positive effect on editor retention. ~Kvng (talk) 16:57, 24 June 2017 (UTC)

I'm not suggesting actually doing this - but it illustrates the point of stopping brand new users from creating pages: Since 80% of pages created by "new editors" get deleted, and 50% of the backlog is comprised of these pages, and we can filter for new users not reviewed in NPP, and reviewers could basically quickly CSD or AfD (or redirect) their way through the entire 11,000 pages in the "new user" side of the backlog. If there is no obvious redirect or CSD to apply, PROD or AfD based on failing WP:N. Even without accepting a single page, such a reviewer would have a better than 80% CSD/AfD acceptance rate which is pretty high indeed. If enough reviewers did this newrly 100% of half the backlog would be gone (deleted or accepted) within a week. Any good stuff would be salvaged with an NPP Accept or Keep votes at AfD. Anything that gets kept at AfD would be accepted and removed from the que. It sounds crazy and very unfriendly but the numbers show we should assume a "new user" article needs to be deleted unless proven otherwise, rather then all articles should be saved unless proven otherwise. The main difference is that 80% of "new user" articles would be deleted in day, week or at most a couple weeks instead of over 4-6 months. Of course stopping the creation in the first place is much better. Legacypac (talk) 02:26, 20 June 2017 (UTC)

  • I'm not expecting much from this discussion as the WMF can't even bother implementing proper quality control. I feel that all of these statistics are really a means of delaying ACTRIAL for as long as possible. Call me a cynic, but no, there are enough statistics to suggest that vetoing ACTRIAL has wasted tons of time. If the WMF is serious about Wikipedia as an encyclopedia, then don't try to cough up a chart, but actually look at the evidence and do something about it. Esquivalience (talk) 02:50, 20 June 2017 (UTC)
  • What exactly stops us from implimenting ACTRIAL? Can we backdoor ACTRAIL by making a rule "every new page submission from an editor with less then 10 Edits gets a) moved to Draft and subjected to AfC or maybe b) sent through AfD unless qualifies for CSD or redirect or maybe some other quality control. We sometimes move new pages by new editors to draft now, and ACTRIAL would force those pages into DRAFT, so why not implement it by moving anything questionable but not CSD eligible to Draft space as standard practice? Legacypac (talk) 04:03, 20 June 2017 (UTC)
Nothing stops us from implementing any of the suggested local scripts that will simply prevent non confirmed users from creating new pages directly in mainspace. Our mistake in 2011 was thinking it needed to be done through an intervention to the MedWiki software and asking the devs to do it. What the devs did not understand is that they are not the policy makers. Their current policy however, is to accelerate the eventual demise of Wikipedia as a quality encyclopedia in favour of publishing impressive stats for new article creations of any kind.
The Draft idea would simply shovel the s*** over the garden wall into the neighbour's backyard, and Wikipedia is already beginning to look like the typical tip of a UK council house garden.
The other problem is that DannyH (WMF) is neither properly reading up on what ACTRIAL was actually all about and what was actually prepared, and not fully following the talk page threads that have ensued from the WMF's own essay. Current essential reading: WT:NPR, this talk page, WP:CANCER, and WP:KNPP. Danny, draw up a comfy chair, make yourself a nice cup of tea, and please do some concentrated reading.Kudpung กุดผึ้ง (talk) 11:25, 20 June 2017 (UTC)
Hi -- sorry, I was away for a few days and couldn't respond to the conversation. I want to figure out what stats could help to make decisions here -- if folks are determined to try ACTRIAL, then I think it's worthwhile to establish baseline stats that we'd compare against. That's something that we can currently offer help with, so I'm focusing on that at the moment, rather than promise something I can't deliver on. If ACTRIAL is an experiment to try out, how would we determine success vs not? Is it just the size of the NPP backlog, or are there other measures?
Also, Kudpung: I've read those pages, and I don't understand what part you're referring to. One of those pages isn't even related to ACTRIAL; it's about WMF budgeting. If there's something important that I don't understand about ACTRIAL, can you help me out and tell me what you think I'm missing? DannyH (WMF) (talk) 22:09, 23 June 2017 (UTC)
We, the community, do not allow our employees to be away for a few days. Absenteeism is grounds for firing. Please ensure that you remain at our beck-and-call 24/7. The reading of the essay by Guy Macon was recommended in order to demonstrate how dysfunctional the WMF actually is and how it fails to address the needs of its stakeholders: the readers, the writers, and the maintenance workers. In short, it has no proper leadership or any staff who are qualifed in general staff management or HR. Macon is a respected editor, is highly experienced in his field of management and his essay represents the views of a large force of volunteers who greatly outnumber even the 300 WMF employees. Macon "...(has) been around long enough to recognize the smell of good developers and good middle managers working hard and getting bad results because of bad top management decisions."
Wikimedia Foundation destroying their own work
ACTRIAL: What the Foundation is missing is that for the trial, everything was planned down to the last nut and bolt. Including the required stats for comparison. Now that your people have found out how to get stats, all that is needed is to look at the list of required stats and compare the 6 months of trial with the preceding 6 months and one month after it. Required stats: Please see WP:ACTTRIAL yet again for the section you missed at Wikipedia:Autoconfirmed_article_creation_trial#Statistics. However, I think perhaps if the WMF is interested in what's going to happen when we roll out ACTRIAL, they should know already what stats will be required without us having to do their homework for them. When you have your suggestions for stats we can compare notes. For our list please see Wikipedia talk:New pages patrol/Reviewers#Comments on Reviewing and Institutional Cultural Disconnect (New Editors), which demonstrates again that despite my persistent begging, you are refusing to read the pages I have linked you too. The paradox is that if Jorm had been allowed to complete his development of a proper landing page, all this today would probably not be necessary, and even if Kaldari picked up the pieces today, you will will refuse him the time and the budget, and another 6 years down the line we will still be where we are today and worse: Imagine a world where all the new page patrollers went on strike - please remain seated with your computers on until Wikipedia comes to a complete stop.Kudpung กุดผึ้ง (talk) 03:30, 24 June 2017 (UTC)
BTW, it appears that interest in your essay is possibly on the wane already - perhaps due to little positive response from the WMF on the very subject they began the discussion on. It seems that the only result to date has been the albeit very welcome updates to the stats Scottywong provided in 2011 - which ironically prove once again that we are right to be rollling out ACTRIAL. Kudpung กุดผึ้ง (talk) 03:30, 24 June 2017 (UTC)
Kudpung: The essay was intended to be a step in a continuing discussion about how to deal with the NPP backlog, and open up more dialogue between the WMF and the NPP community. I think that it's fairly obvious that having the WMF involved in identifying and helping to solve problems is a positive step for everyone involved. It's also obvious that you personally are not interested in talking to me, which is okay. I'm going to continue to read and think and talk about these issues, and work with the people who don't share your scorched-earth philosophy. DannyH (WMF) (talk) 03:47, 24 June 2017 (UTC)
A positive step, yes, Danny, but not when you keep finding reasons to constantly tell the huge community they are wrong, or pretend you don't understand what they are saying. Your essay was clearly a panic measure in view of what your people are calling a 'threat' which was in fact none other than a suggestion made due to your refusal to come up with the goods and find out what causes these backlogs. All we have up to now are updated stats and still no explanations. BTW, I suggest you look up a dicdef of Scortched earth. If you intend only to work with people who support your personal views, that's up to you, I can understand why I am an embarrassment to you; of course I'm interested in talking to you , otherwise I wouldn't still be on these pages, but it's difficult when you keep turning away and pretending you didn't hear. It would be nice if you would stop blocking development in the meantime of some of the features that have been requested. Excuses such as no time, no personnel, no money, tend to wear thin after 6 years, and the WMF will ultimately be responsible for the project's own demise. We, the unpaid volunteers, are the ones trying to prevent it from happening, while the WMF is throwing its baby out with the bathwater and has absolutely zero respect for the thousands of free hours the community provides without any compensation whatsoever. Kudpung กุดผึ้ง (talk) 04:10, 24 June 2017 (UTC)
It is disappointing that the WMF refuses to do something because of personal feuding. Esquivalience (talk) 04:35, 24 June 2017 (UTC)

If you liked the way that the WMF failed to act regarding financial transparency after I posted WP:CANCER, you are going to love the inevitable refusal to act after Wikipedia:Village pump (policy)#RfC: Wikimedia referrer policy closes next week. --Guy Macon (talk) 07:08, 24 June 2017 (UTC)

DannyH (WMF), with all due respect, it's not a "dialogue" if your report begins with an out-of-hand dismissal of the community's long-held consensus on this issue (that new article spam is fuelling the backlog and needs to be brought under control), and then comes up with its own entirely new narrative. The data you guys have provided is very welcome, but as many people have pointed out on the report's talk page, your conclusions bear only a tangential relationship to it. Now you refuse to talk to the one person who has done more work on this problem than anyone else (Kudpung) – is that dialogue?
I've been following this issue with interest for a while. What I've seen is that the community (Kudpung especially) have been asking the WMF for help with three very specific, very straightforward tasks: an ACTRIAL-like restriction on new page creations, an improved 'landing page' for new editors creating an article, and continued development of the page curation tool the WMF started in 2012. And they've been asking for them for six years. I've seen multiple WMF employees breeze by with vague promises that they will happen, but they don't materialise. Instead we get a report telling us that we're all wrong and we should just ignore the backlog. In fact, it's our fault the backlog exists in the first place. Surely you can see where the frustration on this talk page is coming from? – Joe (talk) 18:37, 24 June 2017 (UTC)
Joe Roe: Yes, I absolutely can. I apologize to Kudpung for making this personal on my end. When discussions get heated, it can be difficult to step back and just hear the ideas, but that's my problem, and not his.
I understand and agree with the frustration over the Foundation's lack of investment in issues like this. That's something that's changing, but it's changing slowly, and the Community Tech team talking and thinking about this issue is a step in what I believe is a more positive direction. I'm trying to be mindful not to promise things that we're not sure we can deliver; that's why our team has been focusing on digging in and learning more about this situation, making small improvements, sharing thoughts and ideas, and gathering stats that will help us see the full picture of what's going on.
Unfortunately, the three tasks that you reference are actually not straightforward. :) They're big projects that require a lot of investment from a product team, if we're going to do them well, and I don't have that kind of commitment from the Foundation yet. So our team -- me, Kaldari and MusikAnimal -- are doing what we can with the resources that we currently have.
I knew that some people would agree with the conclusions in the report, and some people would very strongly disagree. The main point I'm making in that piece is that the changes that have taken place in the NPP process over the last couple of years have contributed to the growth in the backlog, and that (in my opinion) doubling down on those changes will not make the situation better. I'm not trying to insult people who are doing really important work for the encyclopedia. I'm suggesting another way to approach the current problem which I think would be more effective long-term. -- DannyH (WMF) (talk) 20:01, 24 June 2017 (UTC)
Respectfully DannyH (WMF), is there someone else at WMF who we should be liaising with who will not color our well-thought-out requests with their own personal opinions. We need an ambassador who will advocate for our needs, not someone who has (apparently) never patrolled newly created articles and who dismisses our legitimate concerns. - MrX 22:26, 25 June 2017 (UTC)
I was personally very sympathetic to the ideas in the report, the problem is that Wikipedia is supposed to be run by a consensus of the people who work on it. The consensus is clearly counter to these ideas and that means that DannyH (WMF) and I are unlikely to get what we want and we need to find a way to realign to that. The best way I've found to do that is to actively participate in the processes we're talking about. I encourage you and your fellow employees to spend some time at WP:AFD, WP:NPP, WP:AFC, etc. Big data is certainly helpful towards understanding what's going on here but so is the day-to-day small data. The consensus to add additional restrictions to WP:NPP and some of our other processes comes from the people who work on it every day. You can't come in and tell them they need to change how they're doing it without first getting down there and seeing what exactly we're dealing with. ~Kvng (talk) 12:54, 26 June 2017 (UTC)

Arbtrary break 2

Commitment from the Foundation - the problem is, Danny, when you speak, you make it sound as if you are the Foundation, leastwise that you personally are deciding he policies as to what issues will be technically addressed. If you are not, then we need to escalate to your superiors and make a case for more of the glut of 300 employees to be drafted in to resolve these issues - either to your team or to another one.Better still, with the huge surplus of funds, create a special team to do it, and I would vote for Kaldari and MusikAnimal to run it. Speaking of which, what happened to Wes Moran and the 1 hour conference Kaldari and I had with hum? Did he walk out or was he fired? Who is in charge of the circus right now? Someone else who needs a whole year to get up to speed? How do we get an audience with Maher and make her understand what's going on? At least Gardner was showing an interest in these issues. Kudpung กุดผึ้ง (talk) 20:29, 24 June 2017 (UTC)
Kudpung: There is a team that Kaldari, MusikAnimal and I run: the Community Tech team. I'm the product manager, Kaldari is the engineering manager, and MusikAnimal is one of the developers. We have the responsibility to determine what our team works on, given the current parameters: 75% Community Wishlist, 25% Community Engagement requests and projects for smaller groups, plus a subteam that's working on Anti-Harassment Tools. We don't have the resources right now to take on the landing page or major Page Curation Tool work. We may in the future, I just can't promise right now because I'd want to be sure we could deliver. Our team has grown a lot in the two years that it's existed, and I'm hoping it continues to grow this year.
We were asked by our bosses if we'd be willing to take on a small-scale project that started at the Wikimedia Hackathon last month, to look at the NPP problem and figure out what could be done in the short-term and the long-term. We were given the responsibility to figure out what that short-term work would look like, and the results so far are the bug fixes that Kaldari and MusikAnimal did, the report that I compiled, the stats that we're putting together, and the discussions that we've been engaged in. I'm not sure if we'll get the resources that would allow us to work on a more long-term solution; it's only been a month, and resourcing decisions can take a long time. If you feel like escalating that you want Community Tech to get more resources, that is absolutely fine with us. :)
Wes Moran left the Foundation at the end of February. The interim VP of Product is Toby Negrin, who's been overseeing the Readers and Community Tech teams; he's been with the Foundation since 2013. Toby organized the Community Tech team, and he's very supportive of the work that we're doing. For the rest of the circus, you can see the Staff and contractors page on the wikimediafoundation wiki. -- DannyH (WMF) (talk) 21:50, 24 June 2017 (UTC)
Landing pages are not hard to program. I have designed a basic yet functional one using only templates at User:Esquivalience/LandingPage. The design of the proposed workflow is described at mw:Article Creation Workflow, and the WMF doesn't need a whole team to program such a trivial function. Esquivalience (talk) 03:18, 25 June 2017 (UTC)
Esquivalience: When my team takes on a new project, it requires research and design work. For the landing page, that would require digging into the problem to make sure we understand the right way to approach it, and that this proposed solution is the right way to achieve the goal. In this case particularly, I'd want to do user testing to make sure that the new feature actually results in more good pages and fewer bad pages. This is a very important workflow, I'd want to make sure that we'd do it right. I know that's frustrating because there's already an existing spec, but that's the way my job works. I don't think you'd find any product manager in this kind of team who'd pick up an old spec and just start building it. -- DannyH (WMF) (talk) 18:04, 25 June 2017 (UTC)
It's quite obvious that when Wikipedia tells people how to create articles, they'll create good articles. A landing page is not an entire product that requires a fifty-page PDF, it's an incremental addition. I know of many open source projects where discussion and a short plan is all that is needed to add features more wide-ranging than a landing page. And for the projects who do need plans (like Python Enhancement Proposals for the Python language), they don't make it seem like writing blueprints for a nuclear power plant. And the "old" article workflow design works just fine with a few modifications. There is nothing broke and little to improve about it. Finally, it is quite absurd really that the most-used reference work in the world does not tell those who creates its entries how to create good entries. That's recklessness on another dimension. Esquivalience (talk) 21:27, 25 June 2017 (UTC)
DannyH (WMF) we know when Moran left. I had a 1 hour Skype conference with him just before he did - remember? What you have done yet again characteristically, is no-too-subtly avoided the question: Why did he go?. Did he go because he was fed up with the chaos, or was he fired, or did he go after a better offer elsewhere? All answers would be understandable. The only other person I know of who is so good at dodging question is Theresa May. Kudpung กุดผึ้ง (talk) 06:56, 25 June 2017 (UTC)
Kudpung: Oh, I can't answer that question. We're not allowed to comment on someone else's employment like that. I think that's a pretty standard HR rule for any organization. -- DannyH (WMF) (talk) 17:19, 25 June 2017 (UTC)
DannyH (WMF), Wes Moran's departure is cloaked in mystery. It will come out somewhere sooner or later if not on Wikipedia, on one of its detractor forums. The very fact that nothing is being said gives rise to conjecture of a negative sort - it's human nature, so for the time being, we'll assume what we will assume, which underlines one more how disorganised and dysfunctional the Foundation actually is, and why we're not getting any cooperation from anything you are apparently in charge of. Perhaps it's because your superiors won't let you help us - we just don't know. But it proves once again what a waste of time the Skype conference with him and Kaldari was. And you've already stated that you are not prepared to work with me, which is a great shame. Kudpung กุดผึ้ง (talk) 06:06, 26 June 2017 (UTC)
DannyH (WMF), is 2016 Community Wishlist Survey/Results an accurate reflection of your team's priorities and is there master plan that shows how each of these is resourced, progress to date, projected release, etc.? It seems like some of the items on that list are maintenance related and that your team would presumably work of several items concurrently. Insight into how our needs are being handled by WMF would probably reduce the amount of consternation that you are witnessing here. What would have the opposite effect is comments like "We don't have the resources right now to take on the landing page or major Page Curation Tool work. We may in the future...". As far as I can tell, these issues have been simmering on the back burner for several years. I hope you can appreciate that we don't want to keep hearing mañana.- MrX 22:01, 25 June 2017 (UTC)
MrX I'm preparing a mid-year report on the Community Wishlist progress. It isn't done yet, but the thumbnail sketch is: two of the top 10 wishes are complete (#9 Mr Z-bot's popular pages reports, and #10 User rights expiration), and three of the top 10 will be released within the next few weeks (#5 Rewrite XTools, #6 Wikitext editor syntax highlighting, #7 Warning on unsuccessful login attempts).
You have all been waiting a long time for the Foundation to give this any attention, and right now you're mostly getting me talking rather than anything that looks like progress. I'm trying to be careful not to promise things that we're not sure we can deliver, but yeah, the "we may in the future" is half-doing that anyway. I may just be frustrating people even more by keeping these conversations going. I hope that I'm not. I'm learning a lot from these conversations. -- DannyH (WMF) (talk) 18:27, 26 June 2017 (UTC)

DannyH (WMF) When my team takes on a new project, it requires research and design work. For the landing page, that would require digging into the problem to make sure we understand the right way to approach it, and that this proposed solution is the right way to achieve the goal. In this case particularly, I'd want to do user testing to make sure that the new feature actually results in more good pages and fewer bad pages. This is a very important workflow, I'd want to make sure that we'd do it right. I know that's frustrating because there's already an existing spec, but that's the way my job works. I don't think you'd find any product manager in this kind of team who'd pick up an old spec and just start building it.

This is tantamount to saying that you will only develop what 'you' want to develop, which is the common WMF policy we are used to, and is only yet another delaying tactic. More importantly, it is contrary to the bold statements you and your colleagues have been making about wanting to work together with the community. Anything would be better than nothing, and if we wait for you we'll be waiting for another six years. You are treating the thousands of volunteers as idiots who don't have clue about communication and software development where we probably outnumber the Foundation on these competencies 10,000:1. Just because you are paid for what you do doesn't mean you are any better. I am telling you that 99% of the conceptual design of Jorm's project was accurate and complete, would perfectly address the urgent and immediate need of today, and just needs a couple of days to write the code. And if you won't do it, you'll only give the community even more reason to go ahead with ACTRIAL without it, demonstrating once more that the Foundation's ultimate goal is to undermine, rather than encourage, the community's volunteer spirit. Kudpung กุดผึ้ง (talk) 06:29, 26 June 2017 (UTC)

Landing page

The 'other half' of the Page Curation project was to create a proper landing page (see the link on the main page) to assist those genuine creators who under ACTRIAL would have to wait 4 days and 10 edits. Kaldari tells us that development was shelved because Notations was more important. Strange however, because even Sue Gardner was bubbling with enthusiasm for it and even remarked what a wall of text the current Wizard is [1]. Kudpung กุดผึ้ง (talk) 12:22, 24 June 2017 (UTC)

Commons' upload wizard is amazing and simple to navigate in my opinion. Would it be possible to piggyback of their design for something like this? TonyBallioni (talk) 13:03, 24 June 2017 (UTC)
The thing is that a few hundred lines of code would be enough to implement such a landing page, and the WMF is treating it like the Manhattan Project. Esquivalience (talk) 15:28, 24 June 2017 (UTC)
"A few hundred lines of code"? Have you seen the workflow? We already wrote over 1000 lines of code back in 2012, and it was half finished at best. I would certainly love to finish this project (although it would mostly have to be rewritten from scratch at this point), but there are lots of competing priorities :( Kaldari (talk) 18:05, 26 June 2017 (UTC)
If you guys spent 1/10th the effort you've spent criticizing the WMF on convincing people to vote on the Wishlist Survey proposal instead, the Landing Page would be finished and we wouldn't be having this discussion. Instead you guys just dismiss the survey as a useless distraction, and none of you even bothered voting for the Landing Page proposal that Kudpung created. If the proposal had received 7 additional votes, it would have made the Top 10 and we would have done it already, and yet none of you even voted for it. You can't complain that the WMF won't listen to the community, when we've provided an open, transparent means of letting the community prioritize the work of the WMF, and yet none of you (besides Kudpung) has chosen to use it. Kaldari (talk) 18:05, 26 June 2017 (UTC)
Kaldari, that is simply not true. If you look at that proposal, you would see that @MrX, Chris troutman, Jbhunley, DGG, BU Rob13, Onel5969, Esquivalience, Blue Rasberry, and Jcc: and myself all voted for it and supported it. That's 10 people who are either NPP/NPR regulars or are regulars to the reform discussions. Everyone who has commented thus far in this section is on that list. TonyBallioni (talk) 18:21, 26 June 2017 (UTC)
You're right, I was looking at the wrong proposal :) Regardless, it's too bad more reviewers didn't support it. Kaldari (talk) 19:02, 26 June 2017 (UTC)
You're right Kaldari. We should have campaigned our asses off to get this pushed to the top of the top ten list of piddling maintenance tasks. I read the 2016-17 plan, and it's obvious that WMF has no intention of allocating anything beyond minimal resources to maintaining the technology platform in a way that supports contributors trying to prevent the encyclopedia from being overtaken by spam.- MrX 18:35, 26 June 2017 (UTC)

Basic question: Who is in charge here?

I have managed hardware and software projects for many, many years, and I can say from experience that there is one development method which inevitably leads to disaster. It is commonly called the "throw it over the wall" method.

This method involves not having a customer who can reject the product if it does not meet the customer's needs. Any time that you have software developers who think they know the needs of the people who will be using the system better than the customers themselves know their own needs you have the beginnings of a disaster. Add the ability for the developers to say "you get what we decide you will get, and no amount of complaining will change this; we are in charge here" and the disaster becomes a certainty.

Back in 2014 the Wikimedia Foundation implemented "Superprotect" -- a new user right that enabled the Foundation's paid staff to overrule Wikipedia administrators. The Foundation then immediately used Superprotect to unilaterally enforce the rollout of the Media Viewer software on the German Wikipedia.

This resulted in a Letter to Wikimedia Foundation: Superprotect and Media Viewer[2] with roughly a thousand signatures. (The talk page for that letter is also quite insightful) BTW, this wasn't the first petition on this general subject; see this one from 2008:[3]

Yet still we have WMF developers insisting that they are our bosses and that we must obey them.

So, who is in charge here? Is it Wikipedia or the Wikimedia Foundation? And will it require something like User revolt#Wikipedia to change that? — Preceding unsigned comment added by Guy Macon (talk • contribs) 02:24, June 25, 2017 (UTC)

The problem is that the WMF software department is like a playground where pointless programs are written so that the WMF's own Ministry of Truth reports don't seem so implausible to the donors (compare shareholders). If the WMF pretends that they're making progress, then couple with sob stories about how the WMF's kitty is gone and they got themselves a steady stream of revenue. Esquivalience (talk) 03:30, 25 June 2017 (UTC)
Esquivalience, did you synthesise your post from the comments I've been making over the years and recently? If not, and you wrote it from the top of your head, I've already given you a barnstar for it! Kudpung กุดผึ้ง (talk) 07:18, 25 June 2017 (UTC)
I agree wholeheartedly with the above. I have also managed software development and systems integration teams, and I'm not accustomed to developers dictating requirements. Moreover, I'm concerned that we have been going around in circles for years trying to move this forward to no avail. Perhaps it is time for something on the order of a revolt to get this going in a positive direction.- MrX 21:41, 25 June 2017 (UTC)
I just could not stand the countless Newspeak excuses for the simplest problems. It's hard not to make a carbon copy of all the editors who have criticized the WMF in the past, it's not failure for specific reasons, it's failure on all vectors, they specialize in failure. Esquivalience (talk) 22:00, 25 June 2017 (UTC)
I don't want to be one of the WMF-bashers because relationships are important, even on the internet. But it's becoming increasingly evident that nothing meaningful will be done to address the substantial issues raised about new page creation. I not sure what more we can do to try to convince WMF that we have serious problems and we really need WMF's help.- MrX 22:10, 25 June 2017 (UTC)
it seems to me that one of our current problems is the amount of time people interested in doing NPP spend in debating how to go about it. From what I see here, it seems obvious that there continues to be a consensus for Actrial. I do not think it will solve all our problems, but based on the figures given, it may give a 5 to 10% smaller backlog. It would seem simpler to me to implement it and see, rather than keep discussing. I do not see that the WMF currently would actually oppose it. I think they;re more inclined to help than to argue, and I hope that's true of us also. DGG ( talk ) 03:54, 26 June 2017 (UTC)
MrX, ACTRIAL can be easily implemented without any help from the WMF. You either flip the switch through the blacklist or through edit filters and boom, its here locally. WMF help on the landing page is needed and would be quite welcome, but the part of the project that gets the most attention can be technically implemented without them. Its a question of if consensus still exists for it, which method we prefer, and when to flip the switch if a consensus still exists. TonyBallioni (talk) 03:59, 26 June 2017 (UTC)
Dummy edit. Kindly ignore this comment, but let it remain here. —usernamekiran(talk) 04:15, 26 June 2017 (UTC)
TonyBallioni, it's looking like we will have to do it locally since we're getting almost no cooperation from the WMF and I don't see that changing in the foreseeable future. I shudder to think what a kludge a local solution will be, but it will certainly be better than the status quo.- MrX 11:43, 26 June 2017 (UTC)
Show that it is a top priority for the community and the WMF will be happy to make it their top priority as well. (Getting 70 editors to vote for a Wishlist proposal should be much easier than organizing a revolt.) Kaldari (talk) 07:40, 26 June 2017 (UTC)
Getting them to include your idea and put it up for a vote, not so easy.[4][5] Unless you can find a place where my idea was put up for a vote along with the other Community Wishlist items, of course -- I might have missed it somehow. --Guy Macon (talk) 08:08, 26 June 2017 (UTC)
What Danny does not (or does not want) to understand, Ryan, is that what we have here is a major critical issue that could well end in a revolt that would be easier to organise. This is not for what I call the users' annual letter to Santa for bots, gadgets and other convenience features. We're talking here about something that affects the very fabric of Wikipedia and its reputation.
None of the 265 items on that list came anywhere near causing so much discussion and polemic or driving a cast iron wedge between the community and the WMF. Eight of the top ten on that wish list have still not been completed or addressed and we're rapidly approaching the 2017 list.
If your team is saturated, what you guys need to do, as you are too low on the pecking order to decide for yourselves, is to lobby the very top for more people or as I said somewhere else, create a dedicated team of devs to work on the things the stakeholders demand rather than what the devs want to do. But even there, chances are they (she) won't understand and with a wink of a wrist will simply say vaguely "Oh, I think I heard something about that. Look into it will you" like she did last time you spoke to her about it. WMF is getting so big now that it's as difficult to get an audience with the top person as it is with the Queen - but at least her Maj still goes on a walkabout and shakes a few hands. When the revolt comes, it's going to be more than hands that will shake. Kudpung กุดผึ้ง (talk) 11:03, 26 June 2017 (UTC)
Kaldari - I'm sure I can collect at least 70 votes (actually, it looks like only 29 votes would make the project highest priority]) from users who will agree with me that defending the encyclopedia against the constant flood of spam articles is far more important than syntax highlighting (already a capability via javascript). Before I do though, I have two questions:
1. If the Community Wishlist Survey is the mechanism for process for getting development projects prioritize, how is it that the 52nd item on the list with 29 votes is done, while New User Landing Page with 69 votes doesn't even have an owner?
2. Does your position in the WMF give you the authority to re-prioritize the development of the landing page project once I collect those votes, or are you speaking for someone else with that authority?
Thank you for your help with this.- MrX 11:25, 26 June 2017 (UTC)
Don't be so sure that you can collect at least 70 votes when the WMF is free to refuse to put your suggestion on the ballot simply because they don't like it.[6] --Guy Macon (talk) 15:51, 26 June 2017 (UTC)
I have no intention of waiting for next year's wish list. This needs to be addressed now, and if we need 29 more people to say so, I can easily find them. I also want to know who is the person at WMF responsible for prioritizing development work, how those decisions are made, and who at WMF is responsible for making sure the work gets done and reporting back to the community. We need transparency, and an open channel of communication that leads to productive results. It hasn't happened yet.- MrX 16:18, 26 June 2017 (UTC)
@MrX: The process for prioritizing the items from the wishlist is documented at meta:Community Tech/Community Wishlist Survey description. The short version is that the Community Tech team takes the top 10 items and works on the easiest ones first. Last year we finished 8 of the 10. This year I think we'll finish 7. Volunteers or other teams are welcome to work on the other proposals. The 52nd item was completed by a volunteer. All of the development teams at the Foundation already have their priorities largely defined for the next year except for the Community Tech team since our priorities are defined by the annual Community Wishlist survey, rather than the annual planning process (which is informed by the community strategy consultations). If a proposal for a New User Landing Page makes the Top 10 this year, I personally guarantee it will be completed. Kaldari (talk) 17:15, 26 June 2017 (UTC)
Kaldari, thank you for answering one of my questions. If I understand correctly, any development requested by the Wikipedia community will not occur unless is one of the top 10 most popular requests (including maintenance requests), requested by users, including developers, across all Wikimedia projects. With nearly 277 employees and a budget of more than $60mm, ten community-requested development projects are allocated for all Wikimedia projects for the entire year.- MrX 18:15, 26 June 2017 (UTC)
@MrX: No, maintenance requests and bug fixes are handled through Phabricator requests, and despite what everyone here keeps claiming, the WMF has done significant maintenance work and bug fixes on Special:NewPagesFeed and the Curation Toolbar in the past year (26 non-localization-related patches since January). Kaldari (talk) 18:55, 26 June 2017 (UTC)
@Kaldari: 'Fix Mr.Z-bot's popular pages report' is not maintenance?- MrX 02:13, 27 June 2017 (UTC)
@MrX: It would have been, except that the reason the bot stopped working is because it was built on old technology (parsing the old pageview logs that we don't produce any more). It had to be completely rewritten from scratch. Your point is taken, however. A lot of the requests are relatively small tasks. In the case of creating a new landing page for article creation, that task is just about the perfect size for the Wishlist. Large enough that it's unlikely to be done without some dedicated focus from a WMF team, but small enough that it can be successfully completed in a few months. Kaldari (talk) 02:32, 27 June 2017 (UTC)
As a developer and technical manager, I just want to say that this whole discussion makes me a bit ill. I understand that the Wikipedia community could collectively be accurately described as the customer from hell but it seems like the WMF's response has been to create the technology development organization from hell. Someone's got to be the bigger man here and it is unlikely to be the chaotic Wikipedia community of unaffiliated volunteers. ~Kvng (talk) 16:27, 26 June 2017 (UTC)
That is actually a very reasonable argument. Kaldari (talk) 19:19, 26 June 2017 (UTC)
  • Flip the Switch Already. If implimenting ACTRIAL can be done locally with edit filters and blacklist then do it already. There is near universal support from all involved in NPP. I don't see fighting over this with WMF as productive. Let them work on what they prioritize and let the active NPP group do what we NEED to do to slow down the Flood of Crap. While we are at it, implement a "near blank" new page filter like the link blacklist one. No saving a new page with almost no content (unless it's a #REDIRECT). That will cut way down on errors, tests, and nonsense pages we process. Legacypac (talk) 17:59, 26 June 2017 (UTC)
    • Obviously an admin needs to do it, but yes, let's get going on this. It is abundantly obvious that the WMF will not help us with this anytime soon. @Kudpung: How do we implement this?- MrX 18:19, 26 June 2017 (UTC)
      • Build a landing page in template namespace then ping me. ~ Rob13Talk 18:42, 26 June 2017 (UTC)
  • Obviously it's probably unsuitable for actual use, but here is what a landing page could look like. Esquivalience (talk) 01:12, 27 June 2017 (UTC)
  • Again there are probably a lot of problems (one of them is that it doesn't work with Wikipedia mobile) but something to build upon. Esquivalience (talk) 03:11, 27 June 2017 (UTC)
@Esquivalience: If this is to be done by edit filter, we only need the latter. Is it really desirable to tell them exactly when they'll become autoconfirmed? That may encourage gaming the system to get the stats/avoid review. This is very common already, with edit filters designed to try to detect it but not always successful. We could just say "new editors". ~ Rob13Talk 10:33, 27 June 2017 (UTC)

Statement by Jimbo Wales

Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia_talk:The_future_of_NPP_and_AfC/Archive_3&oldid=798477683"