The Signpost

In focus

NPP: Still heaven or hell for new users – and for the reviewers

Contribute  —  
Share this
By MB, Novem Linguae, and Kudpung

MB and Novem Linguae took on the task of joint lead coordinators of New Page Patrol this year following the long void left by other coordinators who had moved on. Kudpung was de facto coordinator for many years before retiring from it in March 2017. The views expressed in this article are those of the authors alone and do not reflect any official opinions of this publication.

Just what is NPP? Why does it need the WMF? Why does it need YOU?

Graph of NPP backlog
The slight downward trend in the backlog from the 2nd week of September is due to the work of just one reviewer returning from an absence of several months due to burn out. On average, the New Pages Feed receives 750 - 1,000 articles every 24 hours.Database: Top new article reviewers

Four years ago, in October 2018, the Signpost article "NPP: This could be heaven or this could be hell for new users – and for the reviewers" reported on the New pages patrol process and the challenges it faces. Today, those issues produce new hurdles to keeping the encyclopedia clean – and the need for upgrades to its software is even more acute.

Earlier this month, in a last ditch attempt to draw the Foundation’s attention to the need for developer work on the PageTriage/Curation tools, an open letter from 444 users and admins was published along with notifications to senior WMF staff, and to the members of the Board of Trustees. In the absence of a more detailed hierarchical organigramme, it is assumed that the Contributors Product Management team subordinate to the 'Product' section headed by new CPTO Selena Deckelmann, is responsible for declining or delaying the urgently needed attention.

Out of a total of 165 open NPP tickets at Phabricator, over half are 4 years old or more with many going back to 2012

A Page Curation tool
A tool in the Page Curation interface of Page Triage
A Page Triage tool
The reviewers' New Page Feed in Page Triage

New Page Reviewers use the page curation system to review new articles and pass them for indexing for search engines, or to propose them for further improvement or deletion. Unfortunately, the community has too few capable and competent people at NPP and despite occasional drives to address it, today's backlog still stands at an untenable level. Software improvements, both bug reports and feature requests, that would help and encourage reviewers do this job have been languishing unaddressed at Phabricator for months and even years. Commenting on the WMF's rejection of ACTRIAL in their 2011 article in The Signpost, users Skomorokh, Jorgenev, and Daniel Mietchen suggest that:

Wikipedia has over 700 New Page Reviewers. Only around 100 are moderately active and only 25 or so perform the majority of the reviews.

Knowing that an appeal was in preparation, a WMF director of product engineering made a pre-emptive post on the task force talk page in which they placed the blame for the lack of software support squarely at the feet of the New Page Patrol team for not having submitted requests to the Community Wishlist. The Wishlist is an annual programme since 2015 where contributors from all Wikimedia projects can ask for those small changes that they would most like to see. Indeed, the NPP group did avail of the process in 2018 and completely swamped it, but that was four years ago. Paradoxically, the response continued by listing other WMF engineering commitments as an explanation of why they don’t have the funds or capacity for NPP. There was also an opaque reference to a need to rewrite the tool: "current projects that we've prioritized above rewriting PageTriage." If they believe it is beyond normal maintenance, that may be another reason why they are providing little support. The WMF's post was made before the letter was published and the NPP team does not consider their comments to be wholly felicitous.

Despite the bold claim in their Annual Plan 2022-2023 Goal #1:

the WMF makes no mention of supporting the trusted knowledge in their flagship project, the English Wikipedia.

How well informed is the Board of Trustees?

During a July live streamed 'Conversation with the Trustees' (see transcript below), in answer to a pre-submitted question from Atsme, a New Page Reviewer: "What is the Foundation doing to protect New Pages Patrol and Articles for Creation from system overload?" the response from the board's vice-chair, Shani Evenstein, stated that PageTriage was 'community developed' (which it was not: Engagement strategy - New Page Triage) and that NPP volunteers should wait for at least six months and request the work through the Wishlist.

While researching for a reply, in contrast to their claim of " have honest and meaningful conversations [...] 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".[3] they were apparently either misinformed or they misinterpreted the information provided by their source. Insisting that important technical matters are not at all within the board's 'remit', it is interesting to note how unaware the board appears to be of the importance of the flagship encyclopedia to the Foundation's very existence. In a follow up thread on the reviewers' talk page Ms Evenstein offered some words of encouragement, giving hope for future collaboration:

I cannot promise an immediate solution, but I can promise it will be properly discussed. To make sure expectations are realistic, I will add that this topic / issue / problem statement requires further discussions internally, both with our CEO and our new CPTO. As you may know, we have just hired a new CPTO, who will be starting in August. Whatever operational solutions WMF will come up with for this stated problem, it will have to include her. Till she settles in, and till we are able to strategize around this topic (and other related technologically-related topics) further in collaboration with staff, our Product department continues to be aware and continue to work on it to the extent they can; but I hope it is clear that talking about longer-term solutions, a bit more time will be needed to make sure this is properly discussed.

Ms Evenstein has recently been re-elected to the board, while newly elected Mike Peel comes directly from the English Wikipedia community. A user since 2005 and an administrator since 2007, he was one of two candidates campaigning on a platform of more board involvement on the shop floor:

Although there have been hiccups in the past, some serious, the board generally acts within the movement's best interests, but anyone watching the video of the Board of Trustees responding to Atsme's question cannot fail to recognise the board's reply as inadequate and mildly patronising.

No money?

WMF receipts and expenditure up to 2021

Writing in The Signpost Op-ed Wikipedia has cancer about the Foundation's flow of funds, Guy Macon, a financial consultant, states:

…their poor handling of software development has been well known for many years. The answer to the WMF's problems with software development has been well known for decades and is extensively documented in books such as The Mythical Man-Month and Peopleware: Productive Projects and Teams […] This failure is almost certainly a systemic problem directly caused by top management, not by the developers doing the actual work.

Macon's more recent essay, an updated version of Wikipedia has Cancer, is a testament to the runaway spending on pet projects and non-core functions while neglecting the volunteers, and still leaving the Foundation with current disposable assets in excess of US$200 million The volunteers are the project's major stakeholder and the NPP team on whose work the reputation of the articles depends, is hoping for a substantive response and one that comes from the Foundation's senior executive staff, especially CPTO Selena Deckelmann, and Maryana Iskander, the CEO. In view of the rapidly growing surplus of donations, the stock justifications for inaction are wearing thin; the community is trusting in offers of serious engagement on quality control, rather than further rejections and intangible reasons why it can't or won't be done.

As of press time, the New Page Review team has received an acknowledgement by email from the Trustees, but no official public response to the letter from them or the WMF. We reached out to the board and the WMF regarding this article in the Signpost. The board pointed out a minor error in the staff which has been corrected, but the WMF has not offered any comment.

Related articles
New Pages Patrol

NPP: Still heaven or hell for new users – and for the reviewers
30 September 2022

Much ado about Fox News
30 September 2022

NPP: This could be heaven or this could be hell for new users – and for the reviewers
28 October 2018

NPR and AfC – The Marshall Plan: an engagement, or a marriage?
29 June 2018

ACTRIAL results adopted by landslide
26 April 2018

ACTRIAL wrap-up
29 March 2018

The plight of the new page patrollers
20 February 2012

WMF strikes down enwiki consensus on autoconfirmation trial
26 September 2011

Newbies and patrollers: "Every now and then a nun or a tourist wanders in front of the rifle sights"
28 February 2011

Articles for Creation
14 February 2011

Wikipedians test the water at new page patrol
9 November 2009

News and notes: Page patrolling
19 November 2007

Article creation restricted to logged-in editors
5 December 2005

In this issue
+ Add a comment

Discuss this story

Well, this is very discouraging to read. I know the project has to prioritize some projects over others but it's the patronizing attitude and the ignoring legitimate inquiries that are the most annoying. The problem is that there is no obligation for WMF to pay any attention to editors' concerns, I think they mainly respond because they don't want negative optics that would result to complaints that aren't met with some kind of response, even if it is pro forma. As far as the Board goes, I think they are so involved in big picture subjects that improving tools that some editors rely upon to do their work doesn't rank up there in what gets their attention. But, I've found in similar situations that persistence in asking questions until one gets a truthful answer can bring some results, even if one doesn't like the answer one gets. Liz Read! Talk! 00:52, 1 October 2022 (UTC)[reply]

Not to dispute the larger point that WMF may have trouble prioritizing high-impact work, but what are some concrete examples for the "Software improvements, both bug reports and feature requests, that would help and encourage reviewers do this job [but] have been languishing unaddressed at Phabricator for months and even years"? "Show, don't tell" might also work better in convincing board members, and enable them to ask more pertinent questions in their conversations with the WMF executives that they oversee. Regards, HaeB (talk) 01:59, 1 October 2022 (UTC)[reply]

Hi HaeB, this is a newspaper article about the Open Letter campaign. The WMF is fully aware of the bugs and features that need addressing, and have been for a long time. This article addresses the fact that the board stated publicly that they are not involved in worflows of individual projects. You can watch the linked video yourself or read the transcript provided in the article. Kudpung กุดผึ้ง (talk) 04:03, 1 October 2022 (UTC)[reply]
The worst bug is that it doesn't always create AFD pages. phab:T238025 Another bad one before I fixed it recently was prod tagging was broken; the reason was always blank. These bugs are so bad that many patrollers have to use Twinkle. There are also over 100 tickets open on phabricator. –Novem Linguae (talk) 07:21, 1 October 2022 (UTC)[reply]
Before you fixed it? Hope you billed the WMF for it - are you aware of the salaries and perks their devs are on? Kudpung กุดผึ้ง (talk) 07:58, 1 October 2022 (UTC)[reply]
HaeB, I had to smile at this attempt to complain about the article in an edit summary: ...not all Signpost readers might be steeped in the history of ACTRIAL, they only have to look to the right at the list of articles in the Series table of contents - there are links to two Signpost articles about ACTRIAL. Kudpung กุดผึ้ง (talk) 08:04, 1 October 2022 (UTC)[reply]
The Series ToC doesn't appear in new Vector until the last section and not at all in mobile. Given we've discovered the hard way that this piece is rather sensitive to layout adjustments, what do you think of a note that this is best experienced in desktop mode and a link to force desktop like so? Rotideypoc41352 (talk · contribs) 02:49, 2 October 2022 (UTC)[reply]
Hi Rotideypoc41352, Thanks for pointing that out. Any post-publication issues are best discussed with the Editors-in-Chief or their colleagues such as Bri or Headbomb on the newsroom talk page. A large number of editors have read the article already and you were the only one to point out the fmt bug on mobile. It's certainly worth considering for future articles that have a lot of sidebar content and quotes. Kudpung กุดผึ้ง (talk) 06:34, 2 October 2022 (UTC)[reply]
I echo the sentiments of frustration, and concern. The direction the WMF has taken is reminiscent of a RGW, political soapbox for social justice, and a voice for developing countries. While it is commendable in the context of humanity, it is not what the hand that feeds the WMF is needing. We need a focused tech team to develop the tools we need to do our job. We needs stats so we can determine what areas need the most work and help us further determine the best remedies. The shovels and wheelbarrows aren't working, we need a backhoe, frontend loader and dump trucks. We need the WMF to provide the tools we need or risk the encyclopedia being buried under mountains of garbage. Atsme 💬 📧 16:09, 2 October 2022 (UTC)[reply]
You mean the sidebar headed "RELATED ARTICLES / NEW PAGES PATROL"? It’s not visible on Timeless at tablet-portrait width, either. I didn’t know it existed until I read @Rotideypoc41352's comment. ⁓ Pelagicmessages ) 03:37, 5 October 2022 (UTC)[reply]

Reply from WMF product staff (and invitation to continue the discussion!)

Hello everyone – I’m Marshall Miller; I’m a Group Product Manager at WMF, and I’m one of the people at WMF who has spent time working on and thinking about New Page Patrol on English Wikipedia. I’ve gathered up notes from the rest of the team and checked in with Selena Deckelmann (the WMF's new Chief Product & Technology Officer), so that we can respond to the open letter from the New Pages Reviewers, and this Signpost “In focus” piece. Thank you all for working hard on the difficult task of reviewing pages, and to those of you who have made improvements to the software by submitting patches, and I hope we’ll have good conversations going forward.

We definitely agree that New Pages Patrol's work is important: keeping out newly-created articles that are bad-faith, self-promotion, or simply not ready for inclusion in the encyclopedia. We can see that there's lots of support in the community for improvements to lighten the workload for New Pages Patrol's hard-working reviewers, and so we'd like to invite New Page Reviewers, and whoever else is interested in the PageTriage software and new page processes, to a meeting to talk more about the specific needs and work together to improve the process (more information at the end of this reply).

For those that may be newer to this topic, we’d like to share some information about WMF's support for the PageTriage extension:

We understand that the people who worked on the open letter would prefer ongoing dedicated resources rather than relying on the Community Wishlist. As we continue this discussion going forward, we just wanted to say that the Community Wishlist remains an opportunity. Given the success of the NPP proposal in 2019, and the strong show of support from the signatories of the open letter, it's likely that a proposal to the next Wishlist Survey in January 2023 would be successful, and would result in more improvements to PageTriage.

But thinking more broadly, (like we mentioned above) we'd like to talk further about what the group considers to be the top priorities for improving the extension. When we invest in tools and features, we want to build things that work well for as many people and communities as we can – we’re trying to think about the reviewers, the new editors, edit-a-thon organizers, and the different languages and wikis around the world. We will be inviting anyone interested to join us for a conversation where we can hear about your priorities and get perspectives from other wikis and parts of our communities that are interested in new pages work. We will work to find a convenient time in the coming weeks where we can all meet to discuss and will follow up at NPP's talk page with an invitation. Please feel free to share it with anyone who may be interested in the discussion.

We're looking forward to continuing the discussion, and also please let us know if this reply should be posted in other places! MMiller (WMF) (talk) 01:49, 1 October 2022 (UTC)[reply]

Marshall, thank you for reacting, but this is a newspaper and a place for user comments on the article. It is not the organ of the authors of the Open Letter or the NPP team. You were specifically requested by them some time ago to place any WMF comments on the letter to a designated page for any replies you wish to be considered as official. There is no guarantee that the 444 editors concerned with the Open Letter action are all readers of The Signpost. In fact when I was E-in-C I was surprised to discover just how many editors are not even aware that The Signpost exists, it was only my initiative to include a watchlist notice that brought it to the attention of a few more users. If you want to be sure that your post receives the attention it deserves, you may wish to repost it at the page you were linked to, but the NPP is really expecting some response there from some very senior staff, among whom are those who allocate the vast surplus of funds. Kudpung กุดผึ้ง (talk) 03:20, 1 October 2022 (UTC)[reply]
Now that I have read your long comment Marshall, I do want the readers to know that you and I collaborated on some PageTriage developments a few years ago and how pleased I was of the excellent experience working with you after the abuse and personal attacks we have received in the past at Bugzilla, the precursor to Phabricator. You can read about that collaboration in a Signpost article I wrote in 2018 at NPR and AfC – The Marshall Plan: an engagement, or a marriage? which credited you with that work, but at no times has PageTriage been a community initiative or development. Most scripts written since by community volunteers are workarounds for bugs or to plug holes left in the tools - effectively this has been doing the WMF engineers' work for free. As far as the Wishlist is concerned, the NPP team received a comment from Product that in the same message accused thed NPP of not availing of the Wishlist but that on the other hand hat they don't have time or money to address the NPP concerns anyway - it's all in the article above, Putting ACTRIAL back in its correct context, Mr Horn only acquiesced following a long essay from me and a community threat to end article creation through the implementation of a local filter. The results of the ensuing trial were a staggering confirmation of how completely wrong the Foundation had been in its assumptions. Let's not forget that PageTriage was developed by the Foundation as a consolation for their brutal condemnation of the massive community consensus for the first ACTRIAL attempt (at the time one of the most heavily participated RfC in Wikipedia history). There have been many software developments that the various communities have not approved of and sorry to say, but it is hardly surprising that the communities who have no other voice are sometimes obliged to recourse to radical action. The NPP community still has a blank page waiting for WMF and/or BoT response. Kudpung กุดผึ้ง (talk) 04:03, 1 October 2022 (UTC)[reply]
I will also just add, in an attempt to contextualise your comments about the Wishlist, that readers who are really interested in what went on there should read the whole discussion, including, but not only these responses to FF-11 and Jo-Jo Eumerus who opposed it:

We were not that keen to take up a slot in the community wishlist either (although the wishlist is not just for proposals that help 'all users', this is the 'Admins and Patrollers' section after all). We have been forced to come here as a last resort as we have been told in no-uncertain-terms that they won't even do bug fixes to the existing tools unless we come here. It's a crappy situation, but we don't have an alternative. As for a potential for over-proceduralization of NPP, I'd be keen to discuss in more detail elsewhere (please contact me on my talk page). Cheers, — Insertcleverphrasehere (or here) 16:28, 17 November 2018 (UTC)
[A}s can be seen, many of the various wish list requests do not address all users. However, the impact of New Page Patrolling/Reviewing affects all new users who wish to create new articles. The current level of 'proceduralization' has never changed since NPP was introduced at the very beginning of the Wikipedia. These requsts do not over-proceduralize the process, nor do they even add any more layers of bureaucracy to it. They simply improve the work flow for the users who do this thankless task and who are unable to keep up with the flood of junk that will fill the encyclopedia if they don't. Kudpung (talk) 06:38, 18 November 2018 (UTC)

And the patrollers are still unable to keep up with a new kind of junk (very recent classic examples available, you only need to ask - or do some patrolling yourselves!) Kudpung กุดผึ้ง (talk) 04:42, 1 October 2022 (UTC)[reply]

@MMiller (WMF): I spend several hours reading through, voting on and advocating Wishlist proposals most years, but I find your references to it here to be quite bizarre. Due to the extremely limited proportion of resources that the WMF gives to Wishlist implementation, the project is largely a failure on the WMF side, and year on year you have had to water down pledges to implement the top 10 proposals, to implementing the top 5 proposals, to "no promises but we'll try to implement something, but not the ones with the most votes".
NPP can only succeed in the Wishlist at the expense of some other area of Wikimedia that desperately needs it as much, or moreso. Your suggestion is that the we fights over scraps, dividing our community in the process. My suggestion would be that the WMF stops acting like a dragon guarding a pile of gold and spends its overabundance of money on bugfixes rather than on paying its employees to write written feedback on why the WMF is declining to fix high-priority bugs. Examples of "wishes" are improvements to watchlists, new translation features and expansion of the notification functionality. A non-example of a "wish" is that the button in your software that's supposed to create an AfD page reliable does create an AfD page—this is essential maintenance. — Bilorv (talk) 18:11, 1 October 2022 (UTC)[reply]
In general, I'd love to see the community tech team, the team that does the wishlist, receive around triple the devs it currently has. And also for the wishlist to run twice a year. Tools for editors and power users, requested by editors and power users, should be a very high priority. –Novem Linguae (talk) 19:30, 1 October 2022 (UTC)[reply]
Bilorv, I think most of us on en.Wikipedia will concur with you 100% - you have summarised boldly what the authors of this article avoided spelling out so clearly for fear of retribution - indeed emails from some quarters led us to have genuine fears that the powers that be would order the E-in-C not to publish it; indeed as there had been no reference in the newsroom to the article from the editorial team, we were not sure about anything until we actually saw it in print.
The NPP people were embarrassed in 2019 at having squatted the Wishlist's bandwidth for six months where other users' requests were just as valid but did not have such a powerful voice. This is the failure of the well intended voting system. A reason why use of the the Wishlist was not heavily pursued in later years was to avoid provoking the ire of others by squatting it again - which is probably what would have happened, and partly because at the time, despite huge backlogs NPP did not have a particularly strong leadership - any de facto coordinators were caught up in RL or other Wikipedia stations that demanded more of their attention. It's only now in recent months that a couple of users have picked up the thread and decided to take action which has long been overdue.
The NPPers' main contention however, is the WMF's constant claim of not having enough cash, which everyone, absolutely everyone, knows to be total hogwash. This is why NPP is expecting a few personal words of engagement and reassurance from the CEO and Ms Denkelmann on the designated response page rather than them telling their staff to reply and here in the readers' section of a newspaper. But those at the top are probably too aloof to address us unwashed hordes, and we've seen it all before from their predecessors. It looks very much like a "Someone please do something about this rubbish from the rabble" command (but probably more like: 'Selena, can you tell Danny to tell Marshal to say something to these people'). In the meantime, the BoT is still deliberating what to say on the response page and emails from them suggest only that NPP representatives should take part in their next videoed state-of-the-nation speech. (a bit like the Wishlist all over again...). Kudpung กุดผึ้ง (talk) 01:36, 2 October 2022 (UTC)[reply]
Thank you for replying Marshall. Different managers within the same groups have very different opinions as to the importance of a process such as NPP. That's why the community looks forward to working with you as soon as the Foundation staff who have their fingers on the purse strings commit to funding a few more salaried code writers and not with employees who state these requirements should be ported through the Wishlist and in the same breath say there are not enough resources anyway. The English Wikipedia community knows full well that the Foundation's claims of being short of cash are total hogwash. Kudpung กุดผึ้ง (talk) 01:12, 4 October 2022 (UTC)[reply]


The Signpost · written by many · served by Sinepost V0.9 · 🄯 CC-BY-SA 4.0