Grants talk:IdeaLab/Fast and slow new article review

From Meta, a Wikimedia project coordination wiki
Jump to navigation Jump to search


The top priorities for deletion are attack pages, then probably vandalism and copyright violation. Spam and articles that don't credibly assert importance or significance are not quite as high a priority. WereSpielChequers (talk) 20:17, 2 March 2016 (UTC)

Attack and vandalism can probably be detected using a supervised machine learning strategy like the one I was imagining for spam. I don't see why copyright violation detection needs to be fast given the affordances of the DMCA. Is there some reason we couldn't put potential copyright violations in the slow queue? --EpochFail (talk) 14:58, 4 March 2016 (UTC)
I'm not sure about the logic of prioritising copyvio, and it isn't something I volunteer to get involved in. But I'm aware from multiple attempts to change newpage patrol that there is a lobby who think it should be one of our highest priorities to combat. WereSpielChequers (talk) 15:09, 7 March 2016 (UTC)
The reason for not prioritizing it is that it is non-trivial to check in most cases. Whether a G13 speedy can be safely deleted or whether there's an option should not be based on a quick glance. DGG (talk) 02:46, 19 March 2016 (UTC)


If the aim is to make things easier for the newbie creating a new article, it would probably be easier and less contentious to fix some of the many bugzilla and phabricator requests to reduce edit conflicts. As it is a newbie creating a new page risks losing edit conflicts not just to deletion and maintenance taggers but even to categorisers. WereSpielChequers (talk) 20:17, 2 March 2016 (UTC)

We don't need to choose here. There's lots of beneficial work that can be done. IMO, it's much easier to stand up a new machine learned AI model and experiment with workflows than it is to make a robust, tractable solution to edit conflicts. Given my expertise, I'm the right person to help with this workflow AI, but not the right person to help with addressing edit conflicts. Maybe you could submit another idea proposal about taking on edit conflicts. I'd support it. --EpochFail (talk) 15:04, 4 March 2016 (UTC)
Fair point, yes it should be possible to do both, just don't underestimate how big a problem edit conflicts are to creators of new articles. I've raised multiple Bugzilla requests in the past, but could try to raise this here. WereSpielChequers (talk) 15:09, 7 March 2016 (UTC)


It is a common mistake among new speedy deletion taggers to think that A7 means no assertion of notability. The actual test is no credible assertion of importance or significance. Training new speedy deleters that notability is a criterion at AFD not at speedy deletion is one of the unending tasks at New Page Patrol. Any suggestions as to how we can do that before taggers have tagged multiple articles for speedy deletion because in their opinion there would have been consesus to delete them at AFD? WereSpielChequers (talk) 20:29, 2 March 2016 (UTC)

Note, posted at New Page Patrol[edit]

See en:Wikipedia_talk:New_pages_patrol#Proposed_tool_to_help_you_all Jytdog (talk) 03:41, 15 March 2016 (UTC)

Wrong approach[edit]

Thanks for the suggestion! I agree with the problem statement, but rather than teach a machine to evaluate if an article should be put on the "fast track" or "slow track" for review we need to restrict NPP to humans with this (very basic) level of competence and self control, potentially using a user right. This would also help protect against certain kinds of circumventions of the patrol mechanics ie this. VQuakr (talk) 04:17, 15 March 2016 (UTC)

VQuakr You think it is not even worth trying? That is what this is for, to see if such a thing could be usefully created, and then to find out how to appropriately use it. Jytdog (talk) 19:33, 15 March 2016 (UTC)
Jytdog given that resources are finite, it makes sense to solve problems efficiently. Any NPPer that cannot follow a community-determined rule for waiting a specified period of time before tagging for A7 will also probably lack competence to patrol for other reasons. Conversely, no machine will be 100% correct at determining if an article is subject to a waiting period, so some sort of permission-regulated override would still be needed.
The good news is that no, NPP is not overwhelmed. At the moment at least, our leading problem at NPP is coaching newbies who really should be learning to edit before they patrol. Letting absolutely anyone patrol contributes to problems with editor retention, copyvios/BLP vios/spam slipping through into mainspace, and vulnerability to bad faith attacks. Creation of a user right (or repurposing of the nigh-obsolete rollback right) to allow community control over who uses the curation tools is such an easy partial fix that IMHO it should be enacted immediately, even if just as a stopgap while the more sweeping improvements Kudpung mentions below are developed. VQuakr (talk) 08:21, 16 March 2016 (UTC)
  • I'm not sure if I've correctly understood what's being proposed but any allocation of a grant to investigate the possibility of and/or develop any automated review process for NPP would be a gross misallocation of funds and a typical Foundation belief that everything can be automated. The previous research that has been carried into NPP over the years, under whatever guise, never addressed the real problem, and the results that were extracted were spun by the Foundation to justify work and salaries to keep their 'engineers' in employ. Fortunately, some of those who were involved at the time have now left and we're still waiting for the Foundation to address the bugs in the current Page Curation suite. This is one of those areas where the ideas for solutions MUST come from the volunteer community, and the WMF should simply develop and correctly maintain the software and servers that are asked for. That the WMF clearly wants to interfere with article quality has been demonstrated by the way they constantly nip any community initiatives in the bud, such as for example proposed presentations on the subject at Wikimania.Kudpung (talk) 19:56, 15 March 2016 (UTC)
Kudpung I apologize for having blundered into something where there is a history of tension between WMF and at least part of the editing community. Could you please say a bit more about a) how a collaboratively developed tool to review new articles would be harmful, and b) say a bit more about what is wrong with the current tools? Also, if there are tools you want, would a proposal here be a way to get that done? I would be happy to review and support something else. I supported this because my sense is that the people who work NPP do have an overwhelming about of work, and tools to flag articles might be helpful to you. I am sorry if I was wrong about that. Jytdog (talk) 23:00, 15 March 2016 (UTC)
Pull up a chair, Jytdog, pour yourself a cup of tea, this is going to be long but hopefully not TL;DR. Several years ago, a small group of editors and admins including WereSpielChequers, Scottywong, me, and a couple of others were concerned with what was a huge backlog at NPP (really huge, in the tens of thousands) so after slaving day and night over a hot stove for a couple of weeks we were able to get it down to a mere 10,000 or so and then we set about thinking what could be done about it. Knowing full well that there was (and still is) little interest in this critical maintenance task which is nevertheless absolutely crucial, and knowing that the vast majority of new creations were fit for the trash can, we took the bull by the horns and proposed to restrict the creation of new articles in live mainspace to autoconfirmed accounts only (4 days and 10 edits) in the anticipation that this would stem the tide of spontaneously created rubbish and other disallowed content. Heralded by the community as a great idea, the RfC was subscribed by several hundred participants and consensus was carried by an overwhelming majority to carry out a six-month trial (to appease the few critics), known as en:WP:ACTRIAL. Because this minor change in user privileges required an equally minor operation on the MediaWiki software to which we don't have access, the required intervention was requested at Bugzilla.
The 'engineer' reading the request immediately (and very rudely) rejected this community request on idelogical grounds, which although demonstrated later to be shaky, had reached the notice of Foundation middle management who immediately supported the decision of one of their own salaried employees and hence widening the already growing rift between the volunteer community and the Foundation. Nevertheless, something still needed to be done urgently about the huge flow of inappropriate new content and NPP's inability to cope with it. After lengthy live discussions with en:Eric Moeller and Brandon Harris, at that time the two most senior staff below the CEO, I was able to get them to offer us an olive branch in the shape of the suite of Curation Tools (for which a junior staff contractor promptly took all the credit on behalf of the Foundation, see Wikipedia:New pages patrol/Survey and Research:New Page Patrol survey), and the promise of a completely new welcome page/article wizard.
The truly awesome Page Curation and its New Pages feed and video tutorial arrived and we were left waiting for Brandon's new user landing page, code named Article Creation Flow, to materialise. It was thought that the effect of the two solutions working together would greatly contribute not only to stemming the tide of incoming rubbish but also greatly enhance and speed up the work of the New Page Patrollers, whom it was quite wrongly rumoured (another WMF myth) were suffering from burn-out. Brandon had begun work on his new landing page idea and the first drafts were indeed highly encouraging, and were far better than my earlier attempts to redesign and rescript the Article Wizard. In short, while not making potential new article creators wait 4 days and 10 edits before their work would be published, their articles would be forced through the Wizard and reviewed very quickly at NPP (or, as it turns out later, but unconnected, partly by en:WP:AfC). It quickly became evident however that NPP still proved to be a challenging task to all but the most experienced users and admins who occasionally did a stint at it. New users were still being badly bitten while the creations of spammers and paid editors were gething through with impunity
Five (or is it six already?) years later, due to its needing absolutely no knowledge or experience, NPP remains a magnet to younger and/or inexperienced users who treat the world's largest knowledge base as a MMORPG in the full knowledge that even the millions of run-of-the-mill blogs and forums exercise stricter control over what gets posted. A quick look at my contribs will show that I have asked 100s of 100-edit wonders to please, please, please first get the mandatory 500 edits and get their experience doing something far simpler such as combating vandalism until they have grasped the fundamentals of content control. I've even had to have some users topic banned from NPP or even blocked to ensure that they refrain from causing more damage.
We as a community have to weigh the balance of what is perceived as a dwindling number of new articles against the dangers of spam, hoaxes, copyright infringements, and libelous attack pages being allowed to slip through by well-meaning but thoroughly inept patrollers with a lust for control over Wikipedia pages, and decent new contributors being bitten and scared away by patrollers who lack the aptitude for educated communication because there is still no landing page telling new creators up front what they can and what they can't write.
The underlying cause for the rejection by the WMF of new controls and the way Brandon felt he had to sweep his Article Creation Flow project quietly under the carpet, is the fact that the Foundation is only concerned with the raw statistics of newly created pages, irrespective of what is on them, in order to maintain its socio-political position in today's world of technology and exchange of knowledge and information. Most of these people have now left WikiMedia in the turmoil that ensued since Gardner's retirement and which has peaked with the sudden departure issue of new page control of her successor and the collapse of the Board of Trustees. Now is the time of course to do something about this critical, cross-Wiki issue.
In my opinion, the only way to improve the state of new content is to once more take a broad look at en:WP:AfC, en:WP:NPP, en:Article wizard, en:ACTRIAL, properly grant some funds for the actual codewriting of what we the community needs, and quickly, but very quickly introduce a technically controlled user group to ensure that the Page Curation system is only in the hands of editors who are fully conversant with our NPP, COI, SPAM, and DELETION policies. DGG and I and a couple of others are currently working on it, and I hope to facilitate a discussion about it at the upcoming Wikimania. The junior contractor's constant bleating that the Foundation was floundering on a bootstring budget not only rang hollow then, but would be an absurd statement today. He was installed by the Foundation deliberately to usurp any community initiatives and to stonewall any complaints.
Perhaps now that some of these people have gone we can get some progress in spite of the loss with them of their instutional memory.Kudpung (talk) 07:12, 16 March 2016 (UTC)
Since I've been pinged on this I would like to point out that my views on the solution for the newpage patrol problems don't always coincide with Kudpung's (I've just reread en:Wikipedia:Village_pump_(proposals)/Proposal to_require_autoconfirmed_status in_order_to_create articles#View from_User:WereSpielChequers and still agree with that). In particular I was not part of the consensus for ACTRIAL, though I think the way the WMF overruled that consensus was crass and inept, possibly as damaging in a very different way as the proposal. WereSpielChequers (talk) 07:55, 16 March 2016 (UTC)
At a glance, at least some of the ideas in Kudpung's post, such as waiting before A7 nominations, are not mutually exclusive with the VP/P section you link. VQuakr (talk) 08:21, 16 March 2016 (UTC)
Well, I read all that. And wow. That's all I have say. And maybe. Terrible. and. Yes I see why you find this grant proposal ... difficult. And also. Yes, now!. If there is anything I can do from my end (I work more on dealing with COI in existing articles) please let me know. Jytdog (talk) 08:25, 16 March 2016 (UTC)

April 12 Proposal Deadline: Is your project ready for funding?[edit]

The deadline for Individual Engagement Grant (IEG) submissions this round is April 12th, 2016. If you’ve developed your idea into a project that would benefit from funding, consider applying!

To apply, you must (1) create a draft request using the “Expand into an Individual Engagement Grant” button on your idea page, (2) complete the proposal entirely, filling in all empty fields, and (3) change the status from "draft" to "proposed." As soon as you’re ready, you should begin to invite any communities affected by your project to provide feedback on your proposal talk page.

If you have any questions about IEG or would like support in developing your proposal, we're hosting a few proposal help sessions this month in Google Hangouts:

I'm also happy to set up an individual session. With thanks, I JethroBT (WMF) 00:38, 2 April 2016 (UTC)