User:Alba/Workspace/Cleanup sorting proposal

'Cleanup is broken, and Cleanup process is too slow. The backlog of articles piling up on is threatening the quality of Wikipedia as a whole. To solve this problem, it is proposed that cleanup articles be sent to relevant WikiProjects by a process similar to WikiProject Stub sorting.'

Urgent nature of the crisis


The graph at right starkly illustrates the problem. From May, 2005, to March, 2006, Wikipedia as a whole has grown only about two-fold (blue line). But in the same time period, the number of articles tagged for cleanup per month has grown sixfold (purple line). If we were clearing six articles from cleanup for every article created, this would be fine; but such is not the case. As a result, the cleanup backlog has grown at an faster rate: Cleanup has expanded twenty-five fold since May, 2005 (yellow line)!

The implication is clear: if we do not take action to improve the Cleanup process, the quality of Wikipedia will steadily degrade.

To clear the backlog, Cleanup must work faster than the addition of new articles. So far, the Wikipedia community has failed to achieve a fast Cleanup turnover, despite the attention lavished on the problem.

Previous attempts
The Cleanup disaster is not a new issue, and previous moves have been made:


 * Specialized tags such as, , and exist to classify Cleanup articles by problem. In fact, an entire constellation of specialized tags has been created at Cleanup resources.


 * One early system for dealing with cleanup was the Cleanup Taskforce. Under this model, an article would be assigned to a taskforce member for cleanup. As of today, the system appears defunct.


 * When the backlog began to be noticed, User:Beland directed his bot, User:Pearle, to sort the backlog of Cleanup into . This highlighted the problem but did little to solve it. Requests continued to pour in at a higher rate than interested parties could fix.

Previous discussion of topic-specific cleanup

 * In August 2005, User:Visviva suggested topic-specific sorting of Cleanup articles. While the idea received substantial support, discussion died around October for lack of implementation. The visibility of the Cleanup project as a whole, as well as its growing crisis, was a major problem, but most Cleanup Wikipedians continued to devote most of their efforts to holding the dike against the oncoming flood of cleanup requests.


 * Other historical discussion of the problem may be found at Wikipedia talk:Cleanup, User talk:Jekoko/Cleanup links, Wikipedia talk:Pages needing attention.


 * A currently active idea is to sort articles by month, as Cleanup is currently sorted. See Category talk:Articles that need to be wikified.

Difficult to automate
As desirable as it would be, Cleanup cannot be completely automated. Cleanup, taken as a whole, requires many decisions that are content-specific, such as wikification, bolding of the main topic, grammar and style editing, and clarification. Additionally, since most articles sent to Cleanup are poorly formatted or unformatted, most of the handles used by bots to sort articles into categories are not present. Some classifications can and should be bot-automated. I and User:Eagle 101 are working on one bot to automate some sorting (wikification needs, images for cleanup, etc.). User:Bluemoose has a Bluebot that does some similar tasks. However, the possible applications are few. (See Wikipedia talk:Cleanup for what we think can be automated. If we've missed something, let us know.) Other tasks can be machine-aided with tools such as AutoWikiBrowser, if only we can get enough AWB-assisted editors engaged in the Cleanup process.

Proposed: a clear workflow for article improvement
There already exists one set of pages that address topic-specific editing needs, namely Pages needing attention. These pages are also showing a backlog, and are currently maintained haphazardly. Meanwhile, we have active WikiProjects whose primary goal is to improve articles in active collaboration on a single topic.

I propose that we improve the flow of articles from a vague basket marked "Cleanup" to the editors who are most interested in improving those articles:

Cleanup -> Attention -> WikiProject

 * Wikiproject Cleanup sorting:The purpose of Cleanup would be refocused from trying to fix the articles to sorting the articles. In a system similar to WikiProject Stub sorting and perhaps associated with it, Cleanup would sort articles by need (wikify, redlink removal, copyediting, etc.) and by topic (New York City, say, or Environment, or Star Wars).


 * New role for WP:PNA: The articles from Cleanup would be sorted into the appropriate page on Pages needing attention. Attention would become more active and better maintained, as it would become a regular part of the improvement process. Attention would become one of the primary focus pages for Wikipedians seeking to improve, rather than expand, the content of Wikipedia.


 * Wikiprojects adopt an attention page: It would be a priority to associate Attention pages with their corresponding Wikiproject, if any. The incoming flow of pages from Cleanup to their own Attention page, maintained not just by Wikiproject volunteers but by editors across Wikipedia, would give a rich field of material for individual Wikiprojects to work on, and encourage them (as their contributions are desperately needed).

Expert attention

 * Flow from WP:PNA to : Once a page is well wikified and edited, if it still needs more material, it would graduate to an  tag. This would be point at which outside sources (and even outside collaborators!), as well as expert Wikipedians, would be called in to gin up the article.

Outside review

 * Aim for quality status: Finally, it is suggested that the notional end of the pipeline be a submission to Featured article candidates, Peer review, or other content review systems. While this might be a high bar for articles coming off of Cleanup, it is the ultimate aim of Wikipedia: to have an entire encyclopedia produced to FA/peer-reivewed status. Sure, a lot of articles won't get there, but it's important to remember what the goal is.

Overall
The general idea is to harmonize the many, many projects we have devoted to Wikipedia improvement. Cleanup, Attention, WikiProjects, and other efforts are all effective in their own small way. However, because there is no clear system for deciding who can best work on what, many articles get left sitting for months in backlogs.

For deleting articles, we have a clear pathway for articles to take: Speedy deletion or Proposed deletion, which if contested goes to Articles for deletion, and can be appealed at Deletion review. Likewise, for dispute resolution, we have a clear pathway from discussion and warnings, through Requests for comment and Requests for mediation, to the ultimate Requests for arbitration. Shouldn't we have a similar, clearly defined pathway for article improvement?

Submitted for discussion by Alba 06:38, 18 March 2006 (UTC)

But won't this flood WP:PNA with poor and unwikified articles?
That's not a bug, that's a feature. Key to this proposal is the idea that people interested in a topic will be more likely to improve it than someone working only on a vaguely defined "cleanup", or on "grammar" or "copyediting", which are tedious tasks when you don't care what the text is about. WikiProjects exist for these issues, too; we should direct work on them to people who care about them.

Under this proposal, Attention is split into two levels, with Attention being less refined than Expert Attention. I wouldn't be ashamed to see an article tagged both and, but would be appalled to see  and  together.Alba 06:38, 18 March 2006 (UTC)

How dare you try to impose a straightjacket on my editing habits!
Well, if you don't like my idea, I suggest you be bold and ignore all rules. Some people will anyway! This proposal, though, is meant more to catch all the articles you aren't just all fired-up to fix.

With apologies to MasterCard and the BJAODN folks: ''There are some article fixes Wikipedia can't buy. For everything else, there's Cleanup and Pages needing attention.'' Alba 04:32, 20 March 2006 (UTC)