User talk:Gnome (Bot)/Help/CleanupCriteria

"List" found in title
If the title contians "list", then replace cleanup with
 * Reasoning: Only madmen title a page list if it is not such. Therefore it should be cleaned as such.
 * Safeguards: Currently, against misidentifying articles with list as part of a word, e.g., enlist
 * Hint: search for "List of" instead of "list". — Apr. 9, '06 [01:48] <[ freakofnurxture]|[ talk]>

Article is a stub
If the article has less than  words than replace cleanup with add.
 * Reasoning: Objections have caused us to label stubs and retain the cleanup moniker. While we still feel that stubs are, by their nature, targets for cleanup and expansion, we bow to consensus and retain the cleanup tag. However, cleanup often contains stubs, and this is a simple means of identifying them.
 * Safeguards: don't add stub if it's already present.
 * Settings:
 * X = 225 words
 * Therefore, all articles marked for cleanup shorter than 225 words will be marked as stubs.
 * Hint: ignore infoboxes, tables, images when calculating number of words. — Apr. 9, '06 [01:49] <[ freakofnurxture]|[ talk]>
 * Hint: search for the regex "id\s*=\s*[\"\']stub[\"\']" and, if found, log it for manual intervention: some people are stupid and subst: stub template content directly into the article. — Apr. 9, '06 [01:58] <[ freakofnurxture]|[ talk]>

Recommendations

 * For stubs, remember that most stubs are tagged, not just.
 * For wikification, User:Bluebot adds wikify tags in some circumstances. It might be a good idea to use uniform rules between the two bots.  Also, if one bot is patrolling a certain set of articles, there's probably no need for the other to do the same thing.
 * I don't think average word length is a good indication of whether an expert needs to look at an article. I think this tag should be applied by humans.  Usually this happens when someone is doing cleanup and find they don't have the expertise to complete the job, or if they think that the next random person to come along probably won't.  This depends not so much on what is in the article, but what is missing, which contemporary bots really can't know.

Personally, I wouldn't worry too much about adding tags to articles in the cleanup queue, especially since now it's been agreed that the bot will not remove tags. If you ask me, just slapping "cleanup" on an article that needs to be wikified is enough. Whoever does the cleanup will likely start with wikification - or perhaps the cleanup will make wikification unnecessary (for example, if much of the content needs to be deleted).

It is, however, very useful to identify problem articles that are not yet in any maintenance queue. Articles that aren't tagged will wait a lot longer before being fixed, and tags act as a signal to readers not to trust those articles, and that the Wikipedia community knows they are sub-standard. I know Bluebot was patrolling Recent Changes; I don't know if it still is. Someone should also probably do a scan on an offline database dump, to catch older problem articles, if that hasn't been done already. If it has been done, it should probably be documented on the appropriate maintenance queue description page, to avoid future duplication of effort. I have posted a request for more information on User talk:Bluebot.

As useful as article identification is, it will not shrink the cleanup queue. The biggest mess right now is the static list of articles on Cleanup and previously archived months. Many of these are already fixed. An interactive bot that operated under continuous human supervision might help quickly purge this backlog. A useful thing for it to do would be to check to see if the cleanup tag (or derivative) had already been applied to the article, and whether or not it had been removed. If it's been added and removed, then it's likely that the article is fixed. The human should check to make sure any specific concerns mentioned in the listing have been dealt with. For articles that have not been fixed, it would be nice if a bot or popup system were available to move the comment to the article's talk page in a single click, and add the cleanup-date tag (if not already present) with a month corresponding to the month of the original listing (which is almost certainly not the current month). But I have also seen dedicated editors clear entire months with manual editing. It may be that the time spent making a popup framework to do this would be better spent on actually de-listing articles. Or maybe it really would make this work a lot faster, I'm not sure.

-- Beland 01:11, 15 April 2006 (UTC)