Wikipedia:Bots/Requests for approval/ChzzBot IV


 * The following discussion is an archived debate. Please do not modify it. To request review of this BRFA, please start a new section at WT:BRFA. The result of the discussion was Symbol keep vote.svg Approved

ChzzBot IV
Operator:

Time filed: 17:27, Monday May 16, 2011 (UTC)

Automatic or Manual: Automatic (supervised for now, at least - will become unsupervised)

Programming language(s): c#

Source code available: No

Function overview: For users creating AFC submissions with no inline references at all, it adds a short note explaining that it'd be helpful if they added some - soon after they submit. Their submission is still reviewed in the normal manner.

Links to relevant discussions (where appropriate):


 * Log of manual addition of help, showing review of results, and some input from other user tweaking the idea: User:Chzz/afc_unreferenced_log
 * Project discussion: Wikipedia_talk:AFC
 * Discussion re. DAB: User_talk:France3470
 * Now mentioned on the Village Pump (as advised): Village_pump_(miscellaneous)

Edit period(s): When approved, it would run all the time (or, at least, every 5-10 mins or something)

Estimated number of pages affected:

Could be circa 30 user pages per day, plus each of those AFC's

Exclusion compliant (Y/N): Yes - skips any AFC or user talk which have bots excluded

Already has a bot flag (Y/N): N

Function details:

For every entry in Category:Pending AfC submissions,
 * If it appears to be a submission - ie, it contains " {{afc submission") and the page name begins with "Wikipedia talk:Articles for creation/", and
 * If the page has no references at all (ie no "&lt;ref" case-insensitive), no "{{harv", no "{{sfn" and
 * If it's not a disambig - ie doesn't contain "{disamb") or "{{dab" or "{{dbig" or the title contains "disambig", THEN
 * IF the creator has no user talk page at all, it adds a short 'welcome' -
 * For IP users, it adds User:ChzzBot IV/Welcome anon,
 * For registered users, it adds User:ChzzBot IV/Welcome user
 * Then it adds a message about the AFC needing a reference; User:ChzzBot IV/AFC unreferenced
 * It also adds a note to the AFC, and logs the actions to User:ChzzBot_IV/log.

This has already been tried by hand, and a few tweaks made, e.g. skipping DAB sumbissions. It's been shown to help in some cases, and does not seem to do any harm in any way. In particular, a timely message helps - the AFC may not be reviewed for several days, but whilst the user has submitted (within minutes), if they get a message, they do sometimes improve the submission.

It's also bot exclusion compliant, so skips any AFC/user talks with bots denied.

Discussion

 * You should broaden your regex for "unreferenced", I think, given that numerous refering style that do not make use of ref tags are valid (if unusual in many topic areas) forms of referencing around here. Someone must already have written one though to deal with other bot requests. Worth looking into to save you some time. - Jarry1250 [Weasel? Discuss.] 17:31, 16 May 2011 (UTC)
 * Agreed. In addition, a better method for checking if the page is a dab, may be to use an API query like this, rather than searching the text (although continuing to check the title also seems to be a good idea). Looks like there's a fair amount of support at WT:AFC for something like this, would like to see what (if anything) the wider community has to say, so we'll see if the pump discussion turns anything up. - Kingpin13 (talk) 17:39, 16 May 2011 (UTC)


 * Note that it's possible to reference an article without any <ref markup, like sfn and reflist, though I do realize how unlikely that is for AfC. — HELL KNOWZ  ▎TALK 17:43, 16 May 2011 (UTC)
 * Fair point. I will add a skip for any that contain either "{{harv" or "{{sfn". The submissions do almost always have a {{tn|reflist}}, but commonly with no actual refs. Hopefully that covers that (unlikely) case?  Chzz  ► 17:45, 16 May 2011 (UTC)
 * I do believe it would. — HELL KNOWZ  ▎TALK 17:53, 16 May 2011 (UTC)
 * I've added that, anyway. And (if OK'd) will continue to do as much checking as I can, to pick up other 'exception' cases and refine the conditions for any other oddments.  Chzz  ► 17:54, 16 May 2011 (UTC)


 * Great idea, but I don't think you have the DAB issue covered, not all DABs are named in the format "Topic (disambiguation)" If the topic being disambiguated doesn't already exist, it would be correct to just use the name of the topic. Could it check for other signs of a disambiguation page, such as the "may refer to" phrase, or maybe even the characteristic bullet points with one wiki link each? (I don't know how hard it would be to do the latter) Monty  845  17:49, 16 May 2011 (UTC)
 * It is set to disregard pages containing "{disamb") or "{{dab" or "{{dbig" or if the title of the page contains "disambig". I hope that will cover almost all cases?  Chzz  ► 17:51, 16 May 2011 (UTC)
 * Would it detect Catskill? Granted I think it is unlikely that many such pages are going to come through AfC. Monty  845  17:57, 16 May 2011 (UTC)
 * The bot can have a long list of all the different dab type templates. More relevantly is what to do if there is no such template and the title does not have a "disambiguation" as you pointed above. Though, again, that's unlikely for AfC. — HELL KNOWZ  ▎TALK 18:00, 16 May 2011 (UTC)
 * I can add a further check to see if the AFC is in Category:All disambiguation pages - which should pick up AFC's with templates such as {{tn|geodis}} and others. I'll work on that now.  Chzz  ► 18:03, 16 May 2011 (UTC)
 * I should have brought that up, but I think most people relying on AFC probably wont have the template added, but then, it may be a lost cause, as they are also unlikely to have the DAB formatted correctly enough to be detectable by any automated means. I think the value of the idea is worth a few DABs getting erroneously tagged, but the fewer the better. Monty  845  18:05, 16 May 2011 (UTC)
 * Agreed; it might miss some malformed AFC's that are destined to be DAB, but few. I've added the cat check, anyway.  Chzz  ► 20:03, 16 May 2011 (UTC)

{{BotTrial|days=7}} The AFC community seems supportive of this. Let's get an idea of how this is going to work... - Kingpin13 (talk) 01:29, 20 May 2011 (UTC)

{{tlx|adminhelp|Per the above, please set {{user|ChzzBot_IV}} as 'confirmed' so it can create pages. Ta.}}  Chzz  ► 07:32, 20 May 2011 (UTC)
 * {{done}} - Kingpin{{sup|13}} (talk) 07:34, 20 May 2011 (UTC)

The bot's edit here was completely unhelpful, the submission was in a foreign language so I doubt the author even understood half of the message. Sure leaving a note about referencing is helpful but not when the submission is not written in English. Could it be programmed to do simple tasks like declining foreign language submissions? — James {{sup|(Talk • Contribs)}} • 10:47am • 00:47, 21 May 2011 (UTC)
 * I don't see how the bot could identify whether or not a submission is in English. I have now made it ignore any submissions which have "Template:" in the name - as templates do not need refs.  Chzz  ► 01:30, 21 May 2011 (UTC)
 * Good point there, I don't suppose importing the OED would be recommended then :P thanks for that though Chzz! Could the bot leave a note about primary sources if say most of the references are affiliated with the topic (eg. they contain the subject's name in the URL, John Smith's submission uses www.johnsmith.com multiple times, for example), would that be possible? Thanks again — James {{sup|(Talk • Contribs)}} • 5:52pm • 07:52, 21 May 2011 (UTC)

Just another suggestion, the message the bot posts on the talk pages of the authors is good, but shouldn't it point to WP:REFB or WP:VRS and it should probably include "Accessed on YYYY-MM-DD" and the author of the source's name. — James {{sup|(Talk • Contribs)}} • 2:57pm • 04:57, 22 May 2011 (UTC)

One other thing, references can be parenthetical too, so perhaps it should just "articles on Wikipedia should have citations, parenthetical: eg. X + Y = Z (Blah, 1994) or inline, (the rest of the current message would follow)" and reliable sources should probably be wikilinked to WP:RS. — James {{sup|(Talk • Contribs)}} • 3:00pm • 05:00, 22 May 2011 (UTC)


 * I don't think it is feasible for this bot to work out if there are too many primary sources; I think that must remain part of the human review process. Links to "johnsmith.com" could well be appropriate, and there could be lots; there could also be valid links to e.g. "newyorktimes/john_smith" or invalid links to "facebook/johnsmith" or a zillion other variations.
 * I've changed the example, to try and make it clearer and better; see User:ChzzBot IV/AFC unreferenced and User:ChzzBot IV/Example page with references.
 * I don't want to complicate things by talking of parenthetical referencing. I think it's important to keep this as simple as possible. (It might already be getting too complex). A huge number of AFC's have no refs at all; as long as they add some, that will be a massive improvement, making it easier to help get the article live where possible. Same for linking to WP:VRS or WP:RS or WP:V - those links may not be applicable at all; the AFC might have reliable sources that show notability which are just not correctly formatted. All this message is saying, really, is "please put some inline refs". It's not trying to work out the validity of the submission in other respects; that's down to the review itself.  Chzz  ► 19:17, 22 May 2011 (UTC)

After trial
During the 7-day trial, the bot placed messages on around 150 pages (log), and it will take some time to evaluate how effective it was, and if there were any issues of concern. The nature of AFC means, it will be some time before results can be seen (ie, if the authors add refs, and/or articles are approved or declined, etc). It seems sensible to hold this approval process whilst that happens.

I hope that's OK; once I (hopefully with help) have had a chance to look over the results, I will seek further guidance here. Best,  Chzz  ► 20:25, 27 May 2011 (UTC)
 * Sounds good. Any prelim results?  MBisanz  talk 23:22, 12 June 2011 (UTC)
 * Prelim results: It's quite difficult to evaluate. Even though it posted help on 150 AFCs, a huge number of AFC pages are unlikely to become encyclopaedic articles. It's further complicated by the fact that results don't show up instantly; sometimes article authors revisit their declined AFC week later (or even longer). I haven't seen any cases yet where it could be considered to have had a negative impact, apart from when it placed help re a disambiguation (where refs were not needed) - but that's been resolved now.
 * So far, 25 have been checked; 20 didn't have refs added, and were declined - mostly as 'not notable'. A couple of those probably shouldn't have been declined.
 * Five of the 25 had some kind of refs added. Four of those were declined as not meeting notability requirements; one became a live article.
 * Empiric evidence indicates that giving users prompt and appropriate guidance, as opposed to not giving it, is a Good Thing.
 * Lots of AfC's get rejected for inappropriate or spurious reasons (by inexperienced reviewers), which means we lose valuable contribs / new users.
 * See User:ChzzBot IV/log and feel free to help evaluate the results.  Chzz  ► 19:22, 13 June 2011 (UTC)
 *  MBisanz  talk 14:31, 14 July 2011 (UTC)
 * The above discussion is preserved as an archive of the debate. Please do not modify it. To request review of this BRFA, please start a new section at WT:BRFA.