Wikipedia:Bots/Requests for approval/ClueBot VI

The following discussion is an archived debate. Please do not modify it. Subsequent comments should be made in a new section. The result of the discussion was

Approved.

ClueBot VI

Operator: -- Cobi^(t|c|b)

Automatic or Manually Assisted: Automatic, unsupervised.

Programming Language(s): PHP, my classes.

Function Summary: Clerk Wikipedia:CHUU.

Edit period(s) (e.g. Continuous, daily, one time run): Continuous

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

Function Details: Clerk for Wikipedia:CHUU. User:ClueBot VI/CHU/U is an example run where it grabbed a live copy of Wikipedia:CHUU, parsed it, added the relevant data, and posted the result to userspace.

Discussion

Approved for trial (25 edits). Please provide a link to the relevant contributions and/or diffs when the trial is complete. —Ree dy 10:30, 28 May 2008 (UTC)[reply]

The bot already has problems. On this,this and this request, the bot says the account was created within six months; this is not true. On this request, it does not make note of the number of edits the requesting editor has. (While SUL sometimes makes this irrelevant, there was no such claim here, and after the SUL influx, such notes will still be needed.) The bot makes note that the target account has not been notified; this isn't particularly helpful — people usually do it for them. Here, there are edits of which the bot did not make note, and made note of nonexistent deleted edits (How, by the way, can the bot see these?). It will revert war [1] [2] after irrelevant notices are removed. Example number two. (The CHU bot did this as well.) While I see a bot for CHU as helpful, I'm not so sure it is at CHU/U. There are many more aspects of requests at CHU/U than at CHU, and I don't think a bot can really do everything needed. It will still require humans to examine the requests, which defeats the purpose. seresin ( ¡? ) 17:36, 28 May 2008 (UTC)[reply]

Well, trials are used to find problems with the bot, so thank you for reporting them. However this and this say that the users were created recently. I can make it make a notation of edits, but I didn't see a place in {{CUU}} to do so. The {{CUU}} template has a place for a |notified= parameter. ClueBot VI just checks to see if a signed note by the person asking to usurp was left on the talk page of the account they wish to usurp. I can make it totally ignore this, if you want. About the edits that the bot didn't notice, that was a bug which I've now fixed :) It detects deleted edits by counting the total edits (including deleted edits), then counting the visible edits, then taking the difference. As for the edit warring, I didn't think that someone would remove the CUU template without leaving another, correct one. If there is a {{CUU}} template or a {{Clerk note}} template on the thread, the bot skips over it, otherwise it tries to add its info to it. All this bot does is lookup the information to go in the parameters of {{CUU}}. I am quite open to suggestions and modifications to the way it works. -- Cobi^(t|c|b) 19:14, 28 May 2008 (UTC)[reply]

Well, then there's a problem with the software that needs addressing, as there is no user creation log for those two accounts, which usually means they were created before the logs existed. The template is not the end all be all; other notes are sometimes required (hence my comment about a bot not being able to fully satisfy the role), and number of edits is often one. As for the note on the talk page, it's probably better to make note of it that not, it's just another thing the bot does not do and therefore requires a human. There is not always need for a template, as in the two cases where I removed it. seresin ( ¡? ) 05:11, 29 May 2008 (UTC)[reply]

I have a few observations on this bot proposal.

"The template is not the end all be all; other notes are sometimes required": I think the idea is that, if further, non-routine clerk observations require to be made, a "human clerk" (or indeed, a bureaucrat) could very easily append the notes to the end of ClueBot's version. That a bot clerk is active on a particular request section, does not shut out additional clerk notes from the existing staff of editors providing assistance.

On a general note, I support the introduction of this robot into the changing username process, on the condition that the current creases are ironed out (namely, those pointed out above). The "clerk bot" will drastically reduce the requirement for Wikipedia editors to spend time annotating requests on a fairly low-key process; I understand that a number of current clerks fairly enjoy their work, but there is always a human redundancy with the introduction of IT into a process. :)

As a specific observation, my impression of ClueBot thus far has been very good. The recent influx of requests on account usurpations and changing username has seen a tenfold increase in work load for the existing clerk staff; the robot has bolstered the smooth running of the requests system, and indeed, avoided what could very well have been a small insight into disorganised chaos. :) The task–annotating requests with almost-always identical information fields–requested seems to be ideal for a robot, and indeed, the ClueBots already handle tasks that are much more difficult.

I do hope this robot proposal is eventually accepted: overall, my impressions have been very good.

Anthøny 19:35, 29 May 2008 (UTC)[reply]

Damn bots, replacing us humans wherever they can! Seriously though, I think this is a fantastic idea, especially with the wave of SUL-related requests we're getting. Hope this succeeds, will be a great help once the bugs are fixed. RichardΩ612 Ɣ ɸ 19:47, May 29, 2008 (UTC)

@AGK with regard to the template. The point is that in order to become aware of other notes needed to be made, a human will still need to evaluate the request. Therefore, the bot is not really saving us much effort, because we still need to check the request. There are just many more variables that need to be examined in /U requests, and I do not believe a bot is an effective tool. As a side note, I presume the bot sends the e-mail when it notes that "an email has been sent by a clerk or bureaucrat to notify them of this request."? seresin ( ¡? ) 23:38, 29 May 2008 (UTC)[reply]

I'm afraid I'm not following. The bot is essentially doing half of the job for you, and yet, because there's a little more to do, the bot is not helping? The maths is pretty straight: assuming the bot eliminates 50% of the workload, there's 50% left... but that's still a 50% reduction in workload. How then, can it be said that the bot is not an effective tool? Anthøny 23:51, 29 May 2008 (UTC)[reply]

The bot does not ever fill the email parameter with "sent", only with "yes" or "no". So the bot should never say "an email has been sent by a clerk or bureaucrat to notify them of this request." If it did, then someone modified the bot's comment. -- Cobi^(t|c|b) 03:02, 30 May 2008 (UTC)[reply]

See here for an example of someone changing the bot's comment. -- Cobi^(t|c|b) 03:06, 30 May 2008 (UTC)[reply]

Done - Can I get approval or another trial? -- Cobi^(t|c|b) 20:43, 30 May 2008 (UTC)[reply]

Approved for trial (25 edits). Please provide a link to the relevant contributions and/or diffs when the trial is complete. Note I cannot give a full approval, due to me having a partial COI with User:SoxBot VI. Soxred 93 21:04, 30 May 2008 (UTC)[reply]

Agree with the decision for another trial: just to establish that the minor niggles with the specifics of how the bot's functions are ironed out. Anthøny 21:53, 30 May 2008 (UTC)[reply]

Done - Can I get approval? -- Cobi^(t|c|b) 12:15, 3 June 2008 (UTC)[reply]

What happened here and especially here? Other than those, I don't see a problem approving. SQL ^{Query me!} 07:27, 4 June 2008 (UTC)[reply]

Well, the parse errors occurred because someone *cough*WJBscribe*cough* decided it was a good idea to change the headers from using ====Target ← Requester==== to ====Requester → Target====. Since the bot then saw a malformed request, it noted so. I fixed the code shortly after. As for the second diff, where it changed all the headers, it just trim()'d the headers for consistency. -- Cobi^(t|c|b) 10:07, 4 June 2008 (UTC)[reply]

Seems like an extremely helpful bot to me.

Approved. SQL ^{Query me!} 03:06, 5 June 2008 (UTC)[reply]

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section.