Wikipedia:Bots/Requests for approval/HersfoldArbClerkBot
- 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 Approved.
Operator: Hersfold (talk · contribs)
Time filed: 20:19, Saturday July 2, 2011 (UTC)
Automatic or Manual: Automatic unsupervised
Programming language(s): Java
Source code available: User:HersfoldArbClerkBot/Source
Function overview: Tracks the length of evidence submissions in open Arbitration cases and notifies users when length exceeds limits set by the Arbitration Committee.
Links to relevant discussions (where appropriate): Discussion on clerks-l mailing list and private IRC discussion with Arbitrators; clerks and Arbitrators will be asked to comment here shortly.
Edit period(s): Expected to run once every two hours, continuously. Will not run when there are no open Arbitration cases (ha ha).
Estimated number of pages affected: Est. 5-10 per open case (1 for the evidence page (edited during almost every run), and 1 per notified user (one notice per user per case))
Exclusion compliant (Y/N): Yes, for user talk pages only
Already has a bot flag (Y/N): No, requested
Function details:
The Arbitration Committee has established limits on the length of evidence in words (500) and the number of diffs (50) that may be provided per user per case. These limits are frequently exceeded, which makes it difficult for the Committee to locate and review important and pertinent information. This results in a longer case and may result in a less effective result. HersfoldArbClerkBot is intended to enforce these limits, and depersonalize the enforcement process, by doing the following:
- Keep track of the word length, diff count, and other URL count in each evidence section
- Add hatnote-like messages to the top of each user's evidence section to raise awareness of evidence limits and assist human clerks with enforcement
- Where limits are significantly exceeded, as determined by a configurable tolerance level for each limit, issue notices to users requesting they reduce the length of their evidence (no more than one notice is issued per user per case)
- Maintain a summary of evidence length information within the bot's userspace, again to assist human clerks (this task has already been completed, a report is available at User:HersfoldArbClerkBot/Length reports)
At no time will HACB attempt to truncate or otherwise refactor evidence submissions. The operation of the bot may be to some extent configured by Arbitration Committee Clerks or Arbitrators using the bot's configuration page, where limits for word length, diff count, and URL count can be modified, as well as the tolerance levels the bot uses to determine if a user should be warned for exceeding the limits or not.
This bot is intended to standardize the process of managing evidence length in Arbitration cases and lessen the work on human clerks. I remain fully open to comments and suggestions from current and past clerks and Arbitrators, although suggestions from currently serving Arbitrators will be granted more weight due to the fact that it's their Committee. :-) Hersfold (t/a/c) 20:19, 2 July 2011 (UTC)[reply]
Discussion
edit- Confirming that the bot in this form is a result of internal discussion among the clerks and arbitrators, and that it has our support to run. AGK [•] 20:36, 2 July 2011 (UTC)[reply]
- I too confirm this, since AGK surely can't be trusted. :-P Tiptoety talk 20:43, 2 July 2011 (UTC)[reply]
Approved for trial (1 week). Please provide a link to the relevant contributions and/or diffs when the trial is complete. Or, bureaucracy aside, however much time you need to have it running stable. — HELLKNOWZ ▎TALK 20:39, 2 July 2011 (UTC)[reply]
- I don't expect too much trouble (I've been running it in "no edit mode" and haven't seen any problems) so I'll wait a few days to allow more comments from Clerks and Arbs. I'm out of town at the moment anyway, and I'd rather be physically in front of the computer it runs on (rather than remotely connecting to it) just in case something does go wrong. Thanks for the fast approval, though! Hersfold (t/a/c) 20:51, 2 July 2011 (UTC)[reply]
- Just an additional Arb-chime-in "Me too!". (otherwise known as yes, this is a fabulous idea, we've talked about it and many thanks to Hersfold for putting this together). 23:27, 2 July 2011 (UTC)
- Hersfold, any chance you could code a way for us to manually configure the bot to grant higher limits to particular users (e.g. Betacommand, MZMcBride)? NW (Talk) 00:07, 3 July 2011 (UTC)[reply]
- I'll work on adding that feature prior to the trial run. Hersfold (t/a/c) 02:43, 3 July 2011 (UTC)[reply]
- Would this detect any content that has been temporarily hatted using {{hat}}? I personally would prefer that it didn't, if possible. NW (Talk) 03:05, 3 July 2011 (UTC)[reply]
- Yes... although I need to make a slight fix in that regard. When doing the word count, the bot will not count any content that it contained within a template. However, the diff and link counts are currently done slightly differently, and at the moment would count links within a template. I can fix that, though, it's a matter of adding two lines of code.
- The override update you suggested will be online tomorrow; it's a bit more complicated than I expected, and required a fairly significant overhaul. I'll post brief instructions here and more detailed ones on the configuration page's edit notice. Hersfold (t/a/c) 03:53, 3 July 2011 (UTC)[reply]
- I've added the suggested features, and am about to test them now within the bot's userspace (this is not a full trial). The overrides can be implemented on the bot's configuration page, with the format
*OVERRIDE=user name|case name|word limit|diff limit|link limit|word tolerance|diff tolerance|link tolerance
, where either the user name or the case name can be "all
" to provide a global override for all users in a particular case or all cases for a particular user. - The word, diff, and link counts will now all ignore all content found between {{hat}} and {{hab}} templates. Hersfold (t/a/c) 01:17, 4 July 2011 (UTC)[reply]
- Sounds good. As far as the override goes, I assume it should be placed before "END CONFIG"? NW (Talk) 04:39, 4 July 2011 (UTC)[reply]
- I've added the suggested features, and am about to test them now within the bot's userspace (this is not a full trial). The overrides can be implemented on the bot's configuration page, with the format
- Would this detect any content that has been temporarily hatted using {{hat}}? I personally would prefer that it didn't, if possible. NW (Talk) 03:05, 3 July 2011 (UTC)[reply]
- I'll work on adding that feature prior to the trial run. Hersfold (t/a/c) 02:43, 3 July 2011 (UTC)[reply]
- Please. It won't solve all the associated problems, but it's a great idea and will definitely help us poor clerks. Dougweller (talk) 10:14, 3 July 2011 (UTC)[reply]
- I'm going to start a live trial run now. Because the Tree shaping case is in voting now, I've set the bot to avoid enforcing limits on that entirely; the bot will still add the templates to the evidence page, but the template won't actually display anything. A length report for the case will still be generated in userspace. Hersfold (t/a/c) 05:25, 5 July 2011 (UTC)[reply]
- First trial complete, however there is one problem. The bot is not recording information in the warning log correctly, so if I were to run it again right now, Rd232 and Chester Markel would each get a second notice about the length of their evidence. Everything else appears to be working correctly, however. I'll try to fix this bug then run another test tomorrow. The logs from this run will be available at User:HersfoldArbClerkBot/Trial logs. Hersfold (t/a/c) 05:33, 5 July 2011 (UTC)[reply]
- Found the bug, I hadn't set up an output stream correctly. I'm going to delete the warning log, roll back the bot's previous notices, and run it again. Hopefully nobody will notice; not likely, since two are blocked right now. ;-) Hersfold (t/a/c) 05:50, 5 July 2011 (UTC)[reply]
- That trial has been terminated, because the bot is adding duplicate templates to the evidence pages. Going to fix that now... Hersfold (t/a/c) 05:54, 5 July 2011 (UTC)[reply]
- Also fixed and reset. Trial three... Hersfold (t/a/c) 06:05, 5 July 2011 (UTC)[reply]
- Trial complete. Trial three ran ok, although I do need to look into why the bot is making null edits and try to fix that. This shouldn't be a very major problem, though, and everything else seems to be working fine. I can fix that issue as it runs without much difficulty. Hersfold (t/a/c) 06:18, 5 July 2011 (UTC)[reply]
- Null edits are most likely an encoding issue on either side. — HELLKNOWZ ▎TALK 13:23, 7 July 2011 (UTC)[reply]
- Trial complete. Trial three ran ok, although I do need to look into why the bot is making null edits and try to fix that. This shouldn't be a very major problem, though, and everything else seems to be working fine. I can fix that issue as it runs without much difficulty. Hersfold (t/a/c) 06:18, 5 July 2011 (UTC)[reply]
- Also fixed and reset. Trial three... Hersfold (t/a/c) 06:05, 5 July 2011 (UTC)[reply]
- That trial has been terminated, because the bot is adding duplicate templates to the evidence pages. Going to fix that now... Hersfold (t/a/c) 05:54, 5 July 2011 (UTC)[reply]
- Found the bug, I hadn't set up an output stream correctly. I'm going to delete the warning log, roll back the bot's previous notices, and run it again. Hopefully nobody will notice; not likely, since two are blocked right now. ;-) Hersfold (t/a/c) 05:50, 5 July 2011 (UTC)[reply]
Edits seem fine. I trust you will resolve any issues should any occur. Clerks seem fine with the operator and details can be discussed among yourselves. Approved. — HELLKNOWZ ▎TALK 13:23, 7 July 2011 (UTC)[reply]
- 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.