Wikipedia:Bots/Requests for approval/RichardcavellBot


 * 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 delete vote.svg Denied.

RichardcavellBot
Operator:. I do not envisage that the bot or its code will ever be run by anyone else.

Time filed: 02:35, Tuesday March 8, 2011 (UTC)

Automatic or Manually assisted: Automatic. It takes no input from the operator at all once it is running.

Programming language(s): C, in the C99 standard. It consists of about 5000 lines of source code in about 160 files, written by me. It uses libcurl, and parses XML manually using . I have fully tested the bot in its own userspace using sandboxes etc.

Source code available: I don't wish to publish the source code on Wikipedia, as that would effectively license it under Creative Commons/GPL. I have published the code under the BSD and MIT licenses here : Sourceforge.

Function overview: Scan the Rove McManus page regularly and if quotes have been placed around the word "Comedian" in his biography, remove the quotes.

Links to relevant discussions (where appropriate): Quotes have been inserted around the word "Comedian" in Rove's biography over many years. It has happened 9 times in the last 12 months. The quotes are intended as sarcasm. At one point the bio contained a commented warning. See here for an example. The text of the warning has changed over time and someone has removed the warning, as the modern version of the page does not contain it.

Edit period(s): Continuous, once per hour.

Estimated number of pages affected: The bot can only edit Rove McManus, the talk page of the most recent author, and its own status page at User:RichardcavellBot/Status.

Exclusion compliant (Y/N): Not applicable. The present task request relates to vandalism.

Already has a bot flag (Y/N): No.

Function details: Scan the page Rove McManus to see whether quotes have been inserted around the word "Comedian". (The code handles combinations of spaces, double/single quotes, etc). If there are no quotes, sleep for an hour and run again. If there are quotes, check to see if they were placed by the most recent author by looking at the past revision made by the last-but-one author. If the most recent author is responsible for the quotes, rollback to the previous author's version (assuming all the current author's work is unconstructive) and place this warning - Template:RichardcavellBot-Rove on the current author's talk page. If the current author is not responsible for the quotes, delete the quotes and save the edited version without warning anyone.

The bot simulates rollback by retrieving a previous version of the page and presenting it as the new edit. I am not requesting rollback rights for the bot. The bot can be stopped by any user using this page : User:RichardcavellBot/Shutoff. The bot checks this page every time it loops.

Since this bot framework is new, I intend that a simple task like this also function as a test of the framework. I intend that future task requests will be more sophisticated.

Discussion
How would the bot avoid false positives? For example "When asked about his profession, he smirked and said "comedian"." or "Time has published a song called "Comedian" to honor McManus.". Though unlikely, I have to point it out that the bot would have no idea of the content's context. Although 9 edits a year is hardly a lot.

Secondly, "assuming all the current author's work is unconstructive" is a very broad assumption and in par with vandalism reversion. However, as far as the description goes, no parsing for vandalism is actually performed. The bot should remove the quotes, not all the material. If a human later deems that the rest of the changed content is also uncontructive, then the whole edit can be reversed.

You should strike or explain "The present task request relates to vandalism." in exclusion compliance field. I understand that this is just one page, so the field should say "No, as task is only for 1 page" or something. But I don't see how this is relevant to vandalism?

Finally, per BOTPOL you should show consensus to do this task by bot. Again, I realize it's just one page, but if somebody later raises this issue, I won't be able to point to consensus based on which this was approved. — HELL KNOWZ  ▎TALK 09:43, 9 March 2011 (UTC)
 * At the moment the bot assumes that any insertion of the quotes constitutes sarcasm/vandalism. The bot does assume that anything added by the most recent editor who adds the quotes is unconstructive and reverts it. The bot could be modified to simply remove the quotes, and could also be modified not to warn any user. As to obtaining consensus to do this by bot, there just isn't enough activity on Rove's talk page to start a discussion there. Would I go to Village pump, or to Bot owners' noticeboard, or what? - Richard Cavell (talk) 09:47, 9 March 2011 (UTC)
 * This may be worthwhile in itself, insofar as Mr. Cavell would presumably do the same thing if he were actively watching the article. As a pilot project, I share HellKnows' reservations about false positives; and I would be very careful about the wording of warning messages delivered by bot.


 * One appropriate place for discussion would be the BLP noticeboard. Septentrionalis PMAnderson 15:48, 9 March 2011 (UTC)
 * The warning message is here: Template:RichardcavellBot-Rove. I'd be happy to start a discussion at BLP/N. - Richard Cavell (talk) 20:53, 9 March 2011 (UTC)
 * Done : WP:BLP/N - Richard Cavell (talk) 22:37, 9 March 2011 (UTC)
 * Wonderful. . — HELL KNOWZ  ▎TALK 19:55, 11 March 2011 (UTC)
 * No responses at the now archived discussion. All I can conclude is that this is not a pressing matter. In any case, you still have not answered to the false positive issue of my first comment. — HELL KNOWZ  ▎TALK 10:21, 18 March 2011 (UTC)
 * The bot will give a false-positive for any appearance of quotes (double, single, with or without spaces) around the word Comedian. I think that for this page, any such addition should be presumed to be unconstructive. - Richard Cavell (talk) 10:26, 18 March 2011 (UTC)


 * Ok, I'll bite. Why does it take 5000 lines of code in 160 files to detect quotes around the word "comedian"?  Since this is related to only one article, and has happened on average once every 40 days, wouldn't it be easier to just keep the page on your watchlist and check any edits to it manually?  The down side is that the quotes might exist in the article for a few extra hours if someone adds them while you're sleeping.  The up side is that there is no risk of a bot removing the quotes when they are being used appropriately.  I don't think this is really an appropriate bot task.  &mdash;SW&mdash; converse 16:38, 18 March 2011 (UTC)
 * Or maybe just make the bot check for "comedian" and then send you an email if it finds it or something. That way you can revert it quickly.  You wouldn't need approval for that.  &mdash;SW&mdash; comment 22:32, 18 March 2011 (UTC)

This is a horrible task for a bot, and should not be approved in my opinion. Approving this would enforce one editors stylistic and content issue, and enforce that opinion through an automated process. Additionally, Snottywong's comment above points out how this task isn't even really useful. The operator could easily watchlist the page, and simply check up on it periodically. — V = IR (Talk&thinsp;&bull;&thinsp;Contribs) 18:28, 21 March 2011 (UTC)

The bot is indiscriminate to false positives. There is also no pressing need to implement this, as the page can be easily watchlisted. Some opposition here and no voiced support leads me to conclude that this is not an ideal task for a bot. — HELL KNOWZ  ▎TALK 18:37, 21 March 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.