Wikipedia:Bots/Requests for approval/ImageBot

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 Request Expired.

[edit] ImageBot

tasks • contribs • count • sul • logs • page moves • block user • block log • flag log • flag bot

Operator: Ρх₥α

Automatic or Manually Assisted: automatic

Programming Language(s): python

Function Summary: tag unsourced images, images with no copyright tags, and recently upload images with no fair use rationale

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

Edit rate requested: maxlag = 5

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

Function Details: Bot will search special:log/upload and look for images with no copyright tags, unsourced images, or images without a fair use rationale. The bot will then notify the uploader depending on what information is missing.

[edit] Discussion

Several questions:

1. How does this differ from what User:OrphanBot does?
2. How will the bot decide what an "unsourced image" is?
3. How will the bot tell if an image has a fair-use rationale or not?

--Carnildo 20:59, 9 November 2007 (UTC)

1. I didn't know OrphanBot existed, but I don't see harm in a bot doing a similar task :).
2. Unsourced image is any image w/o any information.
3. If the image has a fair use template, but less then 10 words it tags. It skips if it has the template for fair use rationale. Ρх₥α 21:07, 9 November 2007 (UTC)

No harm in duplicate bots. Approved for trial (25 edits)., see how you get on. Reedy Boy 21:15, 9 November 2007 (UTC)

Y Done No problems. Ρх₥α 23:00, 9 November 2007 (UTC)

It seems to have marked a number of sourced images as "no source": [1] [2] [3] [4] [5] [6] [7] [8] [9]. A 36% error rate doesn't strike me as "no problems". --Carnildo 00:40, 10 November 2007 (UTC)

I must of used the wrong template, no? None of the images had any copyright information (ie CC by 2.0, GFDL, etc). Ρх₥α 06:02, 10 November 2007 (UTC)

Since the bot asked the uploader for source information in each case, it seems to me that it's a logic error, not simply a matter of using the wrong template. --Carnildo 07:04, 10 November 2007 (UTC)

Well, I fixed it. BAG, should I run it again? Ρх₥α 18:03, 10 November 2007 (UTC)

Approved for trial (10 edits). Try another 10! Reedy Boy 23:10, 10 November 2007 (UTC)

Okay, I'm done. No problems with the bot. Ρх₥α 01:36, 11 November 2007 (UTC)

Are you actually looking at what the bot does? It still doesn't seem to be able to tell the difference between "no source" and "no license", and it's still asking uploaders for source information, even if it's applying a "no license" template to the image. Further, it put three messages on User talk:Little miss jenna asking for source information for the now-deleted Image:Bukkake1.jpg, and it seems to have been edit-warring over tagging the image. Is there anything limiting how many times it will apply a tag to an image? --Carnildo 02:09, 11 November 2007 (UTC)

Yes, I am watching it :). The first few new edits were messed up because I forgot to change the template it uses, but [10] [11] [12] [13] the rest were fine. Ρх₥α 02:13, 11 November 2007 (UTC)

• We really need to see a good error-free/very-low-error run. Try again, Approved for trial (30 edits).. — xaosflux ^Talk 16:33, 11 November 2007 (UTC)

If you're only proposing to run this on newly uploaded images, I suggest that any approval be made with that condition. We have had a lot of trouble with bots unapproved for this specific function that go through the old store of images and tag them for lack of source. The source information is often in there implicitly, or using plain English rather than websites, without the magic word "source", etc. A casual comment "obtained from the owner" anywhere on the page, or "from the publisher", "belonging to the company", etc., indicates source. This isn't just a theoretical concern. Efforts to automatically tag images for lack of source have always had high error rates and generated a lot of complaints, and they wreak havoc on types of images like logos, album covers, film posters, where the statement of source is often worked into the use rationale. I don't think there is any way a bot can practically determine this. Going forward it is reasonable to ask people to use the magic word "source=", use a source heading, or something like this, but it should be a policy decision and not a de facto decision necessitated by bot. Wikidemo 19:32, 11 November 2007 (UTC)

{{OperatorAssistanceNeeded}} --uǝʌǝsʎʇɹnoɟʇs(st47) 23:10, 27 November 2007 (UTC)

I'm here, just busy. I plan on trial running it real soon, when I can watch it changes. Ρх₥α 00:22, 29 November 2007 (UTC)

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.