User talk:MPinchuk (WMF) (usurped)

Template test?

Hey Carnildo,

Steven Walling and I have been running some A/B tests on common user talk templates (check out our task force for more info), and I was wondering if you'd let us tinker with the warning templates that your bot sends (not in the technical sense, just with the actual content of the warning). Right now we're working with Beetstra and Versageek on some redesigned warnings for XLinkBot and with Kingpin13 on an SDPatrolBot warning test, and since ImageTaggingBot is another bot that hits a huge number of talk pages, it would be awesome if we could test out some different warnings with it.

Let me know what you think – you can catch me on IRC if you hang out there at all (nick:Maryana), or just play talk-page tag. Thanks! Maryana (WMF) (talk) 20:33, 28 November 2011 (UTC)

I'd be interested in it. The current user messages the bot uses are:

• User:OrphanBot/deprecated: the uploader used a deprecated license tag (extremely rare; the bot's logs don't show it having happened at any time in the past year).
• User:OrphanBot/nosource nolicense: the uploader did not provide anything that the bot could recognize as source or license information. The bot is extremely liberal in what it will accept, so this usually (but not always) means a blank image description page.
• User:OrphanBot/nosource untagged: the uploader did not provide anything the bot could recognize as potential source information, but did provide something that might be license information. I don't think the bot's current logic actually permits it to use this message: anything that meets the bot's criteria of "might be license information" also meets the critera for "might be source information".
• User:OrphanBot/nolicense: the uploader provided something the bot recognizes as source information, but did not provide anything that the bot could interpret as license information. This usually means the user filled out a template such as {{information}}, but did not provide a license template.
• User:OrphanBot/nosource-new: The uploader provided a license template, but did not provide anything the bot could recognize as a source.
• User:OrphanBot/untagged-new: The catch-all situation: the uploader did not provide a license template, but there's text on the image description page. The bot isn't smart enough to tell the difference between a free-form description of the source and license (rare), a vague description of the image's source (common), or a brief description of what the image depicts (very common), so it makes the most conservative assumption and marks the image as not having a license template.

The corresponding templates placed on file description pages are:

• {{no copyright information}}: deprecated, nosource nolicense, nolicense
• {{no copyright holder}}: nosource nolicense, nosource untagged, nosource-new
• {{untagged}}: nosource untagged, untagged-new

These are the actual messages ImageTaggingBot uses, so changing them on the wiki will change what the bot puts on users' talk pages.

What's the best way to have the bot change what it uses? I can modify the bot to randomly select a message from a group, or the pages I linked to above can be modified from time to time. I'd rather not try any sort of fancy wikimarkup to switch messages: I'm not sure how template coding interacts with "subst:", and I'd like to keep the actual markup on users' pages as clean as possible. --Carnildo (talk) 02:59, 1 December 2011 (UTC)

Fantastic! So happy to have you on board :)

When testing on Huggle and Twinkle, we've been using a template randomizer that our summer researchers hacked together. You can take a look at how it works here (substituting is no problem). With bots, we've just let the bot ops change their code as they see fit. I'm not a bot expert, but Beetstra, Versageek, and Kingpin13 are, so you can ask them about how they've done randomization if you're interested. So it's really up to you – the template randomizer is fairly simple and effective and doesn't leave any extraneous code on user talk, but you certainly don't have to use it. The one thing that would be nice to have is a log of all users warned during the test (which is something that bots are great at doing, of course).

Steven and I can take a stab at tweaking the templates and drop you a link to make sure you're happy with them, and then we should be all set to go. Thanks again for volunteering your bot! Very excited to see how this test goes :) Maryana (WMF) (talk) 19:15, 1 December 2011 (UTC)

repectfully

I will not contribute to wiki until the copyright rules are changed so that my work can't be reused by for profits. I just find that this sticks in my craw.68.236.121.54 (talk) 14:38, 12 December 2011 (UTC)

Template testing

Hey there.

Sure, I'd like to help if I can. There is one technical hurdle to handle, however: my bot needs to recognize some of the templates it leaves on articles and on maintenance pages; any randomization would have to be in a parameter rather than switch the template itself to avoid problems.

That said, it substs the actual user notices so those can be safely randomized. The templates it uses are:

• {{csb-pageincludes}}, {{csb-pageincluded}} or {{csb-wikipage}} are transcluded at the top of potentially problematic articles; and
• {{csb-notice-pageincludes}}, {{csb-notice-pageincluded}} or {{csb-notice-wikipage}} respectively to the talk page of the original editor of an article tagged with one of the previous templates. Those are substituted.

Additionally, it will subst {{welcomelaws}} if the editors' page is entirely blank to begin with, and it uses a half-dozen internal templates that only go on maintenance pages in project space we probably don't want to mess with.

Just tell me what you need from my end; it's trivial for me to pick from a set or add parameters to templates as needed. — Coren ^(talk) 02:25, 7 December 2011 (UTC)

Taking a crack at rewriting these templates, and I have a question for you: how is {{csb-notice-pageincludes}} different from {{csb-notice-pageincluded}}? I.e., what's the criteria the bot uses to decide between "material copied directly from" and "a substantial copy of"? Is the latter used for cases of wholesale article copyvio and the former for just individual sentences/paragraphs? No huge rush... I'm still trying to knock out some ImageTaggingBot alts at the moment. Too many nice bot herders letting us play with their bots! :) Maryana (WMF) (talk) 00:43, 9 December 2011 (UTC)

It's a question of which is a subset of the other. Pageincluded is used when the wiki page contains (most of) the external page along with other stuff – all or most of the external page has been copied; pageincludes is when the wiki page is an excerpt from the external page – so it might be one section of a large page, or a very long quote. In theory, the former is "worst" of the two, and was worded a little more strongly. — Coren ^(talk) 00:58, 9 December 2011 (UTC)

Hi Maryana!

Nice reply @ ANI.

Good luck with this work. – SJ + 22:31, 11 December 2011 (UTC)