Talk:Learning patterns/Counting featured, quality and valued content in Commons
Return to "Learning patterns/Counting featured, quality and valued content in Commons" page.
- Thanks User:Jean-Frédéric. This is awesome, I'll put it on the actual learning pattern page! SarahStierch (talk) 23:22, 18 December 2013 (UTC)
- I removed the step 3 which seemed to be redundant with step 6 (now 5).
- I added the template Quality_image in the list and replaced ValuedImage by Valued_image
- From my experiment, it seems sufficient to ask only for Quality_image and Valued_image (without QI, QualityImage, VI) since the all redirect to the long forms
- I removed the QI_seal and other seals from this learning pattern (just revert it if I was too bold) since these are used to indicate other linked files are QI (Example, which lists one linked QI), and there are about 250 such false positives on QI_seal. But perhaps this could/should be mentionned in the general considerations since it gives an indication about the use rate of QI images of the category internally in Commons (although I guess it underestimates the actual number of uses).
- Using templates is basically a bad idea, because of the mess that is Assessments. Categories are, AFAIK, the only reliable way to infer pictures labels. Jean-Fred (talk) 23:41, 22 January 2014 (UTC)
- User:PierreSelim and I are working on tools to do such things (he and I are using different but complementary methods). This is all command line as of now, so even though we have plans to turn them into web apps on ToolLabs, there is no strong incentive for us to do it. But if there is strong interest we might look into it harder :p (and we welcome pull requests too :p). Jean-Fred (talk) 23:45, 22 January 2014 (UTC)