On 17/08/07, Magnus Manske magnusmanske@googlemail.com wrote:
I couldn't resist *that* temptation!!!
Good lord, you respond fast. Commons *really* needs someone to make me some soup, too... ;-)
This would probably be the simplest thing to get up and running - he said, waving his hands - and probably more generally useful than most of the other API stuff. However, we'd still need to find a way of getting author and license information standardised, so that the tool could pull them in.
I'm working on that ;-)
I could probably "screenscrape" the categories and find those that look like a license.
Determining the author might prove harder. There's the upload log (latest uploader if multiple?) and possible screenscraping of the Information template. Tricky.
Certainly my gut instinct would be take whatever's in the "Author =" field of the information template for preference - uploader usernames falls down heavily on the basis that a) a lot of people don't want to be credited that way ("by MrFancyName? Huh?"), and b) much of the Commons material is transferred to us from other projects by third parties. The "author" field, however, is pretty much verbatim how they want to be credited.
As long as we have a link to the image page, we're about as compliant with the license as Wikipedia is, and asking for more might get grumbles ;-)