On Sat, 01 Oct 2011 19:57:40 +0100, Nuno Tavares nuno.tavares@wikimedia.pt wrote:
Yaroslav,
I agree that is bot work. However, in our case, we just followed a somewhat different approach in face of the circumstances: we proposed ourselves to review image by image - please keep your flames out of the discussion :-))
We developed a tool to verify the photo data[1] so users can manually verify legitimacy/correctness of the data. This is obviously a lot of work, so we also did a tool to allow to verify/accept/reject the image with a single click, propose categories or add categories we had already mapped.
[1]
http://www.wikilovesmonuments.org.pt/tools/verify/?debug=0&wlm_id=70523
The verification tool transparently fetches this info and acts upon user decision: reject/accept, but the big thing is that we automatically get the images categorized with a bigger accuracy. Here's an example (with a single click):
http://commons.wikimedia.org/w/index.php?title=File:Castelo_de_S._Jorge_(20)...
This is just another option put on the table countries may consider, so no one relies on a bot we never saw, and barely heard of.
Nuno, thanks for the tip. I am not part of any of the organizing teams, and it is not up to me to decide whether in particular countries the review should be implemented. However, I feel that here we have two different issues:
1. Determining whether the image is correct (the image corresponds to the caption, and possibly FoP issues); 2. Determining whether the image is ok (has a license, is properly categorized).
Whereas 1. indeed is better done by dedicated teams and possible as review, 2. can be done by any Commons user who my have no interest in WLM. I think the scheme I asked about could definitely facilitate 2. and may also facilitate 1. (in the sense that it is easier to review the images of the same monument all together and not in random order).
Cheers Yaroslav