Gregory Maxwell wrote:
My best guess is that he browsed the site during while we were partially down (it's not always obvious to anons, since the squids will work on their own)... and the commons text couldn't be fetched? I'm just guessing at this point.
I guess youre right :) Its the same time i see this particular failure mode, too. Seems to be using Mediawiki:shareduploadwiki and Mediawiki:shareduploadwiki-linktext
My answer:
We're sorry about it. The practise of not putting image sources is similar to the practise of not signing articles (instead, you need to click history).
There have been some ideas about automatic-author-filling in the past, but there's neither the system to do it nor a specific method on how to automatically choose the right author (but we're slowly improving).
We also suffer image-copying, and try to do our best with it. From your screenshots, it seems that he copied the *thumbnailed version* present on the article. Had he go for the full-size one, he would have needed to past via the intermediate page. Sadly, our thumbnails don't keep a pointer to the description page (Exif data had to be removed too, due to thumbnails embedded on them). If you have a good way to do it, or improve our handling, please let us know.
You're also welcome to follow the discussion at http://lists.wikimedia.org/pipermail/commons-l/