On Mar 23, 2006, at 11:04 AM, Mark Wagner wrote:
On 3/22/06, Jesse W jessw@netwood.net wrote:
On Mar 22, 2006, at 11:30 AM, Mark Wagner wrote:
Wikipedia's got a problem with images. 2000 of them are uploaded every day, and most of them have inadequate source information, an incorrect license tag, or an invalid fair-use claim.
I think this is a little premature. It is really important to remember that, with our current methods - *The Number Of Untagged Images Is Going Down*. The images tagged as lacking source or license info no longer *have* a backlog(due to the change in CSD policy allowing them to be deleted, and the hard work of a number of Wikipedians), and the backlog of Un-tagged images *is going down*. Not by much; about 300 per day on average, but it *is going down*.
Give us working on the untagged image project time to get the backlog cleared up, and then we can see the problem more clearly. Quite possibly, the rate of new unacceptable images is not too great for us to handle. Maybe it is, but the current facts don't seem to show that.
And once you finish with the untagged images, which project are you going to move on to? Verifying the fair-use claims on the 150,000 images in [[Category:Fair use images]]? Checking the accuracy of the 80,000 or so "GFDL" images? Finding the few images where the "No rights reserved" claim isn't bullshit? Image tagging is an important task, but it isn't the only place where work is needed, either.
Either of those two tasks are useful. I was merely saying that we are making progress, not falling behind, afaict, so changing the process may not be necessary.
It's also important to note that 95% of the images on the english wikipedia do have tags (although many may be wrong, it's true).
I haven't checked already-tagged images, but new uploads (based on the data at http://mail.wikimedia.org/pipermail/wikien-l/2006-March/041534.html) have at least a 40% error rate for tagging.
OK. That is an issue.