On 6/28/06, Brad Patrick <bradp.wmf(a)gmail.com>
wrote:
I would be interested to know on what data you
rest your conclusions.
GMaxwell will back up his statements about the nature of the problem with
actual numbers (won't you Greg?) =)
Sure,
I think the most compelling bit of backup is the anecdotal one which
every person who has deleted images on Enwiki can tell you: we are
unable to get answers to basic questions for many images and must
delete them. Does anyone here who's done a substantial amount of
image patrolling on enwiki disagree? It's easy to produce examples,
just look for at orphanbot's deleted edits.
However, I like numbers... they can mislead but they can also be food
for thought. Keep in mind we don't current recommend that anyone use
the email this user to contact inactive uploaders simply because few
of the images which are eligible for deletion have a confirmed user.
There were 57,472 images uploaded to enwiki in march.
*21,000 come from uploaders with confirmed emails.
*36,472 come from uploaders without confirmed emails
There were 30,981 total images deleted on enwiki in march.
*5,149 were uploaded by users who had confirmed emails.
*25,832 were uploaded by users who did not have confirmed emails.
I wanted to give stats on what percentage of deleted images came from
users who had recent editing activity, but query is too expensive to
answer off hand.
I think the policy makes sense based both on my experience and based
on the numbers.
Hoi,
These numbers provide a snapshot. They do not indicate how these numbers
relate to the overall number of Commons and they do not indicate if the
problem becomes worse. The argument was that this trend should be seen
in percentages, this is not provided by the data. What was the average
time that deleted pictures were on the system ??
This policy is very much going to happen as you lot seem determined to
push it through.. Never mind ..
What I would be interested in knowing is the usage in percentage of
Commons pictures used in articles per project.
Thanks,
GerardM