[Foundation-l] The status of smaller languages on the Wikimedia Commons

[Daniel Arnold]

Am Freitag, 05. Mai 2006 11:30 schrieb Gerard Meijssen:
> Wrong. InstantCommons has everything to do with the WM Foundation.
> InstantCommons would have been coded by now if this was not the case.

InstantCommons is a nice idea but without a notification framework that 
*automatically* notifies other projects about things that need their 
attention like duplicate files, files missing required info and all kinds of
deletion requests it simply would increase our current problems by a order of 
magnitude and thus would make Commons unmaintainable.

Duesentrieb (http://commons.wikimedia.org/wiki/User:Duesentrieb) is currently 
writing an automated notification framework called CommonsTicker that would 
run on the tool server and would work as follows:

A database cronjob at the toolserver looks for new files that need the above 
described attention. The new files will be send through CheckUsage on the 
same server and this will return a usage list of these images of every local 
project. In the end a pybot will write in every local project a at a 
dedicated page a daily compiled list of files that get used by them and that 
need their attention (like unlinking images and thus reducing image "holes" 
in articles, alerting people on deletion debates in Commons and so forth).

Of course this solution can only be temporarily and needs to be integrated 
into MediaWiki via a RSS feed that can be imported by other MediaWiki 
installations (which will do the automated checkusage filtering locally on 
their own on the imported RSS feed) and thus tracked by them.

So at first we need to solve our current Wikimedia wide communication problem 
and see how CommonsTicker performs, after that it needs to be coded into 
MediaWiki and later if things have stabilized after a time in Commons we can 
think of how to realize InstantCommons.

> A summary of what happened is nice. It is not what I am really
> interested in. What I am interested in is having things announced before
> they happen. This will have, as an effect, that people do not have the
> excuse that they did not know. They will have less reason to moan (they
> will anyway). When you say nobody told us, you mean nobody told you.
> Fine. This has been discussed at previous occasions when we had a flash
> in the commons pan.

See above. We have so many things to do daily it is impossible notifying 
people in local wikis on a regular basis manually. And of course there are 
also other people suddenly jumping into Commons (you never heared of them 
before) and that just put thousands of images onto deletion request without 
any debate or worse simply make mass image changes (and are as well unwilling 
to share their points) and that make really *big* trouble if I stop them. In 
order to get an impression how much energy it took for me making a proper 
debate on a certain single issue see:
 
http://commons.wikimedia.org/wiki/Commons_talk:Licensing#Review_of_license_templates

Well and for sure there are many more such people I can't be aware of and 
ocassionally someone makes a general Commons-admins-flame on Village Pump and 
afterwards we try to find out what he was flaming about at all and who was 
the person that did something wrong...

So there are two sides:
* One side that makes trouble because we talk to much and do not delete fast 
enough.
* The other side that says we talk to less and delete too fast.

Seems to me that at least one side must be wrong...

The facts are:
* Far more images need to be deleted in Commons daily.
* Local unlinking of files and manual local notifications are such a huge task 
no Wikipedian can imagine (really Wikipedians simply don't realize *how huge* 
Wikimedia universe is).
* Many people upload images in Commons without even interested in a minimum of 
communication.
* A tiny fraction of active Commons admins tries to catch up...

> As to my definition of "profound"; changes in policy that affect usage
> need to be communicated widely before they are finalised and implemented.

Well I could have talked a lot previous to my many *very bold* changes in 
Commons and wouldn't have come so far. I prefer in that case a model of 
communication with some people I know, switching in my brain previous to an 
edit of mine and and then simply do the thing and if something is not that 
ideal hey we can sort it out afterwards. That's the only way how things can 
efficiently get done in Commons.

Regards,
Daniel/Arnomane