Op 14 apr. 2012 om 00:01 heeft Strainu <strainu10(a)gmail.com> het volgende
geschreven:
Re-adding the list, as this is of public interest.
Wow. It's generally considered to be pretty rude to just publish a private reply to a
public list. I'm assuming good faith, but please ask next time, would you?
În data de 13 aprilie 2012, 19:42, Niklas Laxström
<niklas.laxstrom(a)gmail.com> a scris:
I do consider users (at least one, me!).
Glad to hear that, but you're not the main user of either
translatewiki or mediawiki in general.
What's the relevance here, or do you feel like Niklas needs an old fashioned and
proper bashing?
If we must
initiate
discussion (if there even is a community) and wait for consensus (that
might never happen) for every change we do, we will never get anything
done.
That's plain wrong, and Wikipedia is here to prove it :) Many user
scripts were developed after consultation with the community. They're
working great and do what users expect them to do, not what the
developers want them to do. I do agree that developement would be much
slower, though.
Hyperbole is the use of exaggeration as a rhetorical device or figure of speech.
Like I mentioned in the email, I proposed that we notify affected
Wikpedias *before* the changes are deployed [1] and executed that
together with Gerard. They were given right to veto any change they
didn't like or wanted to discuss. And some of them did use that
possibility.
Search for "translation" on the Romanian village pump:
https://ro.wikipedia.org/w/index.php?title=Special%3AC%C4%83utare&profi…
The ro.wp embassy:
https://ro.wikipedia.org/wiki/Discu%C8%9Bie_Wikipedia:Ambasad%C4%83
I can see no such notification. No message from you, and only one from
Gerard, when the translation team was founded. If you chose only some
Wikipedias, then let me tell you this is even worse than no
notification at all. The thing is - the wiki environment is by
definition decentralized and hard to follow for small communities. If
you choose to notify only big wikis or the wikis that already receive
support from the WMF (e.g. Arabic, Indic languages), you're basically
having no effect at all.
http://www.mediawiki.org/wiki/Special:Code/MediaWiki/107309 is not about Romanian, so
there's no reason for such a message at this point in time. Also see
http://lists.wikimedia.org/pipermail/wikitech-l/2012-February/058189.html.
As to solutions, I have proposed at Wikimania 2010 to
the person that
was coordinating the Translation hub from meta to put up some
automated notification system when a new translation request appears.
This hasn't happened and translations in most languages are still
happening by chance, as different users go by the translation hub.
It's being worked on. Expect something soon.
I suggest you do the same, at least for
Special:AdvanceTranslation
strings. Some time before a scheduled deployment (ideally, 7 days, but
at least 72 hours), notify _all_ the village pumps and some users
(either all the users that translated the advanced strings, or all the
users that sign-up on some dedicated page) of the deployment. This
way, you don't have to wait for the consensus, but you're giving
communities an opportunity to reach it.
Nah. These changes are usually do rare that we'll just approach communities as we go
along. It seems that MediaWiki core namespace changes are now the most heavily QA-ed and
agreed upon changes in all of MediaWiki.
The sign-up system might seem complicated, but it
works - we use it on
ro.wp for unblocking requests and for responses to warnings: we're
handling all unblocking requests in a matter of minutes, vs. days
before the bot warnings.
See two paragraphs up. You can also consider signing up to translators-l for now. See
https://meta.wikimedia.org/wiki/Special:LanguageStats for current translation requests.
Another way to go is to have regular deployments as a
rule. Right now,
you guys are something like "oh, we'll do it about every 3-4 months,
if the review queue is not too big and if something goes wrong we'll
push it for 2 more weeks". Having the deployment schedule a year in
advance, including the features you're targeting on each deployment,
and enforcing a maximum delay of 1 week used only for deployment bugs
and nothing else would greatly enhance predictability and give
translators a target to aim for.
You are making many assumptions here. Not sure about the scope -- "you guys" is
pretty generic--, but I'm pretty certain that there are a few variables
(understatement) you failed to take into account that prove this theory as implausible.
P.S. My "accusation" was actually an
observation based on my past
interactions with you on this list and on bugzilla. I don't mean any
disrespect, but I stand by it.
Which accusation was that?
Cheers!
Siebrand