Improving Rosetta's relation with upstream translation teams
danilo at canonical.com
Mon Oct 9 11:13:22 BST 2006
First, a reminder: my points are strictly related to problematic
upstream–Ubuntu translation team relations. (Note that this is in the
subject as well ;)
Today at 9:17, Matthew East wrote:
> It's totally unrealistic to expect to find people with the time to go
> around checking all suggestions for an entire distribution. The team
> already works closely with upstream (at least in the case of Gnome) and
> is extremely approachable if anyone from any other upstream identifies a
> problem with a translation. We haven't had any complaints so far, and
> the idea of throttling the productivity and working practises of the
> team sounds to me like a serious overraction to a problem that is not there.
For those who have no problems, no change is necessary whatsoever ;)
Basically, if you had no problems (team members doing unapproved
changes on their own, messing up with things like "translator-credits"
etc.), then even if there are many members in ubuntu-l10n-it, they are
all "highly trusted" (as per my previous email: you would also trust
them with CVS account).
And, we (specifically, Carlos) are working on improving the review
interface. Also, there are other workarounds for an easier review
(eg. using PO files and sending them around instead).
> That's a different story - teams which admit anyone definitely do a lot
> of damage, but there are lots of teams which have found a middle ground
> and have developed good quality control, and limiting those teams by
> making a couple of people check all translations would be a serious
> shame. In my view if you are really serious about that rule, it is worth
> discussing with the Community Council before implementing it so that
> translation teams who have developed good practises can make their case.
The "rule" in question was simply a *suggestion*: if nobody is
complaining, you can go on as you did. It's also still ok for smaller
teams to admit anyone, and think about QA later on. The problem
arises iff the upstream teams are pretty good at their job, have their
own practices and rules, and Ubuntu translation teams allow anyone to
mess with their work.
We are not going to enforce any such "rule" with code or whatever.
(Also, there might be cases where Ubuntu l10n team would be doing
massive changes, such as terminology changes, but such team would have
to be as well organised as upstream, so no problem there, imho.)
More information about the ubuntu-translators