On Monday 22 May 2006 22:40, Javier SOLA wrote: > What we need to specify are the translation processes that we have to be > open to, and - in each one of these processes - which rights are given > to each participant.
Debian translation is currently very open (and thats a Good Thing),
basically anybody that's willing to put in the work is automatically part
of the team and gets a say, in other words Debian translation is a doocracy
> A flexible application would allow a team to chose a process, assign
> rights to each type of participant and to decide which type of
> participant can delegate rights (admit) other types of participants.
right, note that the current debian translation process is very open and
relatively informal, IMO that's a good thing
> Should only the Translation manager allow new people in?...
> or can a reviewer accept new translators?...
all that's needed to become a 'translator' is to send in an ITT and get
started, i.e. roll up your sleaves and get involved, the ITT is mostly to
prevent double work
> or a translator accept new reviewers?.
to become a reviewer, you just start replying to RFR and LCFC mails on-list
with your review comments.
-> you only need to know how to use a mail-client, and send in comments to
the list
-> the review process is informal and free-weeling:
- no 'minimum number of reviews needed'
- instead there's a minimum number of days (3 for the Dutch team) you
have to wait for comments before going to the next stage
(RFR->LCFC->aproved, RFR or LCFC might repeat) along with the agreement
that ongoing discussion blocks going to the next stage.
- ITR's also temporarily block going to the next stage (indeed that's
there purpose)
- final descision after discussion always goes to the translator.
Discussion often involves some back and forth to find better
translations (especially with some of the more technical debconf
strings)
- can branch of DICO discussion to reach concenus on translation
terminology, outcome of which gets logged in the wiki.
once review has ended the translator 'commits', usually by sending in a
wislist bug
--
Cheers, cobaco (aka Bart Cornelis)
pgpvk93cav09p.pgp
Description: PGP signature
