Quoting Daniel Nylander (firstname.lastname@example.org): > > Since we were discussing whether to use Rosetta or not, I guess these > comments from Matthew East come in handy. > > Matthew pinpoints some of the main concerns.. lack of administrative > routines, preceeding, quality assurance etc. > > Good information, spend a few minutes a read it. That's an interesting point which can be easily summarized as "no matter the tool, the work method is way more important". I happen to consider that l10n in Debian can probably be considered jerky and complicated in many matters, especially as we lack this Grail: a common infrastructure.. However, despite this, several l10n teams in Debian have achieved a very good QA process and most translations are of good quality. From an outsider point of view, the Ubuntu/Rosetta approach maybe did put the technical part (the nice non free infrastructure) before the organizational part (setting up QA processes and work methods), opening the teams very widely to any random user....and here they are with 250 French translators...:-) (I think that all translators are "registered there..including Debian translators and probably some upstream translators) The Debian French team consists of about 20 to 40 people and we already need something like 2-3 people to coordinate all this...of course very loosely as usual in Debian. And that works pretty well...because along with time, we built written and unwritten processes to not only translate a lot but also translate well. Too late right now to continue on the topic, but there's probably a lot more to say. Up to you, folks..
Description: Digital signature