On Wed, Aug 16, 2006 at 11:26:46PM +0200, Martijn van Oosterhout wrote: > I thought of all sorts of schemes but I get stuck on a few small points: > > 1. What is the goal of these checks? To have better quality translations. What's the phrase... "given enough eyeballs, all bugs are shallow". Typos, bad translations, translations that do not follow common glossaries, etc. are all "bugs". > 2. How perfect do translations need to be? As perfect as the *team* translating them wants them to be. Since users native to the language should be able to enter a team at any point in time, this actually means "as perfect as the users using the translation wants them to be" > Once we've got a decent answer to that we're done. A translation system cannot (and should not) hinder a translation team that wants no review process either because they are sloppy or their translators are great (or there is just one translator in a team). If it does people will get around it and subvert the system. Likewise, a translation system should be flexible enough to make it possible for the translation team to *choose* the best review process for them, some translations might have stricter review processes maybe because the coordinator is a nit-picky or maybe because they strive for better translations, or because some translators are sloppy and they all know it. If a interface forces a given review process, translation teams are not comfortable with, they will walk away from it and use their own mechanisms. This is, of course, from a generic point of view. I'm not asking this be implemented in the DDTS right now. IIRC, however, this was discussed in Debconf-MX when we were talking about a common translation system (and should be there in Christian's notes sent to this mailing list) Regards Javier
Attachment:
signature.asc
Description: Digital signature