On Tue, Aug 15, 2006 at 07:36:44PM +0200, Martijn van Oosterhout wrote: > >Moreover, I have a question: how do you check for the identity of people > >that translate and review? I mean, how may you check that a reviewer for > >package XX isn't the translator of the same package XX? > > It makes a guess to see if you're coming from the same computer. It's > not foolproof but it works pretty well. Enough so that you don't > review your own translations and to know if you're reviewing the same > translation twice. The funny thing is that people can subvert those kinds of checks if they simply have access to two different systems and they can simply "review" and "approve" (in one) things they "owned" (in the other one). However, for the current setup (experimental and temporary) this authentication is more than enough [1] For future pending setups I would, however, like to: 1.- Not have the application decide how the language team reviews or approves translations (the current system does not cooperate well with one-man language teams). That method has to be decided per-team. Some teams might want people to be able to translate and approve inmediately, others might want a tigher review process (one or more people check out the translation). This could be even be different per translation (more reviews needed for higher priority / higher visibility translations than optional translations). 2.- Provide per-user authentication and have a coordinator be able to set up a "confidence" level per translator (think of it as a rating, your don't review as often thinks from people that you have been working for years and you are confident in their skills as opposed to new people that have not proved themselves). Just my few cents. Javier
Attachment:
signature.asc
Description: Digital signature