Hello Martijin and Christian, On Thu, 2013-03-07 at 07:03 +0100, Christian PERRIER wrote: > Quoting Martijn van Oosterhout (kleptog@gmail.com): > > Last I heard Pootle had been given up due to its performance issues. > The main problem, for which I never came with a good and easily > implementable idea, is "how to get material back in packages", > considering how distributed is package development and maintenance in > Debian. > > Frameworks exist (Pootle, Weblate, Rosetta....). We could probably > easily something to inject Debian material in them. But how would > packages get the material *back* from the framework. That's the equation. Thanks for your answers. I need to test Pootle in order to get more information on performance issues. I think one of the way to integrate Framework (Pootle, Weblate, Rosetta....) in debian is using it in RFR phase and the user gets the material back to debian using the current process. For example: an user edits a PO file by Framework and sends the modified file to the local list for review. In this way Framework could be used for: 1) make easier to review PO files (only for debian PO mantained - information retrieved from Ultimate Debian Database?); 2) unified PO header: define a standard for debian PO header. Example of i18nspector analysis result: no-report-msgid-bugs-to-header-field, no-project-id-version-header-field, ...; 3) dashboard. For the first point there are many client tools (example Virtaal,...) but every installation has different settings: PO header, suggestion, warning/error and many others things. I didn't describe a long term solution/full_integration but a way to understand if a Framework is useful and issues (performance, usability, security, ...). What do you think about it? Thanks
Attachment:
signature.asc
Description: This is a digitally signed message part