> I'll try to explain my point of view on koffice-i18n. It's very simple: if we > have kde and koffice (separate) packages, with different release cycles and > so, why should we package koffice-i18n and kde-i18n together?
You're right, it's a bad solution. The problem is that the other solutions are bad also. If koffice-i18n is packaged as a single massive binary package, it means that as (say) a German speaker, I'd have to download what appears to be at least 8 megabytes worth of i18n package just to get my 200k of German translations. If koffice-i18n is packaged as lots of little language-specific packages, it means we have lots more little packages in debian that add to the already confusing number of packages for users to choose from and for slow hardware to deal with. It's not like just adding another two or three packages, I believe it's more like adding another 30. If, however, we (the developers) can go to the extra hassle of combining the koffice and core KDE i18n files, it means the users have the benefit of getting koffice files as well when they download kde-i18n-de (or whatever), without having to download *all* translations in however many languages that exist for KDE and without the list of available packages having to grow by another 30 or so. Ben. :)