As one may have noticed, no new language has been activated as of now in D-I since etch.
Indeed, given the tremendous number of languages we support (http://d-i.alioth.debian.org/doc/i18n/languages.html), it becomes harder and harder to find and motivate people for the remaining ones. Most of the potential people are already working on other FLOSS localization and are quite overwhelmed by work. One of the problems we have in D-I is the amount and complexity of work. As of now, there are 1686 strings in the POT file. A big number of these strings are indeed very technically challenging, particularly for languages that haven't yet developed a complete technical vocabulary. Most of the future new translators already told me that requesting this to be 100% translated is scary to them and they have hard time motivating ppl to work on very obscure strings translations. To answer that concern, I developed the concept of "sublevels". It can be summarized easily: there are two sets of D-I packages: - priority set: all packages that display somethign in default "common" installs for "popular" architectures (i386, amd64) - regular set: all others Frans and I had some discussions about these lists and this lead to "packages_list_sublevel1" and "packages_list_sublevel2" in packages/po The idea then becomes: to be activated a language has to complete sublevel1 to either 100% or at least a very high percentage (abobe 98% or so). This reduces the required number of strings to 907 as of now, which is much more sustainable. Would this be an acceptable criterion? If so, I will setup a way for me to follow the progress of this and to offer "prospective" translators the updated PO files for sublevel 1 during the time their language is not yet activated. Once this sublevel1 file is completed, the language could then be promoted to an intermediary state called "development" where the status of sublevel1 is still monitored (to guarantee it remains above 98%) while the translator works to complete the complete file (or works on level 2 and other important files). Once the complete level1 file reaches, say 90%, then the new language joins the list of "fully supported" language and the translator has to keep is 100% for the general big file. Of course, that would work both ways: a neglected translation for which level1 falls under 90%, could be downgraded to the "development" state and then, required to stay 98% for sublevel1. Otherwise, the language would be deactivated from D-I. As a summary: - prospective language: - not activated in D-I - works on sublevel1 PO file - goal: 98% in sublevel1 - development language: - activated in D-I - works on level1 PO file, monitors sublevel1 file - goal: 90% in level 1 - criterion for downgrading to prospective: - below 98% for sublevel1 - supported language: - activated in D-I - works on level1 PO File, monitors both files - goal: 100% in level 1 and complete other levels - criterion for downgrading to development: - below 90% for level 1 - criterion for downgrading to prospective: - below 98% for sublevel 1 --
signature.asc
Description: Digital signature
