Django should use an online translation system.

[Stavros Korokithakis]

An online translation system would greatly lower the barrier of submitting translations. The current system of submitting SVN diffs is a bit cumbersome, and as a result the translations are not very complete.

[Marc Fargas]

That's already been raised before. But it's not a priority atleast for current translators; Neither for core developers.

So unless you want to step in to develop such a tool, or set something up to show case I doubt it will go much further from being an open ticket susceptible of being closed 'wontfix' or 'works-for-me'.

[Stavros Korokithakis]

There already are such systems (e.g. ​http://transifex.org/, which is OSS and actually written on Django), and I don't think it would be that hard to set up... Its developers can frequently be found in #django, so one could ask there for any information...

[Stavros Korokithakis]

It might also be possible to just upload the .po files to Launchpad and let users translate them there, its translation interface is very easy and effective.

[Malcolm Tredinnick]

Integrating with an existing online system is certainly a possibility. A requirement for any choice is that it won't *require* people to use it, since a web browser is a horrible text editor, so we shouldn't force people down that path if they're already comfortable with their existing tools and editors. There would also need to be ways to have staging and validation of changes, because we don't want automatic commits of essentially rubbish translations going on.

I've looked at transifex in the past and it's a possibility, certainly, since it is possible to make it worth as an addition to existing processes. Launchpad's system certainly used to be much clunkier than transifex (I would consider the latter to be the leading candidate for something like this, by the way -- it's very well designed).

In any case, though, this isn't really a good ticket, since it's a huge item with no single resolution. Broad feature additions and process enhancements like this are more "TODO list" items than bug reports. Closing as invalid for that reason. After 1.1 is out, feel free to start a discussion on the django-i18n list about this, bearing in mind the above requirements and thinking about any extra things that might be appropriate. A good post that lays out some possibilities and the trade-offs (something that only says "Django should do this" isn't useful, as there's clearly multiple sides and compromises in anything like this) will be interesting to read and discuss. Right now isn't the right time, though, since it really can't get the full attention it needs with Django 1.1 coming up (it won't just be the translators themselves that are involved in the discussion and decisions).