Your monster post was very useful, gave me a few ideas. Also it's a good set of requirements to compare against for things that should be fulfilled by a potential de facto model translation approach/app. Combined with the excellent wiki article mentioned in this thread (http://code.djangoproject.com/wiki/ModelInterNationalization) it's a good basis for evaluating existing model translation apps, to see what is missing or "wrong". Regarding django-datatrans, I've been planning to take a more serious look at it ever since I started using it. This thread might just be the "kick in the butt" needed, we'll see :).
A few issues I'm aware of: - _pre_save handler can cause problems on concurrent inserts - integration with Django admin (or any other functionality that presents the "translatable" fields in an editable context in the frontend - for ex. in a form text input) - the editable widget will get filled with the translated string and if you save it, the *original* will get overwritten. On the "admin" of datatrans, there is no problem, since it works directly with KeyValue-s. This is usually not a problem on frontend because in most cases it makes no sense to have translated content for values which are editable in frontend by users. And admin is a controlled environment, so it's not a "showstopper" (at least in my experience so far). But a transparent solution is definitely needed. In any case I don't think we'll see any model translation app in django.contrib any time in the foreseeable future, taking for example South as a reference on it's progress in this regard (considering it is miles ahead in terms of maturity/stability/usage etc. than any modeltrans app and it has pretty much achieved it's de facto status in it's problem domain). I'm not saying this is a bad thing, just an observation of facts. As long as a 3rd party app is well known, easy to integrate with existing projects, it's stable/mature, well maintained etc., I don't think it makes very much difference if it is actually within django.contrib or not. Botond On Aug 6, 12:01 pm, JK Laiho <jkla...@iki.fi> wrote: > On preview, django-datatrans looks really good, and the approach is > certainly better than any of the existing implementations, including > my abortive one. Still need to give it a run for its money properly to > see what issues remain. Whatever they are, they're probably solvable. > I'm not a betting man, but I'd be surprised if that didn't grow into > the de facto model translation approach in time. > > I'm just glad I don't have to think about the model translation > problem anymore, I was exhausted just writing that monster post > yesterday :-) > > - JK Laiho -- You received this message because you are subscribed to the Google Groups "Django developers" group. To post to this group, send email to django-develop...@googlegroups.com. To unsubscribe from this group, send email to django-developers+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/django-developers?hl=en.
