[development] Multilingual support

[Gabor Hojtsy]

roberto.gerola at speedtech.it wrote:
> I have made this post only for a reason :
> to know what is the future of multiligual support in Drupal,
> from a technical point of view.
> I prefer to not spend time and efforts on solutions for Localizer
> now that wouldn't permit an easy port to the Drupal core solution
> in the next months, and that could leave the Localizer users with
> unusable systems in their hands, or with the need to rebuild their
> sites from scratch.

Roberto, we are not thinking in contrib modules now. We are thinking in 
core patches. It was repeatedly pointed out by Dries that i18n should be 
a core value of Drupal. You probably need to evaluate whether you would 
like to work on "solutions for Localizer" or solutions for Drupal core, 
if you don't have the time for both. I did choose the later.

What core maintainers will be interested in is the following:

  - a list of the problems with Drupal if we need to implement
    i18n features

  - some proposed solutions for core, and how they interact with
    future contrib features

  - easy extension points for contrib authors, so they can easily
    enable their modules for i18n

  - benchmarks of the solutions, so it can be shown that we will not
    slow down Drupal for those, who are not interested in these
    features

So in some sense, these need to be deeply in core, albeit easily usable, 
so they can be reused, and can have an impact on contrib modules 
supporting i18n, but then they need to have as small a performance 
impact if not used as possible.

For this to work, we need core level solutions (patches to be clear). We 
did have a groups.drupal.org group, where higher level discussions were 
done, and now we have an SVN repository which recently became active, as 
it was possible to build something with the new menu system and 
especially the new url()/l() changes, which made the page generation 
code much simpler and more performant. The changes we made would not be 
possible with the old menu system and url()/l() simply for performance 
reasons.

We have a mailing list, where any member can comment on patches made, so 
we can discuss code directly. If you have better ideas for any of the 
patches we make, you can just hit reply. I have seen you starting two 
threads, one of which we closed as far as I have seen. The other did 
have information not new for me, but anyway, I just replied to it. Maybe 
we can go a bit deeper there and discuss the implementation details now 
that we have a useable temporary Drupal fork to work with.

I have looked at the documentation you pointed to written by your friend 
Edward, and it seems to be a nice end user guide. What we need for 
Drupal core is a lower level focus. I did a deep code review of your 
module a few months ago (as part of writing the i18n report), but given 
the amazing pace of the work you do, it is not easy to keep up with how 
and what you change. Your adoption speed to new ideas is very good for 
the Localizer module, but some developer level documentation (or at 
least lower level explanations in discussions) would be much more 
useable in our work.

> Thank you for your work.

Thank you for your work! You did bring some new ideas into the equation, 
so being a fresher mind around Drupal i18n (at least compared to some of 
us) is a very good news for us. Keep collaborating!

Gabor