Looking for Pootle alternative

For some time, we use Pootle as online translation tool for phpMyAdmin. It is indeed great tool, but somehow we seem to be hitting more and more issues with it and I feel it really won't fit nicely once we will want to support translations on two divergent branches. Also it's code base is not something easy to understand so that I would be able to fix bugs myself without major effort.

So this is when I started to look for alternatives. Unfortunately there is nothing I could find what would work in way I want. I've looked at Transifex, Translatewiki, Crowdin and others whose names I've forgotten and I still haven't found what I'm looking for.

The key feature I want from such system is full integration with git. This means that whatever is changed should be committed to local git branch and merging translations back should be as easy as merging remote branch. I've hacked such thing into Pootle (in quite dirty way), so this is basically workflow we have right now and I'm pretty much happy with that.

Another thing (which seems to be existing in some tools) is support for simultaneous translation of more branches - if same string exists in more branches, it should be translated in all of them. And last but not least, it should support import of po files allowing users offline translation.

Thinking about this I'm starting to thing that it should be pretty easy to implement something what I want using existing pieces of software without major effort. I'll probably start to play with Django, translate-tookit and GitPython to see if I can come up with something working reasonably well in few days...

PS: Anybody working on updated GitPython packages for Debian, looks like I will need them soon :-).

PS 2: The solution for me is Weblate.

Written by Michal Čihař
Filed under: English phpMyAdmin Weblate
Published on Feb. 28, 2012, 7 a.m.
Share:

Comments

Maybe you should look at Damned lies as a base for your work.

Check it out at http://l10n.gnome.org/

I thought Damned lies just provides overview about translation status, so it would not be that helpful. Or am I wrong?

Oh Damn Lies looks pretty interesting. Alexandre should have linked you here instead:http://l10n.gnome.org/about/

Ben

Hi Michal,
I'm also working on a an alternative to handle translations. Own importer & exporter can be added. Check out http://otrance.org (demo) and if your interested, feel free to contact me. ;)
Cheers, Daniel

Thanks for the tip, but it still looks like something what I would need to adjust quite a lot.

On the other side, I have basic translating working on my own system right now, it just misses some advanced features (I expect to complete them this week) and some CSS to make it look like some application and not just set of random links and forms ;-).

(Disclosure: I'm founder of Crowdin)

Sorry to hear you did not find Crowdin as a suitable solution for phpMyAdmin.

However I would like to clarify on a few points here:

Regarding GIT integration: yes, unfortunately integration with CVS's is still in stack and we can not offer "out of the box" solution, nevertheless the current API allows you to integrate closely translation process with your GIT repo.

Maintaining translations in several branches: yes, we god this.

We will be glad to talk with you regarding your localization needs in Skype or via email, please ping us at contact@crowdin.net

We are open for new ideas and certainly ready to implement new featuers that will make Crowdin better.

Well I've used Crowdin (I contributed quite a lot to Flattr translations), however I always felt it is too complex for what I'd like to have (for example I never understood how voting is supposed to work and how many votes are needed to make translation approved).

Usually we have one person working on one translation and things like review or voting does not make much sense for us.

All the features we have are quite essential for effective collaborative translations. We think a lot before adding something. Could you be more specific and describe what features or functionality are excessive or useless, in your opinion?

Say, glossaries are important even if you have only one translator maintaining one language, same things with comments section since it's important to discuss or clarify something regarding current string with translators to other languages.

Regarding votes, so this feature can be used on your own. With crowdsourcing the votes usually help proofreaders to choose which translation is better. Also (if defined workflow does not include translations validation) our build script consider votes number on each translation suggestion and decide in this way what to put in resulted file. There are no any minimal number of votes required for suggestion to be used as translation.

I don't say they are excessive or useless in general, but I hardly find usage of them for our workflow - we generally have one person taking care of single translation and in such case there is no use for voting or comments as there is only one person which would do that. Generally we want to put barrier as low as possible to allow people contribute directly...

And about voting (on Flattr translation) I simply failed to see how many votes are needed to make translation stable and I'm even affraid it is so many votes that there never will be so many people doing Czech translation.

This just exposes the fact that voting might be nice for huge languages with many possible translators, but for not so widely used languages it just puts barrier too high to complete the translation (in this case it is 100% complete for quite some time, it just misses votes).

As I said voting is not necessary and can be just ignored in your case.

Anyways, we still would love to see you as our user. If you decide to use our service feel free to contact me directly or our support team.