Re: What about DDTSS do you (dis)like?
Martijn van Oosterhout píše v Čt 09. 06. 2011 v 08:44 +0200:
> As threatened, I've started building a django project which aims
> integrate the DDTSS more closely with the rest of the system. Other
> than performance it should improve maintainability. Since the new
> system will use a real database with a real schema, it becomes easier
> to do certain things. But also because it doesn't use the email
> interface it has more possibilities and some of the current features
> are no longer relevant. Here is a list of things that will needed
> changing. I've you have any comments, suggestions, etc please respond.
> If there are features people want/need, now is the time to make say it
> (I don't promise to build every suggested feature).
Thank you for your working on DDTSS.
>
> 1. The whole "pending translation" becomes irrelevant, since we can
> peek directly in the database. Instead I was thinking of always
> showing 10 descriptions, but allowing people to filter by
> distribution, tags, priority, popcon, etc. This seems like an
> uncontroversial improvement to me.
>
Very good idea. Here is real model of translating package description
for our team (Czech - cs). I'm fetching descriptions and give it to the
translators to translate. Then I review their work, so pending
translations isn't needed for us, but there are others.
> 2. Similarly, the "recently translated" becomes irrelevant, because we
> can poke it directly into the DB. However, I can imagine people would
> want to keep this. It contains submitted and reviewer information and
> timestamps. Do people have thoughts about what info they'd like to see
> here? Are the logs sufficient?
>
Yes, it would be relevant only in case, if system have the ability to to
announce previous translator if original changes, but this is not
necesery.
> 3. The wordlist can be properly integrated. With an interface to
> add/remove words. A bit trickier since you don't necessarily want
> everyone to be able to do this. Which leads to the following:
>
I thing, that l10n teams has their own wordlist for any kind of
translations. Terms, names etc. must also correspond with for example
debconf templates. DDTSS translator must not be a debconf translator.
> 4. User management. The current system had it bolted on but here is an
> opportunity to expand the possibilities. You can make access rights to
> submit/review changes/wordlists/translations, etc. This would
> naturally lead to "language managers" for controlling rules for a
> particular language. And superusers, for adding languages.
> Deleting/banning users, etc. What I'd like is some concrete proposals
> about would people would like to be able to do/configure. Or does
> everyone like it just the way it is?
>
That's great. All as you wrote. External reviews and e-mail interface in
most cases spam, or translations from not experienced and anonymous
users, that translates a very few descriptions and mostly with bad
results, and it's very hard to find and control/correct this.
> 5. I wasn't thinking of altering the translation process itself. It
> has I think evolved to the point where it works well and I don't want
> to make major changes here. But perhaps tweaks can be accommodated.
>
+1 Only small suggestion further to above. Language
manager(s)/coordinator(s) should have the final word in approving
translations.
> 6. It's actually trivial to extend this to the whole DDTP website. You
> can get the current website as a drop-in replacement, except with
> templates for the HTML. This may be interesting for people who would
> like to restyle the website but don't feel like digging through the
> perl code to do it.
>
> I am currently working in a local git repository but if people are
> interested I can post the code somewhere. The idea is anyone can check
> it out and with a few commands have an installation running locally.
>
> So, fire away!
> --
> Martijn van Oosterhout <kleptog@gmail.com> http://svana.org/kleptog/
>
>
--
Michal Šimůnek
Reply to: