[Localization] Localization: What works and what does not (Sayamindu Dasgupta)

[Carlo Falciola]

- maybe line in source code is too fine grain, and is not really easy to understand the context from unknown code... I can actually already do it by myself in "very" special cases by grep, but takes a lot to derive what the code snoppet is doing... maybe to automatically get the name of the class/object/procedure where the string lies may be a better hint.

- to know when and if strings are available in released code is extremely timesaving, you can start testing accordingly...

ciao carlo


--- Ven 18/12/09, Sayamindu Dasgupta <sayamindu at gmail.com> ha scritto:

> Da: Sayamindu Dasgupta <sayamindu at gmail.com>
> Oggetto: Re: [Localization] Localization: What works and what does not  (Sayamindu Dasgupta)
> A: "Carlo Falciola" <cfalciola at yahoo.it>
> Cc: localization at lists.laptop.org
> Data: Venerdì 18 dicembre 2009, 22:09
> Hi Carlo
> Thanks for the feedback. I agree that context information
> is a tough
> issue. Do you think it will be useful if each message in
> Pootle linked
> to the particular line in the source code (so that when you
> click on
> it, you get to see the particular line in the code from
> where the
> string was extracted) ?
> Regarding "in place translation", Scott had some code lying
> around - I
> have been meaning to look at it, but haven't had the chance
> to do so.
> 
> On a more immediate basis, I have some good news for
> developers. I
> managed to hack message checking in the lines of msgfmt
> --check for
> individual messages into Pootle [1]. This means that the
> occurrence of
> builds getting choked due to syntax errors in PO files
> will
> drastically get reduced.
> Thanks,
> Sayamindu
> 
> [1] http://people.sugarlabs.org/sayamindu/pootle_gettext_check.png
> 
> On Wed, Dec 16, 2009 at 11:25 PM, Carlo Falciola <cfalciola at yahoo..it>
> wrote:
> > My 1/2 cent:
> > 1. Context: many times is difficult to get the right
> term because the original one is out of context and the
> original language term has multiple translations with
> different meanings depending on context, so
> >
> > 1.1 adding context information, if is possible, ease,
> accelerate, and reduce errors.  (es. "play" in english has
> several reasonable translation in italian: as noun "gioco",
> "opera teatrale", ...;  as verb. "giocare", "suonare",
> "eseguire", "recitare"....)
> > 1.2 avoid big chunks of strings: it takes time to
> figure out from where any string belongs, many time could
> happen the same same strings get repeated in the software
> and it should belong to different translations because of
> the contexts..(es. etoys has 1 file, 27454 words 4412
> strings)
> >
> > 2. on the process side it would be useful to have back
> information on when the strings (or .po file granularuty)
> actually landed in releases.
> > (where they should be in GIT would be an intermediate
> useful information too)
> >
> > 3. probably the translator nirvana would be to be able
> to have the option to translate "in place" the strings that
> has more heavy space constrains in the GUI.
> >
> > ciao carlo
> >
> >
> >
> >
> > _______________________________________________
> > Localization mailing list
> > Localization at lists.laptop.org
> > http://lists.laptop.org/listinfo/localization
> >
> 
> 
> 
> -- 
> Sayamindu Dasgupta
> [http://sayamindu.randomink.org/ramblings]
>