[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-i18n-doc
Subject:    Re: kbabel in kde4
From:       Yukiko Bando <ybando () k6 ! dion ! ne ! jp>
Date:       2007-08-15 15:43:52
Message-ID: 200708160043.52602.ybando () k6 ! dion ! ne ! jp
[Download RAW message or body]

Gudmund Areskoug wrote:
> Hello Yukiko,

Hello Gudmund,

> Yukiko Bando wrote:
> > Chusslove Illich wrote:
> >>> [: Gudmund Areskoug :]
> >>> [...] to see if someone already thought of back-checking, to see
> >>> whether a term/string in the target language has been used as
> >>> translation for different things, not just that one source has the
> >>> same translation everywhere.
> >>
> >> This is a bloody excellent idea in my opinion :) Same as Kevin has
> >> mined the adjectives out of the i18n call in the code itself, we
> >> could mine this info out of the full language checkout. And not with
> >> too much effort at that.
> >>
> >> Before I or someone else gets onto it, does someone perhaps see a
> >> caveat with this approach?
> >
> > We use a same Japanese word for settings, configuration/configure and
> > preferences. What will happen to those translations? Get fuzzied?
> > There are some other English words that we have only one word in
> > Japanese...
>
> main thing with the feature is IMO that translators/QA people get
> alerted about it, to make an intelligent decision.

Okay, but for translations I would prefer to run a check locally when I
feel like it rather than get warnings continuously. For the current KDE4, 
I get 1596 fuzzy and 74957 translated messages after doing this.

1. Copy all PO files in l10n-kde4/ja/messages/ to a folder
2. Do msgattrib -o $file --translated --no-fuzzy $file for those files
3. msgcat -o translated.po *.po

Some of these fuzzies result from inconsistency but others have different 
translations with reason. The number of such fuzzies would decrease with 
more contexts added, but probably not completely. For example, I used 
different words for the "Addition" in kalgebra.po and kalcul.po because 
the latter is for younger kids.  

> Once the term is set, it might e. g. be possible to check it against a
> termbank and have the app not give an alert for what's already in the
> termbank. That termbank might e. g. in turn be a candidate for the same
> procedure, but at longer intervals.

I think it would be nice to be able to exclude certain terms per file from 
consistency check...

> > I downloaded Mircosoft's English-Japanese corpus from
> > ftp://ftp.microsoft.com/developr/msdn/newup/Glossary/ in May 2006 but
> > the folder had been emptied when I checked again after Vista was
> > out...
>
> Haven't checked in a while, but AFAIR they've still got those resources
> available online - somewhere. Possibly, they'll only be available in
> some proprietary format (e. g. Trados), but there are AFAIR free (even
> as in speech) tools that can handle/convert that to something free.

Chusslove found them in README.TXT in the Glossary folder. :)

For those who don't know, you can use recode to convert the CVS file to 
UTF-8 or whatever like this: recode ucs2..utf8 filename
It's another Japanese translator (Okushi-san) who told me all about this. 

Yukiko
[prev in list] [next in list] [prev in thread] [next in thread]