Re: Conversion of debian-history to Docbook
Hi,
Osamu Aoki <osamu@debian.org> wrote:
> Hi,
>
> On Sun, Mar 10, 2019 at 03:57:15PM +0100, Holger Wansing wrote:
> > Hi,
> >
> > some months ago, there has been the conversion to Docbook 4.5.
> > Now I notice that translations are in a very bad shape. :-(((
> >
> > Looking at German for example, that was 100% translated before the conversion,
> > and now I have 274 translated, 1 fuzzy, and 52 untranslated. While many of
> > the untranslated strings are rather long.
>
> That's because English content was updated but German wasn't updated.
>
> > Additionally to that, I noticed that copyright hints from the old files were
> > completely wiped out.
> >
> > This all is very bad!!!
> >
> > Why not do this better?
> > When I take the German po file from just before the conversion, insert it into
> > the current project and run "make po", I get a file with status
> > 184 translated, 134 fuzzy, and 9 untranslated.
> > So many fuzzies, but the changings are simply one of the following four cases:
> >
> > <em> -> <emphasis>
> > <prg> -> <command>
> >
> > <ulink url="http://www.debian.org/social_contract";>Debian Social Contract</ulink>
> > ->
> > <url id="http://www.debian.org/social_contract"; name="Debian Social Contract">
> >
> > insert of additional spaces
> >
> > Many of this can be done by simple find-and-replace.
>
> Did you look into the git changelog? I recorded what I did there.
>
> SGML to XML conversion isn't as simple as you describe to do it all
> automatically for all languages.
>
> If you had uncommitted changes, let"s discuss.
No, I had no uncommitted changes (and if I had, that would have been my own
fault.)
Don't get me wrong.
I didn't want you to do all language updates automatically, but there would
have been another (a better) way to deal with the translations:
Going the way I described, I managed to get down to 1 fuzzy and 5 untranslated
strings with the work of 1 hour.
And I will be able to do that for many other languages too.
>
> > So that does not justify to wipe away so much translated content IMHO!
>
> That's a side effect of the use of conversion script. It can be
> re-added manually.
The way I'm doing it, the content is still there, but only set to fuzzy.
I only need to do some small unfuzzying, but not translating the whole from
scratch.
>
> > Maybe I'm missing something, but if noone objects shortly, I will do the above
> > for all translations and commit the result into git.
>
> Please don't.
>
> > Will see how it goes, maybe I will also do some unfuzzying, to get translations
> > back to as good as possible.
>
> Post-xml conversion diff attached:
The diff shows the changes for English, which are perfectly ok.
But that's not the problem, see above.
Holger
--
Holger Wansing <hwansing@mailbox.org>
PGP-Fingerprint: 496A C6E8 1442 4B34 8508 3529 59F1 87CA 156E B076
Reply to: