[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Asking for a new pseudo package in the BTS: l10n-french



Let's keep on debian-i18n until we (translators) all agree on what is needed.

First of all, I changed my mind a bit, and now think that the right name for
the pseudo package to be created should be qa-<language>, with the language
name being written with all letters, not iso code. I guess that it will make
clear for maintainer what it is good for, and is easy enough to remeber for
users (IMHO).

On Mon, Jan 27, 2003 at 01:54:25PM -0200, Gustavo Noronha Silva wrote:
> Em Mon, 27 Jan 2003 14:06:09 +0100, Javier Fernández-Sanguino Peña
> <jfs@computer.org> escreveu:
> 
> > On Fri, Jan 24, 2003 at 04:03:39PM +0100, Martin Quinson wrote:
> > > As coordinator of the french translation team, we would like to ask for the
> > > creation of a new pseudo package in the BTS. It would be called l10n-french,
> > > and would be made to repport problems in french translations.
> > 
> > 	I'm not convinced this is a good idea for several reasons. The main
> > one being that the BTS already holds a place for translation bugs, but that
> > really depends on what the translation is _about_:
> > 
> > 1.- is it a debconf template? then the package that included it should be
> > bugged.
> > 2.- is it a document that it's included in a package? then the package
> > should be bugged (sample: doc-debian-fr)
> > 3.- is the document not available in the package? Then file a bug against
> > debian-doc
> > 4.- is it a manpage? Is it packaged with the program: bug the program, is
> > it not? bug manpages-XX (XX=es|fr|pt...)
> 
> That would work if we had official translators for every piece of software,
> doc or template we translate. Someone who could keep monitoring his "packages".

For that, we need that dpkg supports some translator-XX: fields, and
something like proposed in
http://lists.debian.org/debian-devel/2001/debian-devel-200108/msg02329.html
to be implemented. But for now, it looks like science-fiction to me.

> Our reality is quite differente... we have quite few translators for a lot
> of resources and a centralized way to receive and store translation bugs
> is something that would help us a lot.
> 
> Michael Brammer helped us a lot by creating the DDTP, it has a nice centralized
> bug reporting interface. But still, people need to know how to use the DDTP
> to "report a bug".

And bugs against the DDTP can't port on manpages or web pages, or did I miss
something?

> As reporting bugs to the BTS is already known to lots of users it would help
> us a lot if we had a l10n-pt pseudo-package, for example, with
> debian-l10n-portuguese as the maintainer.
> 
> So, a user could report:
> 
> "manpage X on package Y has 'usuários' spelled 'usurios'"

They could even repport against qa-french something like
« la page de manuel X du paquet Y orthographie 'errreure' au lieu de 'erreur' »

No other system but a specialized pseudo-package can allow the users to
repport translation bug in their own language.

> And the translators would quickly know about that bug, and we would have
> sane archiving of old and new bugs. If, instead, someone reports this to
> the package Y, we would depend on the maintainer doing work for us, and
> we already know not all maintainers are helpful on i18n matters =P.
> 
> We do not have manpower to monitor lots of bug pages.

Agreed. And even more, maintainers faced to a translation bug could reassign
the bug to the corresponding qa-xxxx package, then the translators provide a
patch and reassign it to the original package for patch integration.

> > I would like you, first, to read the proposal on how to bug documentation
> > packages (and translations) we've been working on at debian-doc. Please
> > read http://www.debian.org/doc/manuals/ddp-policy/ch-feedback.en.html
> >
> > This is not yet policy (it's a draft).

I did read this proposal, and it looks very good for documentation issues.
FYI, we already send ITT mails on the debian-l10n-french, but we never
trusted ourselves to send them as bug against the WNPP, since most of the
time those mails are in french, and we don't want to polute the already
heavily loaded WNPP bug page with our cruft. As result, we have no list of
made ITT, and I guess than some of them were forgotten.

I do not agree with the following sentence of your proposal: "Translators
<em>must</em> be subscribed to the BTS for the translated package versions."
It's already hard enough to recruit translators and reviewers, but if I ask
them to monitor a whole bunch of unrelated bug repports, it's an impossible
task. The remark of Gustavo about lacking manpower applies well here.

The graph at the end of the chapter seems to suffer of buggy
TAB-to-space-conversion.

Moreover, this graph do not take the updating of translation when original
is update into account. You may want to check po4a for that. Such a graph is
presented at:
http://www.nongnu.org/po4a/#graphical_summary

As a conclusion, I *think* you should remove everything about translation
from this proposal, since what you say here is not enough, and wait until we
agree on methods shared by all teams on this list before we try to document
them. 

I naturally tend to think that po4a may be a really usefull tool here. ;)
http://savannah.nongnu.org/projects/po4a

> > However, I _do_ see the need for a new BTS tag: 'translation'. This could
> > probably make package maintainer's life easier as well as PTS subscribers
> > to packages which only want to see translation bugs.
> 
> That would help, but a 'translation' bug could happen in any language,
> and it would suck having to browse translation bugs for all languages
> to find bugs reported on mine.

I already asked for this tag, back in october 2001. This is #114221.
But AJ was not convinced, and since no translators gave their advice, I
convinced myself that I was wrong. Now, a year and half after, I tend to
agree with Gustavo. Such tag could be helpfull, but that's not the
definitive solution. I guess i'll close my old bug repport.

Bye, Mt.

-- 
If the facts don't fit the theory, change the facts.
          -- Albert Einstein



Reply to: