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

Re: Spanish translation of maint-guide

On 4 December 2010 18:11, Innocent De Marchi <tangram.peces@gmail.com> wrote:
> I just saw that the translation has been replaced by another
> incomplete today. I do not understand why this substitution.

Hi there,

The reason I inadvertently changed the translation is because I was
provided an alternative one by members of the Spanish translation team
in the team mailing list. When I was going to commit, I reviewed the
PO header and found this:

# Copyright (C) YEAR Free Software Foundation, Inc.
# This file is distributed under the same license as the PACKAGE package.
msgid ""
msgstr ""
"Project-Id-Version: \n"
"POT-Creation-Date: 2010-12-01 22:35+0900\n"
"PO-Revision-Date: 2010-12-01 22:43+0900\n"
"Last-Translator: Osamu Aoki <osamu@debian.org>\n"
"Language-Team: \n"
"Language: \n"
"MIME-Version: 1.0\n"
"Content-Type: text/plain; charset=UTF-8\n"
"Content-Transfer-Encoding: 8bit\n"

This was clearly an indicative of an *improper* PO file for many
reasons. And I (wrongly) assumed the PO file was not current and
overwrote it. As it turns, it had a 100% finished translation, even if
it does not follow the rules for translations defined by the Spanish
Translation Team.

Some facts:

a) I originally translated this document [1] 11 years ago, but have
not worked with it for the last 3 years.
b) There was ongoing work at the debian-l10n-spanish mailing list to
update this translation
c) I was notified of this translation by personal email, but alas, I
missed it and focused in the debian-l10n-spanish work
d) Bug #604710 was open but the debian-l10n-spanish was not notified.
e) Osamu, being document maintainers (but non-native Spanish speaker)
commited the file provided into SVN.

So we actually had both an independent translator (Innocent De Marchi)
working outside of the main group and people in the translation group
that were (at a slower pace) updating the transltion. The first
translation was finished and was not approved by *any* of the previous
translators before it was commited.

In any case, I appreciate the work done by Innocent De Marchi to
update the translation. As well as Osamu's work to convert it to the
PO format and commit it. I have reverted the change I made in SVN to
the previous translation and added a *proper* PO file header. I'm also
starting to review it and will ask the translation team to do so too.

However, I would like to add some comments here on how, from my point
of view, translations for Debian-specific documentation should be
managed upstream (and I'm adding debian-doc to get this also across to
other documentation maintainers):

1.- When a document maintainer gets a new translation for a document
that was previous translated he should at least contact *both* the
previous translators as well as the language team mailing list.

2.- Maintainers of documentation should ask translators  who contact
them if they are working within an exisiting  team. If not (which
might occur just because they don't know the team) they should be
asked to work within the team, whenever possible.

3.- Commits to translated material should be consensuated with the
team or previous translators. At the very best, if a known volunteer
of the translation team or a coordinator exists and has access to the
repository, he should be in charge of the update.

If done by the documentation maintainer, and  the documentation
maintainer is not knowledgeable of the language and the translation
rules followed by the team he should only approve updates after
waiting for a reasonable time for an answer.

As a documentation maintainer myself of a few Debian-specific
documents, I've followed these rules myself in the past and would
appreciate if others do so too.

I know that 100% translation is the holy grail for some, but for some
others we would really have high quality translations and that means
that peer review is required. Having a standing member of the team do
commits to the repositories, instead of a maintainer which is *not*
acquainted with the language (or, even, the new translator himself),
should guarantee that there is at least a quality review gate before a
translation from a new contributor is accepted.

This should prevent issues like the ones that happened with this
translation: two independent works being done in parallel with
different quality levels which leads to confusion, clash and some
general chaos.

Best regards


[1] Unfortunately, the CVS history was lost with the migration to SVN,
but can be found here:

Reply to: