Re: [perldoc2-developers] Meet the team

[Nicolas <nek...@gm...>]

Hello all,

I'm resending this mail, with the "From" address subscribed to the list.
Please ignore my other answer (no need to moderate it).

On Thu, Nov 09, 2006 at 01:57:46PM +0100, Joergen W. Lang wrote:
> Hello and welcome everybody,

Thank you J=F6rgen for setting up this project.

> Nicolas Fran=E7ois, aka 'nekral':
>=20
> *Nicolas*, also know as 'nekral' on sourceforge, is the current=20
> maintainer of po4a, a tool that converts documentation from multiple=20
> formats (POD, to mention just one) into a gettext-compatible format and=
=20
> back.
>=20
> perldoc 2.0 uses po4a extensively to convert the original .pod-files=20
> into gettext-pot for ease of translation and mainteneance.

I've made a Makefile to call po4a and update the POTs, POs, and
translations.

        http://nekral.homelinux.net/perldoc2/

You can have a look at it and see how it works (I took the translations
from the existing French translations

I also have the French POs for version 5.6.2 and 5.9.4, but I would like
to make tests for merging from a version to the other (I did it manually,
but would like to do it in the makefile). The makefile also lack a rule t=
o
detect strings translated differently from one PO to the other.

The Makefile has rules like all-pots, all-update-po, all-translate

update_pots_fr, translate_fr, update_po_fr
or even update_pots_fr_5.8.8, translate_fr_5.8.8 and update_po_fr_5.8.8

It can also show statistics: all-stats, stats_fr and stats_fr_5.8.8

The versions and languages are auto detected.

I will commit this proposal if we agree on the hierarchy it creates in th=
e
'translations' directory.

> Paul Gaborit, aka 'gaborit':
>=20
> I am very happy to see *Paul* join the team. Paul seems to be the only=20
> survivor of the french translation project which has started some years=
=20
> ago as a joined effort of the french perl newsgroup. He still maintains=
=20
> the french repository.

In addition to helping in the infrastructure, I will also happily
participate in the translation and review process of the French
translation.



Answering Herbert's question, po4a extracts the strings from the English
pod, put them in a PO file, and once the PO file is translated, it can
reinject the translated strings in the POD structure (hence generating a
translated POD).

It eases the maintenance of translations because it automates the
update of a translation from one version to the other.

When po4a is run on a later version, if some of the extracted strings are
the same, then these strings are already translated. If they are just a
little bit different, then they are marked as "fuzzy" (which means the
translation needs to be reviewed), and new strings are also detected.

It can also ease the translation because it can detect strings used in
different files.


Other advantage inherent to the use of a PO file are:
 * it is a common format for translators,
 * I like it for reviewing the translations (you do not only review the
   translation looking for typos, you check that the translation matches
   the English original text).
 * You can translate parts of a document (the translated document will
   then be a mix of English and translated text). This means that if a
   translator vanish, we can still distribute an accurate documentation.
   (This point is configurable; some people do not like it)

Another note: if an existing translation exist, po4a can try to reuse it
to initialize the first POs (that's what I did with the French
translation to test the infrastructure).

I'm stopping here my po4a advertisement.

Cheers,
--=20
Nekral