RE: [xliff] Uniqueness of language pair

[Yves Savourel <ysavourel@enlaso.com>]

Hi Stephan, all,

> Bryan gives a pertinent use case here.
> Since I was not able to follow the discussion
> from the beginning, this question may be redundant:
> * What is the rationale to not extend the standard
> to multilingual contents?
> I believe that nicely aligned multi lingual xliffs
> are beneficial for revision and proofreading tasks.

Personally I think an XLIFF document should address only a bilingual case because that is the smallest unit we have to deal with. For example getting a multilingual XLIFF from a CMS system would force a lot of processing before we (i.e. an LSP) could dispatch it to the proper translators. And same thing in its way back.

I think multilingual packages are perfectly fine, but they should be represented at the package level, in other words at a higher level than XLIFF.


> It can also be helpful for consistency when e.g.
> an Italian translator gets in his xliff previously
> translated and approved Spanish texts.

Having the user doing the Italian translation see the approved translation of the Spanish would need a major adjustment of the XLIFF tools to treat different translations as hints. Another way would be to have such translation hints provided using <alt-trans> and that is supported.

Cheers,
-yves