Re: Translation of manuals (was: SES manual French translation)

[Eli Zaretskii <eliz@gnu.org>]

> From: Stefan Kangas <stefankangas@gmail.com>
> Date: Tue, 2 Jan 2024 20:38:27 -0800
> Cc: vincent.b.1@hotmail.fr, emacs-devel@gnu.org, rms@gnu.org
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >> Date: Tue, 02 Jan 2024 13:51:38 +0000
> >> From: Jean-Christophe Helary <jean.christophe.helary@traductaire-libre.org>
> >> Cc: Vincent Belaïche <vincent.b.1@hotmail.fr>, stefankangas@gmail.com, emacs-devel@gnu.org, rms@gnu.org
> >>
> >> > I meant markers specifically for translators, so that it would be easy
> >> > to detect chunk(s) of text that was/were modified, and update their
> >> > translations.
> >>
> >> Don't you think that would be a bit burdensome for authors? Or do you
> >> mean semi-automatically?
> >
> > If there are markers, then finding the chunks whose translations
> > should be updated could be scripted, yes.  It could be done with an
> > Emacs Lisp program, for example.
> 
> A "chunk" here means something like a sentence, or even smaller.  It is
> therefore impractical to add such markers to the source text.

If chunks were sentences, that would be impractical, indeed.  But I
very much doubt that this is a good idea for translating manuals
(message catalogs use short phrases because most messages in programs
are themselves short phrases or single sentences).  Translating a
sentence out of its context is much harder than translating a whole
paragraph, because different languages have different ways of
explaining technical issues, and a good translation might well require
rearrange or rewrite sentences.

So my initial thinking was that whole paragraphs, and sometimes more,
should be marked as a unit.  And that is much more practical than
marking short phrases or single sentences.