Re: The translator credits

[Pablo Saratxaga <pablo mandrakesoft com>]

Kaixo!

On Mon, Oct 09, 2000 at 09:39:35PM -0400, François Pinard wrote:

> > > Also; all names that use a non ascii letter in their native scripts
> > > must be "translatable"; even for programmers etc.
> 
> > That a good remark.  The headers lines are considerd to be encoded like
> > the line in mail messages -- no 8 bit characters are allowed here
> > (msgfmt doesn't check it, though...).  Unfortunately, the PO file format
> > definition is a little bit hidden.
> 
> I lack the context of the original message, so I do not understand all
> issues under discussion.  I do not understand what is a "translatable" name.

I wrote that I think.
What I mean by "translatable" is that a name must be able to be shown in
the proper script, with the proper accents, when it is possible.
For example some person whose name is normally written in a cyrillic script,
should have its name in cyrillic for the translations in languages using
cyrillic script; but of course should have it romanized for others (and in
the original English strings).

> My feeling is that Korean PO files, for example, are written for Korean
> people much more than for Americans, and I would surely welcome Korean
> people to document them in proper Korean, including names.

That is.
And I also think that Korean programmers listed in the credits window should
also have their names properly shown.

> Some scripts
> have encodings which conflict with ASCII,

No, that doesn't exist.
All encodings have the 2*26 letters (even if some, like EBCDIC, may need
a conversion).

THe problem is the other way: only ascii is common to all encodings;
for non ascii letters, the source should show a romanized ascii version;
and the po file(s) must have the proper version, of the name.

> > > So if someone whose name is normally written in Arabic is listed in
> > > the credits, then his name should show in Arabic for the Arabic
> > > translation.
> 
> > Yes, but concerning the header lines, please Mime encode it.  The about
> > box (or something) should display the names decoded.

The header is not to be used by programs for anything (only possible
exception: retrieve the po charset); so it doesn't really matter. 
The header is to be read by humans; and humans are smart to catch situations
that computers are completly unable to.

> opinion about it.  So var, a PO file is not a MIME multi-part construction,
> and there is probably no strong need to have the PO file header having a
> neutral encoding like for MIME messages:

Indeed, that would only make it more complex, without need for it.

> I would be tempted to say that
> we could assume, by default, and for simplicity, that the PO file header
> use the same encoding as for the remainder of the file.

Yes.

-- 
Ki ça vos våye bén,
Pablo Saratxaga

http://www.srtxg.easynet.be/		PGP Key available, key ID: 0x8F0E4975