Karlsson Kent - keka wrote on 2000-10-10 16:31 UTC:
> > The ISO TR 14652 transliteration mechanism (which is already partly
>
> There is no such TR!!! There have been drafts for such a TR, and I guess
> the editor is still working on it. But there is opposition to progressing
> 14652 any further. It was demoted from standards track to report track
> to make it less of an issue. However, it is not a TR yet, and it might
> never become a TR.
I myself have problems with parts of these drafts (it contains lots of
utterly useless stuff on postal address formats, etc.), but the general
idea of having a transliteration mechanism (automatic wide string
substitution applied before the actual multi-byte encoding in the
multi-byte encoding functions) seems to me highly useful and desirable
for a variety of purposes. There certainly are ways in which the precise
wording in the document on how it should be implemented can be made much
more precise.
A fundamental problem that I have with transliteration is that I might
want to restrict its application narrower than just the locale. For
example, I am happy with the idea that LC_CTYPE sets my character
encoding for everything (files, filenames, terminal I/O, environment
variables, etc.), because think/hope that all this will go to be UTF-8
soon anyway. However, I might not want to have transliteration applied
everywhere, because it is a lossy conversion process. Transliteration is
fine for texts at the end of the display pipeline, but it might be much
less desireable to file I/O (depends on the application context
however). A facility that allows the application to selectively activate
and deactivate transliteration would seem sensible here. This could be
done as a stream attribute, but it could also be done by separating
transliteration out of LC_CTYPE into LC_TRANSLIT as Keld suggested.
Keld, what is the current status of the draft, where does the latest
working copy live (URL?), etc.? Can we make this issue a bit more
transparent for the members of linux-utf8? What are the 14652 mailing
lists that we can join? Thanks!
Markus
--
Markus G. Kuhn, Computer Laboratory, University of Cambridge, UK
Email: mkuhn at acm.org, WWW: <http://www.cl.cam.ac.uk/~mgk25/>
-
Linux-UTF8: i18n of Linux on all levels
Archive: http://mail.nl.linux.org/lists/
