Markus Kuhn writes:
> Glen Perkins wrote:
>
> > Is it likely, then, that POSIX compliance is going to *require* the
> > ability to handle UTF-8 transparently? Is it going to further require
> > UTF-8 as the default system encoding, or can three different
> > POSIX-compliant machines have three different default system encodings?
>
> At the moment, we have only two standardized locales: "C" and "POSIX".
> All the rest is manufacturer defined. These locales have been defined
> long before we had Unicode.
Well, most of the publically available POSIX locales are defined
on UCS.
> Suggestion:
>
> Let's add to the POSIX standard another standard locale called the
> "ISO" locale. This locale would specify
>
> - wchar_t variables always contain Unicode numbers
> - the multibyte character encoding is UTF-8 (this is what is used in
> text files, file names, environment variable name and content,
> text pipes, IPC ressource names, etc.)
> - string collation is in accordance with the new ISO sorting standard
> - the locale dependent date/time notation is from ISO 8601
> - the numeric notation is in accordance with ISO 31-0
> - and whatever else is fixed by international standards ...
>
> I.e., it would be a sort of international default locale that takes
> into account the level of international standardization that we have
> today as opposed to the ancient ASCII world in which the "POSIX" and
> "C" locales have been created.
>
> Anyone here who like this idea?
Well, something like it. It is done in ISO CD 14652
Keld
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:36 EDT