[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Li18nux Locale Name Guideline Public Review



On Tue, Jan 22, 2002 at 08:41:50PM +0900, Stephen J. Turnbull wrote:
>     David> But we're talking about locale charsets, the charsets that
>     David> every program can be expected to handle, the master
>     David> charsets for a user.
> 
> That may be what you are talking about.  That's not what the standard
> is for.

>From the Q&A document -

Q. Why are the standard values for codeset limited? [...]

A. [...] For example, UTF-32 and UCS-4 are not included in the standard
values because they are not used as parts of locale names. [...]

Most of the codepages listed aren't used as parts of locale names.
VISCII can't be used as part of a locale name.
 
>     David> Users should be able to expect that you can send a file
>     David> from one Linux box to another in the same locale without
>     David> having to recode it.
> 
> They should, but they can't.

Why not? The main exception is going to be the Euroizing nations, which
are split between ISO-8859-1 and ISO-8859-15. Everyone else has more or
less one charset standard.
 
> The point is that at present, there's no guarantee that any of the
> names they thought they knew for their locale will work on any other
> Linux system, even if the locale itself exists!  Thus the Locale
> Name Guidelines.

In practice, when does it not work? We can write up paper tiger after
paper tiger, but that does not a guarantee make.
 
>     David> in an area without multiple implementations and hence the
>     David> need for a standard.
> 
> But we do have multiple implementations.  The spew iconv --list
> presents you with is proof positive that we support "multiple
> implementations" of names for most character sets.  

iconv isn't relevant here. If it were, we'd be missing ISO-2202-JP,
UTF-32{BE/LE}, UTF-16{BE/LE} and probably a number of other charsets.
iconv is IANA's baliwick.

> I'm sure Uli
> Drepper would love to have an excuse to get rid of 3/4 of it, even if
> it means changing the main glibc name for ja_JP.eucJP to ja_JP.EUC-JP.

And every distribution would have to patch it right back on in, less you
screw up every user's login/enviroment files. Unfortunately, in the
computer world, I don't think you ever get permission to get rid of an
alias until you can personally prove no one's using it. 

And more of a note to self, when I write up a version of this for
comments - alphabets are not individual characters; for example, the
English alphabet has 26 letters, but English does not have 26 alphabets.

-- 
David Starner - starner@okstate.edu, dvdeug/jabber.com (Jabber)
Pointless website: http://dvdeug.dhis.org
What we've got is a blue-light special on truth. It's the hottest thing 
with the youth. -- Information Society, "Peace and Love, inc."



Reply to: