Re: New external dependecy request (ICU)

[Keld Jørn Simonsen <keld dkuug dk>]

On Wed, Jan 07, 2004 at 12:37:22PM -0800, Hideki Hiura wrote:
> > From: Keld Jørn Simonsen <keld@dkuug.dk>
> > On Wed, Jan 07, 2004 at 10:44:46PM +0330, Roozbeh Pournader wrote:
> 
> As far as I witnessed as a member of the U.S. national body, the
> biggest concern I have been hearing from the U.S. natinal body members
> on TR 14652 is its quality.

For its quality, then judge for yourself, as it is available in
http://www.dkuug.dk/JTC1/SC22/WG20/docs/n972-14652ft.pdf
Also look at the employment it has got in glibc locales.

TR 14652 is basically POSIX locales with a number of enhancements, 
and the text is taken directly from POSIX, and then enhanced,
so the quality of the text is mostly as good as POSIX.
I don't know if you consider this good or bad.
We got a lot of opposition for just chosing a POSIX-compatible
syntax, as most of the industry was (and still is, but not for long:-)
not running POSIX-compliant systems.

> It's been too immature for so long and
> dragging so long and national bodies who are asked to continue
> debugging such immature draft TR did not find the way out.  Some
> national bodies seem not to care on its covarage and quolity as a
> whole, but the national bodies who really care about its implication
> on multilinguality and future integrity with the trend on
> multilinguality strongly against it.

Well, there has been a lot of time to improve it, and basically
what is in there is the same as it was when Ulrich Drepper implemented
it for glibc, I think it was back in 1996. I think it has served the
glibc community well since then. But there has been a lot of badmouthing
from other sources. And finally there are now competing specs, like ICU.

I have not seen many requests from the USA or Japan to improve the
14652 specification, and those that were proposed were almost all approved.
> 
> > Another thing is that the US national body in this respect is very
> > dominated by Unicode people, which has some tendency to just want
> > their own specifications.
> 
> I do not find any evidence that this statement is true.
> 
> It is true that many U.S. national body members are also Unicode
> Technical Committee voting member, as I am, however, it does not
> necessarily mean that they just want "their own specifications"
> even if there is any.
> 
> As Unicode technical committee has been clearly communicating that
> they do not have a plan to create "their own specifications" beyond
> Unicode Standard itself, which is just a coded character set.
> 
> Which "their own specifications" are you refering to??

Well, the Unicode standard and its associated specifications.
I once proposed that we included the character classes 
from Unicode in 14652, but that was not accepted by the Unicode people.

So I feel, but this is only a feeling, that the Unicode people were only
interested in developing their own specifications, so that this could be
the best, and in an embrace and enhance way they then would promote
their own standard over the ISO standard. This strategy has amongst
other things resulted in the IETF now referencing the UTF-8 from Unicode
as the normative definiton of UTF-8, although ISO has an almost
identical specification, and UTF-8 was an ISO invention (actually it was
a POSIX invention, done for AT&T plan 9).

Best regards
Keld