Re: how should we localize locale names?

Subject: Re: how should we localize locale names?
From: Paul Rohr (
Date: Mon Mar 12 2001 - 16:04:59 CST

At 11:04 AM 3/10/01 -0000, Tomas Frydrych wrote:
>> I have to admit that I've briefly flirted with the idea of creating and
>> shipping a single Unicode font with just enough codepoints to render the
>> text in this dialog. Talk about hacks!
>This is really getting out of hand.
>(1) There are no Unicode fonts that have adequate coverage (this
>became clear from some of the recent emails from the CJK people
>on other issues).
>(2) Further, the only Unicode fonts that work at the moment are ttf,
>and we cannot expect that all people who use AW can/want to use
>(3) These fonts are huge. The most complete unicode font is
>probably code2000, which is 2.3MB in size; big engough to crash
>my machine under Linux (I have got 128M of memory) when I try to
>use it with AW.


Please don't think I'm *that* stupid. :-)

I was jokingly suggesting that we ship a custom-built font which had *only*
enough codepoints to render the names of each language in each language.
For example, to render:

  "English - United States"

would only require 14 discrete glyphs (if I did my math right), not all of
Latin-1. Clearly a full-coverage Unicode font would be huge, but the hack I
had in mind should theoretically be much more reasonable. For example,
that'd mean an upper bound of, say, 200 locales * 20 discrete glyphs = 4000

This is clearly a huge hack, and I labelled it as such. You'll also note
that I immediately discarded this thought experiment, because it'd set the
expectation (at the GUI level) that we *could* support those other languages
and charsets only to discover that in fact we couldn't (once you hit OK).

>Is it really not obvious that by refusing to do the one logical and
>clean thing, to translate the languages to the respective locales,
>you have already created yourself hours and hours of unncessary
>work, at the end of which you will get a user interface that does not
>produce a satisfactory result on any locale at all, and only on
>Latin1 it will make any sense at all? Please, please, go back to the
>very beginning and look at the whole issue afresh.

Please reread my original post. I was checking to see whether we could swap
translator effort (the N*N translation "problem") for hacker effort (a
one-time job). I identified the likely problem as charset issues, and
requested feedback on that.

Which you're providing in spades. :-)


This archive was generated by hypermail 2b25 : Mon Mar 12 2001 - 17:06:47 CST