Date: Tue Aug 19 2003 - 12:18:48 EDT
From: Rick McGowan [mailto:email@example.com]
Sent: Tuesday, August 19, 2003 3:43 PM
Subject: Re: Hexadecimal again (was RE: Clones)
My apologies if I have offended you (though I'm not quite sure how I might
have done). Let me reassure you that we're all "on the same side" here, as
it were. I know I'm a newcomer to this list, but I'm a programmer, and an
implementor. I want to promote Unicode, and I'm helping to make Unicode
Look at it from my point of view. I'm English. I don't speak Chinese,
Japanase, Bengali, Indian, or ANYTHING that requires me to use a non-English
keyboard. I have _absolutely no idea_ how the Japanese write hex. For all I
know, maybe the Japanese use Japanese symbols for hex? For all I know, maybe
the Indians use Indian symbols for hex? I don't know. And it's precisely
BECAUSE I don't know that it would be arrogant of me in the extreme to
assume that all around the world, in every nation, in every language, and
for every alphabet, everyone, without exception, uses Latin letters for hex
digits. Why should I make an assumption so arrogant? I know for a fact that
people don't all use the Latin digits 0 to 9, so what on Earth gives me the
confidence that they will all use A to F? For all I know, the absence of
Bengali hex digits from the current Unicode standard _COULD_ be an omission
which will be corrected in the next edition. I wish to write libraries of
code that will work for everyone, and which will continue to work for
everyone long into the future.
> > How do I write an algorithm which will convert Unicode hex characters
> > to hexadecimal which is guaranteed to work for all future versions
> > of Unicode?
> Use 0-9, A-F, a-f.
With all due respect, that algorithm doesn't even meet the CURRENT Unicode
I am very, very grateful to John Cowan, who emailed me to say that he is
going to propose adding to the standard the chart he posted. With that, this
quest effectively comes to an end, and I appreciate that.
You may think I'm some sort of moaning nitpicker, but these are real world,
non-trivial problems. The need for internationalisation demands that we drop
our assumption that Western standards will always be used by everybody. An
algorithm must rely on the tables alone. It cannot rely on anything else,
particularly not on an assumption of the universality of Western standards.
Thanks for answering, but please try to understand why I ask.
This archive was generated by hypermail 2.1.5 : Tue Aug 19 2003 - 13:10:13 EDT