From: Hans Aberg (haberg@math.su.se)
Date: Wed Feb 02 2005 - 13:22:48 CST
At 18:31 +0000 2005/02/02, Peter Kirk wrote:
>>So, it does not matter, from this theoretical point of view, how the Unicode
>>characters are assigned to non-negative integers. ...
>If it doesn't matter, why don't we just stick with the current mapping
>of abstract characters to code points? After all, it really doesn't
>matter that a small range (the surrogate range) has been arbitrarily
>excluded.
I do not make myself any judgements of this. But I got some email where
folks speak about a variety of different character and numbering schemes. I
think also that the topic will evolve, so that people will start
experimenting with different numbering schemes and perhaps more efficient
character sets as well, fulfilling different needs. I merely indicate method
how such a thing can be done fairly easily, by merely pushing concepts like
characters, character numberings and encodings, already present in Unicode,
a bit further to become even more independent.
Once I have started walking down this path, it strikes me that a most
efficient numbering, according to some need, may have lots of spaces in it:
Instead of trying to cover a specific range, one tries to group the
characters more logically. In the process, it might help to have a lot of
holes in the character numbering domain.
Hans Aberg
This archive was generated by hypermail 2.1.5 : Wed Feb 02 2005 - 13:25:50 CST