From: Hans Aberg (haberg@math.su.se)
Date: Fri Apr 29 2005 - 10:49:29 CST
At 15:33 +0100 2005/04/29, Marion Gunn wrote:
> > ...there should perhaps be a notion
>> of input abstract characters. And so on. Unicode mixes all these
>> together in the notion of "abstract characters", without explicitly
>> clearly separating them.
>
>Also true.
>
>Would it help to simply (a) use the words/term 'graphic character' for
>the usuals/majority (i.e., characters which display on the screen as
>alpha and/or numeric for use in text), and (b) 'non-graphic' for what
>are sometimes now referred to in Unicode as 'abstract characters' (such
>as ESC)?
The notion of an "abstract character" is the comprehensive notion
that includes all other characters. I suggested the notion of a
"rendering character" as one that is put there specifically for
rendering purposes alone. A character such as "DIGIT ZERO" will
surely display something in normal usage, but it is still not a
graphical character. By contrast, any variation of the "0" used in
various scripts might be viewed as graphic in nature; alternatively,
one might view them as a combined character, the semantic "0" plus
the semantic information of a type of script that it appears within.
There is may thus be more than possible solution.
>I speak here only as a terminologist, and make this suggestion with no
>anticipation at all of its being accepted (only to help, if I can).
As such, I only suggest logical structure, as I am a mathematician. I
leave it to the linguists to come up with a good classification. In
general, I think these problems are tricky, and Unicode may need to
find a way to transition to a better classification. One simple way
to do that, if the number of available character numbers is
sufficiently rich, is to merely add these as additional characters
with a few more property fields.
-- Hans Aberg
This archive was generated by hypermail 2.1.5 : Fri Apr 29 2005 - 10:51:24 CST