From: Patrick Andries (firstname.lastname@example.org)
Date: Fri May 20 2005 - 14:18:28 CDT
John H. Jenkins a écrit :
> On May 20, 2005, at 9:13 AM, Tom Emerson wrote:
>> And why not encode the Phaistos script regardless. Is the sheer
>> uniqueness that stands in the way of any encoding effort? Comparison
>> against thousands of random, accidental, historical Han variants seems
>> disingenuous. Why was it rejected back in 1997? The proposal isn't
>> even available on the Unicode site any more.
> Phaistos wasn't rejected for encoding; it was "not accepted." That
> was a minor semantic difference that I insisted on because it didn't
> close the door with quite as loud a slam.
> The problem with encoding Phaistos is that we just don't know enough
> about its repertoire and the identity of the characters to make any
> serious encoding, and there's very little need to do it, or at least
> that was the reasoning.
And I believe this is a very valid reasoning.
How can one respect the character-glyph model, essential if we don't
want Unicode to be a glyph repository, and code all the signs of a
"script" whose characters are not known ? Which also tends to prove its
very relative usefulness. I don't agree with the idea that we should
generously code and sometimes on purpose err by excess In other words,
while we know that tomes signs are dubious which should code inexistent
or duplicate characters just to be done with it. I believe the
traditional view should be respected : be prudent, delay doubtful cases
-- we are not in any hurry in most dubious cases --, and code what can
be ascertained to be a character.
As far as the numerous spurious kanjis, although their presence
discredits Unicode in the eyes of some uncharitable critics, they at
least don't pollute the character-glyph model, I believe.
This archive was generated by hypermail 2.1.5 : Fri May 20 2005 - 14:19:34 CDT