RE: New draft Unicode technical reports available for review

From: John McConnell (johnmcco@microsoft.com)
Date: Tue Nov 10 1998 - 20:46:53 EST


Roman,

TR9 was never intended to be fast -- the whole purpose was to be as clear
and as close to the text description as possible. A commercial implentation
of the algorithm should never use the code as is, regardless of the
language.

Mark Davis has just released the revised text for the bidi algorithm. I'm
updating TR9 to match and verifying that it generates the same results.

John
-----Original Message-----
From: Roman Czyborra [mailto:czyborra@cs.tu-berlin.de]
Sent: Sunday, 08-Nov-1998 05:08
To: Unicode List
Subject: Re: New draft Unicode technical reports available for review

> we will now be posting proposed draft technical reports as they
> become available....

Thank you!

Looking at http://www.unicode.org/unicode/reports/tr9/, I find it
disappointing that the Reference _Implementation_ is in some byte code
that can only be executed by slow HumaniumPro :) 60 Hz neural network
processors instead of some more standard coding like C (ISO 9899).
Will that perhaps get added before TR 9 loses its draft status?

Don't get me wrong - I am very glad that the BIDI description has
finally made it online! This is already a good start for promoting
the BIDI function.

Playing with http://czyborra.com/unifont/mirrored also made me wonder
if Unicode's "mirrored" really means plain mirrored in right-to-left
context or something more intelligent: Is the little exponent "3"
sitting on the U+221B CUBE ROOT really supposed to be mirrored into a
round "E" instead of just mirroring the root it sits on?
Is the U+2232 CLOCKWISE CONTOUR INTEGRAL really supposed to turn
anticlockwise or isn't just the integral sign supposed to be mirrored?
And why is the negation stroke mirrored in U+2224 DOES NOT DIVIDE and
U+2226 NOT PARALLEL TO but not their common component U+0338 COMBINING
LONG SOLIDUS OVERLAY? And why are U+FE59 SMALL LEFT PARENTHESIS & Co.
not mirrored like U+0028 '(' and the CJK braces U+3008..U+3017?
What does the small punctuation have to do in fullwidth cells anyway?
And why are the small fullwidth symbols and wide Hebrew letters like
U+FB21 HEBREW LETTER WIDE ALEF & Co. listed as "narrow" in
http://www.unicode.org/unicode/reports/dtr11.html?

Cheers, Roman http://czyborra.com/unifont/



This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:42 EDT