RE: PDUTR #27: Unicode 3.1

From: Peter_Constable@sil.org
Date: Mon Jan 22 2001 - 12:49:02 EST


On 01/22/2001 10:28:42 AM Mike Lischke wrote:

>This is very interesting. How would you encode, say:
>
>2F9EA;CJK COMPATIBILITY IDEOGRAPH-2F9EA;Lo;0;L;927C;;;;N;;;;;
>
>to UTF16?

The workings of UTF16 are well known and clearly defined on p. 45, section
3.7 of TUS3.0, specifically in definition D28. (Available online at
http://www.unicode.org/unicode/uni2book/ch03.pdf.)

>I'm currently extending my Unicode library for Delphi and I wonder how to
>deal with all those new characters which are > 64K? Particularly the
>character properties causing me headaches because of the memory
>requirements. Using a linear search through the categories and a binary
>search through all ranges in that category is one solution to minimize
>memory footprint but it is also very slow. Do you know any better
>implementation which I could look through to learn a better way?

These are the kinds of issues Mark Davis discussed in his "Bits of Unicode"
presentation at the last Unicode conference (IUC 17). See
http://www.unicode.org/iuc/iuc17/papers.html and look for session B3.

Peter



This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:18 EDT