From: Philippe VERDY (verdy_p@wanadoo.fr)
Date: Sat Mar 19 2005 - 17:42:41 CST
> De : "Jony Rosenne"
> > -----Original Message-----
> > From: unicode-bounce@unicode.org
> > [mailto:unicode-bounce@unicode.org] On Behalf Of Mark Davis
> > Sent: Saturday, March 19, 2005 10:03 PM
> > To: Doug Ewell; Unicode Mailing List
> > Cc: Ahmad Gharbeia
> > Subject: Re: Incorrect names for Arabic letters
> >
> >
> > >Sorting of character data generally
> > > should not be done simply by comparing code point values.
> >
> > Make that always. I don't know of a single language that uses
> > absolute code
> > point order.
> >
> > ?Mark
>
> Absolute code point order happens to work quite well for unpointed Hebrew.
And quite well also for *basic* modern English (but not for the whole modern English language that also contains non ASCII letters: some wellknown words with accents, notably in American English that includes words imported from other languages like French and Spanish, or with traditional English that has supplementary base letters, sometimes still used in Britain...). Try using codepoint order and it will only work well with technical English or international English...
(May be this works for Swahili, or Bahasa Melayu, or Bahasa Indonesia, which still only use basic ASCII letters in their modern Latin transcription; I have found no occurence of any accent or diacritic in those Latin-written languages; may be I'm wrong and they also exist, also because of imported words from Spanish, Portuguese, Dutch, French... or even with Latin transcriptions of Arabic words)
Definitely you need collation for all languages, unless you want to sort technical languages...
This archive was generated by hypermail 2.1.5 : Sat Mar 19 2005 - 17:43:54 CST