Re: Encoding polytonic Greek

Date: Fri Aug 27 1999 - 16:58:04 EDT

       Thanks for the info. For the particular situation (a temporary
       use), I've chosen to follow practice B, but I've told the user
       that in the long term this may not be the best solution. I'm
       still interested in hearing more debate on this in order to
       know how to make that longer-term decision when the time comes.


       From: AT internet on 08/27/99 03:22 PM

       Received on: 08/27/99

       To: Peter Constable/IntlAdmin/WCT, AT
       Subject: Encoding polytonic Greek

       Peter Constable asked:

> - U+1F0C (capital alpha w/ psili and oxia) has a canonical
> decomposition of 0391 0313 0301
> - U+1FCE U+0391 has a canonical decomposition of 1FBF 0301
> which has a compatibility decomposition of 0020 0313 0301
>Q: Which approach should be considered preferable?

       Greek users have always used both practices for input of
       diacritics+capitals. However, practice A (U+1F0C) is the rule,
       while practice B (U+1FCE U+0391) is th
       exception, caused by font-specific problems (lack of accented
       capitals in some codepages/printer fonts or ugly glyphs in
       low-res fixed fonts). I would go for practice A, but leave
       option B available for input if possible. For example, the
       texts in <> use practice B.

       and Rick McGowan commented:

> The clearest way to encode polytonic Greek text is to avoid
       using all the
> extended precomposed stuff. Use fully decomposed sequences.
> Some people might have a different opinion, of course.

       Well, it is true that the tendency in Greece is to use the
       precomposed character
       of Greek and Greek Extended and not decomposed sequences. For a
        number of reason
       I don't see decomposed sequences for Greek implemented in the
       near future (at least in Greece and Cyprus where the large user
       base lives).


       Constantine Stathopoulos,
       Iris Media Internet Solutions.

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:51 EDT