Re: Granularity of Unicode Conformance

Date: Mon Sep 27 1999 - 16:27:37 EDT

       Ed Hart wrote:
>From the perspective of a user of Unicode products,
       conformance to Unicode is very important. However, the degree
       to which a product implements the required Unicode functions
       and supports various scripts is an important topic not
       addressed by the Unicode conformance chapter.

>Yes, qualification by script is a better way to start
       analyzing Unicode "support"...

       [a lot of useful parameters applying to Arabic snipped]

       In some ways, this is a step in the right direction. Things can
       be more rather more complicated, though, depending upon what
       one is needing to do. For example, MS Office 2000 running on
       Windows 2000 promises to provide a very good level of support
       for a number of scripts, provided you are using those scripts
       for representing a limited selection of languages. If your
       language happens not to be in that list, you may experience a
       lot of frustration trying to use one of these apps.

       As Joe Becker correctly observed, metrics for support of any
       given script is not well defined, unlike Unicode conformance.

       To take the case of Arabic, for example, it's one thing to say,
       "Our app supports the Unicode Arabic range", but it's a little
       more to say that it does so for Classical Arabic, for Morrocan
       Arabic, for Mandinka (about 1,000,000 speakers in Gambia,
       Senegal, Guinea Bissau), for Hausa (Ajami style Arabic script),
       for Kanuri (>3,000,000 speakers in Nigeria, Ajami style), for
       Mwani (~100,000 in Mozambique), for Farsi, for Kurdish, for
       Azerbaijani, for Baluchi, for Kurmanji, for Turkmen, for Uzbek,
       for Kazakh, for Kirghiz, for Uyghur, for Urdu, for Sindhi, for
       Hindko, for Pashto, for Kashmiri, for Panjabi, for Balti, for
       Purik, or for Tausug, to mention just some that are written
       using Arabic script. Each of these might have different needs
       in terms of input methods, details of rendering (different
       ligatures used, different character inventories, different font
       styles, etc.), and certainly would have *very* differing needs
       in relation to things like spell- and grammar-checking, and
       also sorting (as Ed mentioned).

       The same issues are repeated for many other scripts, e.g.
       Cyrillic and Devanagari. Even to say that an app supports Latin
       script leaves a lot unknown: e.g. does it support diacritic
       positioning including stacking of diacritics, as needed by
       various languages and especially for IPA?

       There's a lot that we might want to know about all of these
       apps. The biggest question, though, is whether any one of us is
       willing to start compiling all of this info?


This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:53 EDT