Re: Granularity of Unicode Conformance

From: Kenneth Whistler (
Date: Mon Sep 27 1999 - 13:51:40 EDT

Jony Rosenne asked:

> As I understand it, there is no requirement to render the whole repertoire,
> and if the rendering Hebrew and Arabic is not supported there is no
> requirement to handle bidi.
> I[s] this correct?


The basic requirement for Unicode conformance can be stated this way:
*If* a process interprets a particular Unicode character, it shall
do so in accordance with the specification of the standard.

So if a process does not interpret any Hebrew or Arabic (or Syriac
or Thaana or bidi formatting) characters, it is not interpreting any characters
with the strong right-to-left property or any specific bidi formatting
interpretation. Therefore, it can dispense with handling bidi, because
there would be context under which it would make any sense for that
process to implement it.

You can make the same argument, mutatis mutandis, for all kinds of script
support. If a process is not interpreting Indic characters, then it is
not required for it to understand the processing of virama, reordering
of vowels on rendering, etc.


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