[arabic-vip] Some (ignorant) questions about particular code points (was: Singapore)

[Alireza Saleh]

Andrew
The linguistic effect of ZWNJ is that it can completely change the word! This is the case for languages that need ZWNJ (e.g., Persian and Urdu both of which use Arabic script); as far as I know it is not needed in Arabic language.

Example from Persian:
[U+06AF, U+0644, U+0647, U+200C, U+0627, U+06CC] means < a flock >
 [U+06AF, U+0644, U+0647,                  U+0627, U+06CC] means  < flowers of >

--
Alireza




On Jun 14, 2011, at 10:36 PM, Andrew Sullivan wrote:

> Dear colleagues,
> 
> Speaking as someone who understands IDNA and DNS quite well, but who
> does not speak Arabic, I have a couple question about the items below.
> 
> On Tue, Jun 14, 2011 at 10:18:17PM +0500, Dr. Sarmad Hussain wrote:
>> 
>> Thanks for your comments.  Yes, the optional combining marks refer to
>> diacritics or aerab.
> 
>> From: Manal Ismail [mailto:manal at tra.gov.eg] 
> 
>> -          Technical 
>> 
>> o   Required  combining marks - extra Unicode normalization
>> 
>> o   Optional combining marks - 
>> 
>> o   Joining characters - ZWNJ
>> 
>> 
>> 
>> Manal: Thanks for the useful summary .. Does 'Optional combining marks' in
>> bullet 2 above refer to 'Diacritics' ? 
> 
> IDNA2008 introduced the CONTEXT rules for permissibility in the
> protocol.  It also relies on Unicode, and in particular a character is
> not allowed if it isn't stable under Unicode Normalization Form K
> (NFKC).  I understand that some combining marks are likely to be
> eliminated by the NFKC stability rule, but I don't really have an idea
> of how many cases are left after the stability rule is invoked.  It
> would be very helpful to me to have such an idea.
> 
> Similarly, ZWNJ is ruled out except in the case of a CONTEXTJ rule.
> What I have not understood are the linguistic effects of the rule (see
> Appendix A section 1 of RFC 5892:
> http://www.rfc-editor.org/rfc/rfc5892.txt).  If someone could clue me
> in, that would be a big help.
> 
> Obviously, I'm not expecting a quick answer; I'm rather hoping to
> highlight an issue that I can see for implementation, especially at
> the top level.
> 
> Thanks and best regards,
> 
> Andrew
> 
> -- 
> Andrew Sullivan
> ajs at anvilwalrusden.com