CONTEXTO Proposal

Alireza Saleh saleh at nic.ir
Mon Jul 20 23:22:46 CEST 2009


I don't agree.
Prohibit mixing btween numbers was proposed just because the BIDI 
properties of ASCII  digits and Extend Arabic-Indic digits are EN and is 
different than Arabic digits which is AN .  I really don't know why now 
your are proposing that  2 different digit sets with same BIDI 
properties and complete different shapes should be prohibited from being 
co-exist in the label ?

Besides, Removing the EN,AN mixing prohibitions will be a very good move 
because the JAWI script can have their reduplication words. in Jawi 
script the functionally of Arabic digit 2 is like alphabets and it must 
be written complete the meaning of the word. E.g : rama٢  means  
butterfly.  they are using ASCII digits for writing the numbers.

This restriction was added to the  IDNA-BIDI document because it  
violates the label uniqueness condition  however I think 1) After fixing 
the UNICODE TR-9 report there won't be any problem with that and 2) the 
label uniqueness can be also violated as long as we have homographs  
but  I don't know why  the document says that  labels must look 
different than another label by re-ordering the  characters  but two 
labels can have the exact same look even by having different  characters.


Alireza


>     B. ARABIC-INDIC DIGITS
>
> The rule on these overlaps with Bidi, and would be simpler and more 
> appropriately moved there, since these are all only of concern with 
> Bidi processing. So remove:
>
>     * http://tools.ietf.org/html/draft-ietf-idnabis-tables#appendix-A.13
>     * http://tools.ietf.org/html/draft-ietf-idnabis-tables#appendix-A.14
>
> And change BIDI from:
>
> 5. If an EN is present, no AN may be present, and vice versa.
>
> To:
>
> 5. If an EN is present, no AN or EXTENDED ARABIC-INDIC digit 
> (U+0660..U+0669) may be present. If an AN is present, no EXTENDED 
> ARABIC-INDIC digit (U+0660..U+0669) may be present.



More information about the Idna-update mailing list