Potentially redundant context rules

[Chris Wright]

I agree that everything that has been said below is factually correct.

The context rules prohibit Arabic Indic Digits (which have BIDI property AN) from co-existing in a u-label with Extended Arabic Indic Digits (which have BIDI property EN)

BIDI says that ANY DIGITS with BIDI property AN cannot co-exist in a u-label with ANY DIGITS with BIDI property of EN, so the context rules are a redundant subset of the BIDI rules! (Which is why we are arguing for the BIDI rules to be the ones that stay)

I accept that the two rules are not strictly identical, however the argument that the context rules allow ASCII digits to co-exist with Arabic Indic Digits (AN) is irrelevant because as soon as the label contains Arabic Indic Digits it would be subject to the BIDI rules and thus the digit mixing in the label would not be allowed anyway.

Thanks
c.

-----Original Message-----
From: Patrik Fältström [mailto:patrik at frobbit.se] 
Sent: Wednesday, 29 July 2009 10:09 PM
To: Matitiahu Allouche
Cc: idna-update at alvestrand.no work; Chris Wright
Subject: Re: Potentially redundant context rules

On 29 jul 2009, at 13.48, Matitiahu Allouche wrote:

> Patrik Fältström asked:
> "What about the other way around?
> Is there anything that is covered by Tables context rules that is NOT 
> covered by the Bidi?"
>
> Sure!  The context rules for ZWJ and ZWNJ, Geresh and Gershayim have 
> no equivalent in idnabis-bidi-03.txt.

This discussion is about the digits.

> By the way, even the rules for digits overlap between the 2 documents 
> but are far from identical:
> - In the Tables document, only Arabic-Indic digits and Extended 
> Arabic-Indic digits are mutually exclusive (but can coexist with 
> regular digits in the same label).
> - In the Bidi document, regular digits (U+0030..U+0039) are mutually 
> exclusive with Arabic-Indic digits, while there is no explicit mention 
> of Extended Arabic-Indic digits, so by default they will be handled 
> like regular digits since they have the same Bidi type (EN).

Ok, so this seems to imply the rules for the digits are definitely not the same or redundant, as you say Chris?

    paf

> If we call EN the regular digits, AN the Arabic-Indic digits and XN 
> the Extended Arabic-Indic digits, the combinations
>
> - EN and AN is disallowed in Bidi and allowed in the Tables
> - EN and XN is allowed in both documents
> - AN and XN is disallowed in both documents
>
> Seems to me that some more work is needed here.
>
> Shalom (Regards),  Mati
>           Bidi Architect
>           Globalization Center Of Competency - Bidirectional Scripts
>           IBM Israel
>           Phone: +972 2 5888802    Fax: +972 2 5870333    Mobile:  
> +972 52
> 2554160
>
> _______________________________________________
> Idna-update mailing list
> Idna-update at alvestrand.no
> http://www.alvestrand.no/mailman/listinfo/idna-update
>