Tamil Numerals in IDNA - Re: WG Last Call for Four Primary IDNABIS I-Ds

[Kenneth Whistler]

I concur with Andrew Sullivan's analysis of this request
and am not in favor of making a protocol exception for
this particular set of digits.

Gihan Dias is correct that the Tamil digits are not in
general use and that some of them have appearances which
are confusably similar to Tamil letters. But the same
argument could be made for *other* sets of traditional
digits included in the standard. For example, the argument
would apply to Malayalam digits. Or Balinese. And almost
every traditional set of digits in Brahmi-derived
scripts has one or more lookalike problems: Myanmar digit
zero versus Myanmar letter wa is a notorious example.

Making an exception for this set is not a direction which 
should be taken in the protocol documents, particularly not at this late stage
in consideration of them. These kinds of restrictions can
be handled by the registries, I believe.

--Ken


> On 2009-08-20, at 21:22, Gihan Dias <gihan at cse.mrt.ac.lk> wrote:
> 
> >
> >  0BE6..0BEF  ; PVALID      # TAMIL DIGIT ZERO..TAMIL DIGIT NINE
> > These characters are not in general use, and several of them are  
> > identical or similar to other Tamil characters (see attached image).
> 
> I believe that on previous occasion, the WG has rejected both the  
> premises "not in general use" and "confusably similar" as reasons to  
> make a code point DISALLOWED.
> 
> > I am aware that Tamil digits have the Unicode property "Nd", in  
> > common with other digits, and it will be difficult to segregate them  
> > on the basis of Unicode properties.
> >
> > Leaving it up to registries to not register names with these digits  
> > is unlikely to work, as registries would normally use the RFC as  
> > their reference.
> 
> The IDNA2008 document set explitictly urges registries to make policy  
> beyond the IDNA2008 documents. I therefore reject the above premise,  
> and I conclude that these characters should not become exceptions.
> 
> Best regards,
> 
> A
> -- 
> Andrew Sullivan
> ajs at shinkuro.com