Follow-up to Monday's discussion of digits

[Martin Duerst]

At 23:50 08/11/28, Andrew Sullivan wrote:
>On Wed, Nov 26, 2008 at 08:26:52AM -0500, Vint Cerf wrote:

>> It appears that many of your colleagues find the ban on mixing of  
>> European, Arabic-Indic and Eastern Arabic-Indic either reasonable or  
>> even necessary.
>
>It seems to me this exchange is landing us squarely back into the
>issue of confusability, and we already adopted a principle that
>confusability rulings ought to be made by registries and handled by
>policy, rather than getting bound up in the protocol.  Confusability
>isn't enough for the restriction in the protocol.

I have to agree. The confusability between the two series of Arabic
digits is very high (out of 10 digits, 7 of them look the same and
mean the same thing), but I still (as I said earier) don't see the
need of prohibiting combinations of them in the protocol (except
maybe for bidi).

>John's argument for some restriction (in Minneapolis) was, I thought,
>dependent on the premise that a majority (or siginficant minority) of
>user agents do so many funky things with the encoding in between the
>keyboard and the wire that it's all but impossible to guarantee
>results without this change to the protocol.

There has been some promize that the arguments given orally in
Minneapolis would be written up. I haven't seen that writeup yet.
[If I missed it, please send me a pointer.] The above summary
also doesn't provide any relevant information. If keyboards
are unpredictable, what's the difference between a restriction
in the protocol and a restriction by a registry?



>Without that premise, I don't think there's any principled way to
>include the restriction in the protocol.  Therefore, we need to know
>whether the premise is true.

Agreed. And not only the premise, but also how the premise implies
the conclusion.

Regards,    Martin.


#-#-#  Martin J. Du"rst, Assoc. Professor, Aoyama Gakuin University
#-#-#  http://www.sw.it.aoyama.ac.jp       mailto:duerst at it.aoyama.ac.jp