[Idna-arabicscript] Reasons for disallowing Arabic script digitmixing at the protocol level

Eric Brunner-Williams brunner at nic-naa.net
Wed Mar 11 21:44:41 CET 2009


Ram,

There is no technical substance to your note. You can correct that, but 
that's a wicked good reason not to send it to any IETF WG list, though 
to be fair, lots of junk is sent to some lists, including the IDNAbis list.

I'm more interested in making sure that the scripts bit of the UTC is 
well informed about characters adapted from Arabic, and either not yet 
in Unicode, or usefully designated in the IDNAbis tables draft, than I 
am in pushing an arbitrary, and rather vaguely argued partially correct 
fix for what, to be honest, is yet-another-glitch in the UTC's work 
product, the partial unification of Arabic Script, and the dangling 
directional semantic difference between two sets of L2R digits in a 
predominantly R2L script.

Ram, if you think I'm not a colleague you need only delete my address 
from the mailing list. It is as simple as that. In fact, I think I'll 
ask Aliraza or Dr. Shahshahni to delete my address from the list. I've 
enjoyed some very smart and considerate exchanges from individual 
members of the list, but the list itself is less useful to me than the 
people on it, and in my opinion, it hasn't produced much since last Fall.

Because transparency is important I copied both lists on my original, 
and I'm copying both lists on my reply. I'm deleting the portion that 
contains your note, which you can send to the IDNAbis list or not, as 
you please. Vint can always write "no squabbling", but silence would hid 
the difference between two views, promoting one, and concealing the 
other, which would not be useful no matter who held what view.

It seems to me there is a lot of work to be done, just on getting Arabic 
right, from the Unicode bidi bug of treating "." as punctuation when it 
is a label separator and should have no semantic value whatsoever, let 
alone provide a vehicle for directionality to leak across label 
boundaries, to typography, as Arabic isn't Latin done with a brush and 
ruler, but is multi-level and could be slightly less ugly than at 
present. Then there are the written languages which use some characters 
from Arabic, in Arabized Africa, and Asia, and these are more useful, 
and interesting, avenues of work than the one currently pursued to the 
exclusion of all others.

The bottom line is that the IDNAbis WG would be ill-advised to hum in 
favor of the "uniqueness and homogeneity of use of numerals" proposal as 
a globally scoped policy implemented as a protocol. It may still, you 
know how those things go, John was sold on the idea in Cairo, and 
Everson "supports this proposal completely", though why he does so is 
left to the reader's imagination, and I expect you'll sell your advocacy 
position to the best of your abilities.

Eric

Ram Mohan wrote:

(some text deleted)


More information about the Idna-update mailing list