CLDR data (Re: Comments on the IDNA2008 document)

JFC Morfin jefsey at jefsey.com
Tue Jan 13 17:20:16 CET 2009


At 16:37 13/01/2009, Mark Davis wrote:
>I agree with Harald about not using CLDR in the way you describe 
>(and I'm the chair of the CLDR group). What are and are not 
>characters in common use in a language is rather fuzzy. In CLDR we 
>try to deal with that issue by having a "core" list (eg A-Z for 
>English) and an auxililary list for characters that are not core, 
>but in common customary use in books, magazines, etc. It is 
>notoriously difficult to draw a bright line even there -- so you 
>don't want that baked into a protocol. That is best left up to the 
>registries, like DENIC.

Correct. This is documented in the IANA ccTLD tables. The problem is 
crossupporting every language entity as WTO TBT rules call for 
several millions tables. This was the ISO/NWIP proposed by UK/BSI and 
Debbie Garside that was denied. This is why some other solution is to 
be found at the end of the day.

The Internet architecture is based upon end to end information 
conservation. IDNA is a non-end-to-end entropic process based on two 
namespaces having different cardinals. To obtain a fully robust 
architecture we need to change the underlaying code, restore an end 
to end process or/and provide a controlled smart negentropic source. 
This is not what the WG is chartered for. So, let LC what we have, 
give a try to a general LC and get it published. So, we can 
eventually start working on something stable.

jfc



More information about the Idna-update mailing list