Lessons from the Chinese Second Level

["Dillon, Chris" <c.dillon@xxxxxxxxx>]

Dear colleagues,

The practice of the Chinese community at the Second Level seems to be an 
especially interesting model. I will outline key parts of it below with 
references to the report and suggestions for spreading them to other scripts:
- Three variants (just about as low a number as it gets) are allowed; two 
automatically from a single table used across the script, and one user-defined. 
Note that the "automatic" bit of this goes against 6.1.1.1. ("The approval of a 
variant TLD must not be automatic, but initiated upon the request of an 
applicant ...) but does promote consistency.
- The primary label and all the variants form an indivisible set. This should 
address many of the issues in 5.1. Internationally, it may be desirable for the 
set to be referred to using the A-Label of the primary label in case old 
software cannot handle U-Labels.
- As far as I can see there is no intention to use different LGRs for generic 
and country code 2LDs and TLDs. This approach reduces the work necessary in 
6.1.6.1. for ICANN to evaluate second level LGRs. If it is ever necessary to 
have separate LGRs at the second level for a script, a table used at all levels 
with a short list of characters such as the hyphen also available at the second 
and below would be better than maintaining two tables.  Some existing 2LDs may 
need to be grandfathered.

Regards,

Chris.
--
Research Associate in Linguistic Computing, Dept of Information Studies, UCL, 
Gower St, London WC1E 6BT Tel +44 20 7679 1599 (int 31599) 
ucl.ac.uk/dis/people/chrisdillon