Re: [idn] UTC feedback

[Keith Moore <moore@cs.utk.edu>]

> In a global system such as we are discussing
> here, one needs a global ruleset for normalisation/canonicalisation,
> not different rules for France and Viet Nam.  Different rules that
> are country-dependent mean that the same input could map to different
> values of IDN, which would not be interoperable.

if you expect client-side normalization/canonicalization to do the
job by itself, then I agree wth you.  

but I don't see why it would fail to interoperate if the comparison
against different spellings were done on the server.  this is similar 
to what UTC is proposing except that instead of maintaining separate 
sets of RR records for each different spelling, the servers for a 
specific zone are configured to recognize the query for each of those 
different spellings and return something reasonable.   this could even 
be configurable on a per-label basis (different comparison rules for
different labels) as long as collisions between labels within a 
zone could be detected.

the real issue is how to make this work with DNSSEC.

Keith