Eszett and IDNAv2 vs IDNA2008

[Andrew Sullivan]

On Thu, Mar 12, 2009 at 11:13:55AM -0700, Erik van der Poel wrote:

> I didn't say anything about a new RRTYPE. I suggested a convention in
> authority names to indicate a display preference. There may well be
> reasons not to have so many different authority names. (I.e. one
> authority name per name-with-display-preference.)
> 
> You and John may call this a "distraction". Others might call it
> "considering all possibilities".

And how, exactly, are you going to get this "conventional authority
name" back with the resolution of the A-label-including name, without
inventing a new RRTYPE that includes special processing?  Or is your
plan to just ask the DNS, in separate transactions, for these
different names?  If it's the latter, we're even further from a
solution.  Eric B-W and I were just discussing this on-list today.

This is "considering all possibilities" in the sense that "all"
includes "even the things that will occur after heat death of the
universe".  I get the reasoning -- and I myself have asked whether
what we're really trying to do is solve the problems of DNS, and
whether DNS2 is a better goal.  But if this WG is to produce a useful
spec any time soon (where "soon" sounds like it is measured in months,
not years, and surely not decades), "all possibilities" need to be
whittled down to the ones we really can do.  And the proposal you seem
to be chasing, I am sure, does not fit in that category.

A


-- 
Andrew Sullivan
ajs at shinkuro.com
Shinkuro, Inc.