Another Transition Plan Proposal

Andrew Sullivan ajs at
Fri Dec 11 16:39:51 CET 2009


John did a good job of outlining the issue (and making me not
hysterical), but it seems some additional details might be helpful.
Warning: more DNS details below.  This WG is the wrong place for
discussing how the DNS might be modified, but I think we had better
understand exactly what the technical capabilities are if we're
planning to recommend their use.

On Thu, Dec 10, 2009 at 04:33:12PM -0800, Gervase Markham wrote:

> Thanks for the clue. Various things have now dropped into place. I must 
> say I am somewhat surprised that the DNS has not evolved a way to say 
> simply "this domain is actually that domain" - DNAME seems close, but 
> not quite, that, as Vagellis mentioned in in what I now realise was his 
> request for just this ability.

Part of the reason this hasn't evolved is because the notion of "a
domain" from the POV of the DNS is not actually what many (most?)
users think it is.  (The DNS community has, I must day, done an
incredibly bad job at either explaining itself or else changing its
mind to meet the actual use of the DNS.)  There are some tricky
problems around the zone cut (the place where delegation happens) that
make this awkward.  There is in fact some discussion going on over in
dnsext, though, and if anyone is interested in the topic I encourage
you to join namedroppers & try to provide some information on the use
cases to the DNS protocol community.

> Can those measures be taken by the .com owner, or do they have to be 
> taken by the haß registrant?

Either.  DNAME is an RR like any other, and it can therefore appear at
just about any place in the DNS tree.  Around a zone cut, this means
that it can appear at the parent side or the child side.  For the
example, I'll stick with hassle/haßle, but use the versions one will
see in the DNS.  According to libidn, that oughta be

More information about the Idna-update mailing list