Re: [idn] UTF-8 as the long-term IDN solution

["Adam M. Costello" <amc@cs.berkeley.edu>]

"Eric A. Hall" <ehall@ehsco.com> wrote:

> Applications store/display names that they are given by DNS, so DNS
> must provide them with the data in the appropriate format.  Although
> the on-the-wire protocol can be ACE as you suggest, the resolver still
> needs a way to remember the desired disposition format of the query
> results.

Agreed.

> I suppose that it's technically feasible to ask resolvers to maintain
> state,

They already do.  gethostbyname() blocks and retransmits the query until
the response comes back or the time limit is exceeded.  That's quite
stateful.

> > and creating a new query format just for UTF-8 is certainly more
> > trouble than it's worth.
>
> No.  The EDNS message is already defined and is required/used for
> things like A6 and DNAME already.  This would just be another usage of
> that message.  It also allows us to resolve the sizing problems.

Okay, creating a new query format just for UTF-8 would be more trouble
than it's worth, but it's not necessary, because the new format already
exists for other reasons, and we might as well specify that it uses
UTF-8.

Whether A6 and DNAME are useful enough to motivate wide deployment of
EDNS remains to be seen.

Marc Blanchet <Marc.Blanchet@viagenie.qc.ca> wrote:

> fyi, there is a draft for idn over edns (draft-ietf-idn-idne-xx.txt)

That looks good (after a cursory reading).

AMC