Re: [idn] report of the straw poll

["Eric A. Hall" <ehall@ehsco.com>]

Keith Moore wrote:
> 
> I'd be curious to know who would support idna-nameprep-ace in the
> short term with the assurance that a native utf-8 query protocol
> would be worked on after idna-nameprep-ace went out the door.

Can you turn that statement around and still be happy with it? Rather than
saying, "we commit to do something better in the long term" can we say
"ACE is the legacy interface component of XXXX long-term strategy"? IOW,
can we define a long-term strategy that ACE plugs into?

I've advocated EDNS as the best technical solution to the problem, but I
also recognize that the use of UTF-8 data breaks older systems (I favor
EDNS because it forces the break, and because it offers the broadest
future features). If we can agree that EDNS (as an example) is the
appropriate long-term solution AND begin designing around it, AND then we
include ACE as a way to provide fall-back backwards compatibilty to legacy
systems, then we have a holistic solution.

Let's take email as an example of this principle. Just for arguments sake,
we agree that EDNS is the way to go and begin desingingh an ESMTP
extension to negotiate the use of IDN, and specify the use of ACE if the
negotiation fails.

In that example, ACE provides necessary backwards compatibility to a
long-term approach, rather than being a destination in its own right. We
can design for long-term features and capabilities from the start, without
asking designers to do everything twice (once for ACE, once for something
else).

FWIW, I did not vote in the straw poll because neither option represented
my feelings on the matter.

-- 
Eric A. Hall                                        http://www.ehsco.com/
Internet Core Protocols          http://www.oreilly.com/catalog/coreprot/