At 1:24 AM -0500 11/6/03, John C Klensin wrote:
It was my understanding of the agreements when the IDN
documents were approved that, when they went to Draft, the
Draft versions would incorporate a better statement of
applicability and scope than the original versions and, in
particular, would incorporate the gist of the "IESG
Statement" on IDN applicability and missing pieces. That
agreement does not appear to be reflected in the text of the
new drafts.
Just before the problem statement in
draft-hoffman-rfc3490bis-01.txt, it says "The IESG issued a
statement on IDNA [IESG-STATEMENT]." I didn't want to put the
statement in directly because the IDNA authors didn't write
the IESG statement, and it could dilute the value of the
statement by making it appear to be something we did. However,
seeing your concern, I'll ask the IESG about what they would
prefer on this. I'm happy either way.
Since the purpose of IDNA is to deliver native-text
characters to end user applications and the presentation to
the user, there is a case to be made that the "interoperable
implementations" condition needs to be demonstrated with
actual, end-user-oriented, applications that deliver
non-ASCII characters to users.
Fully agree.
That is, interoperability between test environments that
can demonstrate the ability to prepare, code, and decode
strings is not sufficient to demonstrate that interoperable
and conforming implementations are possible.
That is being debated in other parts of the IETF right now,
but fortunately, it isn't an issue here.
I can't tell from the IDNConnect "final report" whether
that stronger condition was met by those programs but, if
it was not, some serious community discussion on this issue
is probably in order.
Right. We specifically didn't list the participants because
doing so made it easier for more organizations to test. Having
said that, I can certainly say that many of the participants
were testing "actual end-user-oriented applications".
In my previous message, I said:
Subsequent to the event, I have validated one (unnamed,
pre-release)
IDNA system passes all the tests that it should pass, and
fails all
the test that should fail, as specified in the test
description.
That implementation is still unnamed and pre-release, but it
is very definitely a "actual end-user-oriented application".
There are many other such applications available, although
some of them are useful only in particular regions.
--Paul Hoffman, Director
--Internet Mail Consortium