IDNAbis session 1 at IETF 74: mechanism/policy, looking for test materials

Vint Cerf chaired the session and revewed status/goals.
The current work ("bis") is called the 2008 design, as opposed
to the 2003 design. The 2008 design aims
to be Unicode version independent, for example.
There was also an interesting presentation on CJK details.
The presentation materials are (to appear?) in the IDNAbis
section of the IETF 74 meeting materials
https://datatracker.ietf.org/meeting/74/materials.html#wg-idnabis

I'll excerpt a bit from the jabber chat logs that
illustrate some points I picked up...
   http://jabber.ietf.org/logs/idnabis/2009-03-24.txt


I (re-)learned about the mechanism/policy divide:

[01:29:42] <dan.hoopyfrood> (I'm struggling to appreciate the aversion 
to policy issues in this design space. the security/policy issues seem, 
to me, to dominate. If policy is Somebody Else's Problem, is it clear 
who Somebody Else is?)
[01:29:55] <Andrew Sullivan> Dan: yes
[01:30:00] <Andrew Sullivan> the zone operator(s)
[01:30:03] <Andrew Sullivan> and maybe ICANN
[01:30:54] <dan.hoopyfrood> remind me who the .com and .cn zone 
operators are? (if you have bandwidth)
[01:31:19] <Andrew Sullivan> .com's zone operator is Verisign, and .cn's 
is CNNIC
[01:31:35] <dan.hoopyfrood> thanks
[01:31:41] <Andrew Sullivan> and the zone operator of crankycanuck.ca is me
[01:32:12] <Andrew Sullivan> (Note that lower-level zones are also 
important in this dicsussion -- important not to ignore that)
[01:32:45] <dan.hoopyfrood> so it's up to verisign whether to allow 
p<alpha>ypal.com to be registered. I guess that makes sense.
[01:33:30] <Andrew Sullivan> that's the idea, yes
[01:37:15] <dan.hoopyfrood> (a quick check of -rationale confirms that 
this is discussed there; e.g. "... zone administrators..." in 3.1.2.1. 
Contextual Restrictions in 
http://tools.ietf.org/html/draft-ietf-idnabis-rationale-09 )


And I got discouraging signals about the state of testing:


[01:02:41] <dan.hoopyfrood> is anybody maintaining an IDNAbis test suite?

[01:10:03] <dan.hoopyfrood> (searching for IDNA test materials yields 
something from 2005. 
http://www.gnu.org/software/libidn/draft-josefsson-idn-test-vectors.html )
[01:10:14] <dan.hoopyfrood> (oops; rather: from 2003)
[01:10:58] <klensin> There is lots of stuff for the 2003 version... and 
some issues about what "test suite" means.  Can supply later -- too 
distracting to try to do while following the discussion

[01:15:11] <dan.hoopyfrood> (I see various perl/python library test 
materials, including, from Feb 2008, 
http://testers.cpan.org/show/IDNA-Punycode.html )

[01:38:18] <dan.hoopyfrood> indeed, that is counter-intuitive that 
Bucher != bucher (where the u's have umlauts)
[01:39:09] <dan.hoopyfrood> this is *exactly* the sort of question that 
test suites are ideal for. I wish I could get the IETF to do more test 
suites. (I tried to convince CALSIFY to do a test suite, without 
success, a few years ago)


p.s. dan.hoopyfrood is a name that I didn't really intend for
permanent records, but I haven't figured out how to get my
jabber client (iChat) to call me something else.
see also http://esw.w3.org/topic/JabberChickenEgg

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Tuesday, 24 March 2009 03:48:55 UTC