Domain names and URLs

Hey hey,

So one part that's missing in http://url.spec.whatwg.org/ is domain
names. As far as I can tell no browser has moved beyond IDNA 2003
(Opera "regressed" when it adopted Chromium) other than updating their
Unicode implementation and nobody is interested in implementing IDNA
2008 (whatever that means, it doesn't exactly define it all the way
from code points in an unparsed URL to host bytes).

Given that, my plan is to put that in the specification. IDNA 2003,
its label separators, the Unicode normalization it uses (but without
restrictions to a particular Unicode version), ...

Some registrars have moved to IDNA 2008, but this creates security
issues as pointed out in http://unicode.org/reports/tr46/ and people
are unlikely to register names that do not work. I suspect long term
the IETF will have to revisit this, but in the mean time it would be
good to have accurate documentation for how we want everything around
URLs to work.

If there's anything I'm missing here, I'd love to hear about it. If
you want more background reading, I wrote this after doing some
research last year: http://annevankesteren.nl/2012/11/idna-hell

Kind regards,


-- 
http://annevankesteren.nl/

Received on Thursday, 15 August 2013 10:43:32 UTC