Stability of valid IDN labels
John C Klensin
klensin at jck.com
Mon Apr 21 22:47:13 CEST 2008
--On Monday, 21 April, 2008 16:20 -0400 Andrew Sullivan
<ajs at commandprompt.com> wrote:
> On Sat, Apr 19, 2008 at 06:51:55PM -0400, John C Klensin wrote:
>> In a sense, the real stability criterion one would like is
>> "if this was valid at the time it was registered, then it is
>> valid forever". Unfortunately, that criterion is impossible
>> to state, much less enforce, given the many millions of
>> zones in the DNS.
> So for practical purposes, isn't it enough to say, "Once a
> character is valid, it is valid forever," irrespective of
> whether it is in fact ever registered? This solves the
> problem that you can't practically know all the names actually
> registered in the DNS.
I believe that, if one is going to start making "forever"
statements, that is the one you end up with. At the same time,
Eric is correct that registrations have never been treated as
immutable, even though concerns about archival URI references
are moving us down that path.
> (I realise it entails the cost that,
> if we're wrong in some serious way about a character, we have
> to live with it forever. That just seems to me like a good
> incentive to be as conservative as possible when adding things
> to the exceptions list.)
Yes, but where that argument leads is also to being as
conservative as possible when adding things to what is now
called the Protocol-Valid list, regardless of origins.
Fortunately or unfortunately, we got rid of "MAYBE" as too
complicated and not sufficiently justified, and our only real
lever for being conservative about, e.g., things classified as
letters went away with it.
More information about the Idna-update