Standardizing on IDNA 2003 in the URL Standard
marc.blanchet at viagenie.ca
Thu Aug 22 17:32:06 CEST 2013
Le 2013-08-22 à 11:17, Andrew Sullivan <ajs at anvilwalrusden.com> a écrit :
> On Thu, Aug 22, 2013 at 01:17:58PM +0100, Anne van Kesteren wrote:
>> On Thu, Aug 22, 2013 at 12:38 PM, Mark Davis ☕ <mark at macchiato.com> wrote:
>>> 2 Stay on IDNA2003.
>> 2 as deployed is not stuck on an archaic Unicode version.
> Right. 2 as deployed instead has _new_ compatibility problems as new
> registries and names compatible with IDNA2008 but that don't work
> correctly under IDNA2003 come online. Since that's where all the
> growth in Unicode is, this position represents the trade off of
> preventing a few things breaking right now (including a number of
> names that are impossible to type, like those with smileys and so on)
> at the cost of breaking future things more and more, as the IDNA2003
> assumption of Unicode 3.2 shows more and more strain.
right, as one of the authors of nameprep/stringprep and co-chair of idn (v1), we did talked about future versions of Unicode (I was the one that brought this topic at that time), as many on this list remember. However, the basic IDNA2003 design was based on more openness/liberal to new codepoints. IDNA2008 is more tight into which new codepoints are allowed, and therefore, has better rules to manage the Unicode new versions.
So continuing on using IDNA2003 will put more potential wrong codepoints into the system, which results into more compatibilty problems in the future.
I agree with Vint that, for the good of the Internet DNS future, we ought to move as soon as possible to IDNA2008 in order to avoid increasingly more compatibility problems in the future.
> It seems to me that one possible explanation for the success of IPv4
> was the early willingness to say, "These early ones didn't work.
> We're breaking them, even if you're using them. Sorry." I very
> strongly agree that preserving compatibility is extremely important.
> But if you're going to have to break things -- and given what we've
> learned, _some_ stuff needs to break -- the time to do it is as soon
> as possible. The problem will only be worse over time.
> Andrew Sullivan
> ajs at anvilwalrusden.com
More information about the Idna-update