Re: [idn] NSI Multilingual Testbed Information (fwd)

[Mark Davis <markdavis@ispchannel.com>]

I would definitely agree with Patrik. At an absolute minimum, the provisional
registrations should restrict themselves to characters that are distinguished
under caseless NFKC identifier rules. While that may not be what the IDN group
finally resolves, it has a far better chance of not causing collision problems
than if arbitrary character choice is allowed.

There is a file containing those characters on
http://www.unicode.org/unicode/reports/tr15/data/LC-NFKC-Identifiers.txt.

For IDN use, one would substitute
  002D; 002D # P: -             HYPHEN-MINUS
for
  005F; 005F # P: _             LOW LINE

Mark

Patrik Fältström wrote:

> At 07.11 -0400 00-08-25, Hollenbeck, Scott wrote:
> >We fully intend to keep the test bed in synch with this group's efforts.  If
> >something other than RACE encoding is what eventually becomes a proposed
> >standard, we'll change the test bed to keep up.
>
> The problem is definitly not the encoding. The problem has to do with
> the question of what is "equal".
>
> Today "foo.com" and "FOO.com" are equal.
>
> For international domainnames, the IDN wg is not even close to know
> what is equal or not, so it will most certainly be a risk for anyone
> deploying internationalized domainnames.
>
> I.e. if users A and B register two domainnames which today according
> to whatever rules are _NOT_ equal, and the IDN wg decide that they
> later should be treated as the same. Do you, Scott, call A or B and
> tell that user that he can no longer have his domainname? Which one
> do you call?
>
>     paf