Re: [idn] LDAPv3 and IDN

[Eric Brunner-Williams in Portland Maine <brunner@nic-naa.net>]

> Also, there is a chance that one domain name can be represented by two
> different DNs; one using UTF-8-encoded IDCs and another using ACE-encoded
> DCs. That would be a Bad Thing indeed.

In the problem domain in which we work, several distinct encodings in ASCII
have been proposed, RACE, LACE, MACE, DUDE, ...

In the problem domain in which we work, several distinct extensions of ASCII
have been proposed, UTF-8, UTF-16BE, UTF-16LE, binary, ...

In the problem domain in which we work, several distinct implementations are
apparently extant, using SJIS, 8859-1, BIG5, EUC, 8859-11, GB, ...

Personally I'm not concerned about chance, or certainty across the value sets
of each distinct approach (encoding in, extension of, alternative to). VGRS
has the problem of rolling a set of identifiers from RACE to whatever, as a
product concern, others have a similar concern with the same, or equivalent
encoding in ASCII, but it is, for want of a better term, "registry private".

Personally I'm certain I'm interested in certainty between one element of the
first class, and one element of the second class (transition), as a minimum.
I don't agree that its a Bad Thing, in theory or in practice. I'm open on the
issue of equivalency between elements of the third class, and with one of the
first or second.