Erik van der Poel
erikv at google.com
Wed Apr 8 04:19:57 CEST 2009
I believe one way to think of these things is as supersets and
subsets. The set of "strings" is infinite, and many of them are much
too long to be DNS/IDNA labels of any kind. In IDNA2003, the set of
"internationalized labels" is the set of labels that can be converted
successfully using ToASCII:
In IDNA2003, "ACE labels" are labels that start with "xn--" and are
the result of a successful ToASCII. If ToASCII succeeds but the label
does not start with "xn--", it is a plain ASCII label.
So an IDNA2008 A-label corresponds to an IDNA2003 ACE label, and an
IDNA2008 U-label corresponds to an IDNA2003 ACE label that has been
run through ToUnicode.
So, one possibility for M-label is to make it correspond to an
IDNA2003 "internationalized label", which would make the set of
M-labels a superset of U-labels. (The other possibility mentioned by
John is: [<set of U-labels> + <set of M-labels>] corresponds to <set
of IDNA2003 internationalized labels>.)
Does this make sense?
On Tue, Apr 7, 2009 at 5:31 PM, Paul Hoffman <phoffman at imc.org> wrote:
> At 4:59 PM -0700 4/7/09, Lisa Dusseault wrote:
>>I'm not certain a new term needs to be introduced. If we're talking
>>about a string that is invalid as a label, giving it that term seems
>>to legitimize it. If we're talking about a string that may or may not
>>be valid, that's just "a string"
> Good catch. John's second option was:
>> "An M-label is a string that can be mapped into a
>> [valid] U-label. It may be a U-label, since those
>> trivially map into themselves. The category of U-label
>> is a proper subset of the category of M-label." (Mark
>> and others).
> When I supported it, I was thinking of it as "...a valid U-label", that is, without the "valid" being optional.
> Idna-update mailing list
> Idna-update at alvestrand.no
More information about the Idna-update