Single-letter names

Vint Cerf vint at google.com
Sat Jul 5 19:10:34 CEST 2008


John,

it is my belief (which can be readily challenged in this list ) that  
the question of minimum length of a U-label form is NOT an issue for  
IDNA-bis.  The basis for this view is that I have not understood  
there to be a protocol-based reason for this WG to establish  
restrictions on minimum length. The punycode function clearly imposes  
potential maximum lengths of U-labels based on their lengths express  
in the xn-- format. There is also, presumably, an implied restriction  
on the total length of a domain name expressed as a sequence of  
labels, some or or all of which are encodings of U-labels.

vint

On Jul 5, 2008, at 1:03 PM, John C Klensin wrote:

>
>
> --On Saturday, 05 July, 2008 12:55 -0400 Vint Cerf
> <vint at google.com> wrote:
>
>> no, I was specific that my remarks were oriented solely to the
>> IDN   question. I do not believe that our work on IDNs deals
>> with the ASCII   LDH question except insofar as we are
>> constrained to assure that our   mapping of IDN's expressed in
>> some form of UNICODE does not conflict   with established
>> practices associated with ASCII domain names. It is   clear
>> that a valid IDN is a minimum of some 6 ASCII characters I
>> believe, and that means none of the IDN work will directly
>> implnge on   questions about single character ASCII domain
>> names at top level or   elsewhere.
>
> Vint,
>
> There is also an almost-separate discussion of whether there
> should be minimum length requirements on the U-label form of
> IDNs (as well as LDH labels).   The current ICANN
> recommendations apply such minimums to TLDs, as discussed on
> this list and somewhat more extensively on the IETF list.
>
> I interpreted your note as indicating that such minima were not
> an agenda item for the IDNAbis WG, whether LDH, U-labels, or
> A-labels.  For the latter, as you point out, the prefix length
> plus some punycode encoding implies that we are well above the
> limit.
>
> Is that correct?
>
>     john
>
>
>



More information about the Idna-update mailing list