an interesting ICANN development on similar domain names

Vint Cerf vint at google.com
Sat Aug 9 22:04:14 CEST 2008


John

I fully agree - I just thought it was worth showing that the  
similarity issue
was being addressed outside the WG per our general hope and  
expectations.

vint


On Aug 9, 2008, at 3:16 PM, John C Klensin wrote:

>
>
> --On Saturday, 09 August, 2008 08:02 -0400 Vint Cerf
> <vint at google.com> wrote:
>
>> tring Similarity Algorithm Update -- ICANN staff recently
>> completed a workshop with SWORD, the partner who is assisting
>> ICANN with the creation of an algorithm that will help
>> automate the process for assessing similarity among proposed
>> and existing TLD strings. SWORD's verbal search algorithms are
>> used by various patent and trademark offices throughout the
>> world. SWORD has completed a beta algorithm and reviewed
>> several test cases with ICANN staff. This is being done in
>> order to refine the parameters and discuss how the algorithm
>> could be successfully integrated as a tool to help implement
>> the GNSO's recommendation that new gTLD strings should not
>> result in user confusion with existing TLDs.
>
> Vint,
>
> While I look forward with much curiosity to the outcome of this
> work, I believe that it is unlikely to have any bearing on the
> work of this WG for several reasons, among them:
>
> (1) At present, there are only a few hundred TLDs, all of which
> are relatively short ASCII strings.  Comparing them for
> similarity to a large collection of applications is relatively
> simple.  By contrast, comparisons to all possible existing
> strings (ASCII and IDNs in both U-label and A-label forms) that
> might appear in zones further down the tree would be a much
> greater challenge.
>
> (2) When names can be presented to the user in a wide variety of
> fonts and artistic styles, confusability may be very much in the
> eye of the beholder -- an eye conditioned by what the beholder
> is expecting to see when side-by-side comparisons are not
> available.  If ICANN staff and SWORD can solve that problem...
> well, I will be very interested.
>
> Most important, we have already agreed (I think) that confusing
> similarity of strings is a matter for zone administrators and
> not an objective of this WG.
>
>      john
>



More information about the Idna-update mailing list