RE: [gtld-council] Outcome of discussion on string checks on Wed 30 Aug in Amsterdam

["Bruce Tonkin" <Bruce.Tonkin@xxxxxxxxxxxxxxxxxx>]

Hello Ross,


> 
>  "Confusingly similar" does have the 
> benefit of being settled US law, but I don't see the value in 
> building a process that requires dragging US law into the 
> fray as a matter of standard process (i.e. "is this new 
> string confusingly similar to others? dunno - we need to ask 
> a judge."). 


In the discussion in Amsterdam - the term was actually derived from
European Commission law - the US law was not directly referenced.

I think the point was that the term may be "broadly" understood.  Of
course different legal frameworks probably have different
interpretations.

Regarding using a "judge" - the discussion was more along the lines of
how a "trademark" office makes a determination.   A judge is used really
as part of dispute resolution rather than as a first step.   The idea
was that using a panel of experts would assist ICANN in making a
determination.   There would then be various dispute processes - which
may ultimately involve a "judge".

There is also the argument that introducing this 
> criteria lends credence to the emerging notion that these 
> TLDs are the intellectual property of the proposer. I don't 
> believe that this is the case, and I agree with the 
> assessment that we should avoid fostering that impression 
> through our policies.

Agree.   It is not the intent.  We are essentially dealing with a
security issue that results when strings look sufficiently similar to
cause significant confusion amongst the general public - mainly from a
visual perspective.   If there is another way of wording this then
please suggest appropriate wording.   There is also the issue of IDNs
emerging and the intent was to use text that allows some flexibility.   

If it would help we could add additional text to clarify what we mean
here.   

E.g "The creation of a new gTLD string and the assignment of that string
to a registry operator does not imply any intellectual property rights
around that string.   The DNS is a public resource, and the registry
operator has the right to use that string for a specific period of time
etc".

Note that the same concept applies to domains at the second level - ie
the registrant has a "right to use" the second level domain, but does
not "own" the domain.


> 
> You've also made the statement that:
> 
> ""typo" confusion - is more related to the use of particular keyboards
> where it is easy to make a mistake in typing.   E.g mistyping "n" for
> "m".  Ie "tonkim" instead of "tonkin""
> 
> I also agree that the term "typo confusion" is inappropriately vague.
> Rather, instead of focusing on confusing similarity (legal 
> basis), or typo-confusion (error based), we should instead 
> use "typographically similar" as the criteria.
> 
> This would allow the process to execute in an objective 
> typographic context, i.e "that by which something is 
> symbolized or figured..."
> Typography has very little to do with the meaning, semantics 
> and intentions of words and characters, and everything to do 
> with how they look.

Sounds fine to me.  Can I just confirm that your definition would cover
example and examp1e?

The use of the term "similar" in your definition still involves some
subjectivity.   

> 
> Don't we want to implement simple processes that can be 
> predictably repeated with very little overhead? Or are we 
> doomed to a future of specialist committee's that will make 
> subjective determinations about various parts of the 
> application based on what their opinions "as experts"?

That is indeed the intent.   Note however that I think you probably
still need some level of expertise once you get beyond the latin
character set.  For example two Chinese characters may look the same to
me, but be seen very differently for people that use the Chinese
language.    

Regards,
Bruce