RE: [gnso-idn-wg] One comment on techno-policy details

["Tina Dam" <tina.dam@xxxxxxxxx>]

Sophia, all,
I guess what I was aiming at is that uniqueness is an important factor for
the DNS. That means both at a user standpoint and a technical standpoint.
Strings must be unique from a technical standpoint, which they are if they
have different codepoints...this is the way computers can tell the
difference between what character it is you have entered and as such the
entire idea behind for example Unicode.  As you know ip-addresses and the
allocation of such by IANA and the regional registries ensures that
uniqueness. The fact that we have domain names is simply because it is
easier for humans to remember such names as opposed to ip-addresses.
 
>From a user standpoint, the same thing applies...if strings are not
considered unique how do you know where you are going or for example who you
are  sending an email to? The problem on the user side is that users cannot
"read" characters the same way as computers to. Some characters that are
different (perhaps based on different scripts) look exactly the same and
then what the user enters depends on how his/her system is set up/ keyboard
etc. That is a problem that we need to make sure is solved. Part of this
will be solved via the protocol by limiting the characters that are
available in IDNs, but other parts need to be solved via policy decisions.
 
Does that makes sense?
 
Tina
 
 


  _____  

From: Sophia B [mailto:sophiabekele@xxxxxxxxx] 
Sent: Wednesday, February 07, 2007 7:59 AM
To: Tina Dam
Cc: Mawaki Chango; olof nordling; chun@xxxxxxxxxxxxxx; gnso-idn-wg@xxxxxxxxx
Subject: Re: [gnso-idn-wg] One comment on techno-policy details


Tina,
 
Did you then partially answer Mawaki's ?.  is it the "root server" we are
trying to resolve for the 'confusingly simiar' situation or is it the
"user'"community?
Sophia
 
 
On 06/02/07, Tina Dam <tina.dam@xxxxxxxxx> wrote: 

Maybe this helps:

.com in Latin looks awfully like .com (with Greek o) and there are tons of
examples like that. Take a look in Unicode. 

The root server will see the difference because the punycode version is
different because the Latin o and Greek o has different code points. For the
user though it looks similar.

> -----Original Message----- 
> From: owner-gnso-idn-wg@xxxxxxxxx
> [mailto:owner-gnso-idn-wg@xxxxxxxxx] On Behalf Of Mawaki Chango
> Sent: Tuesday, February 06, 2007 10:59 PM 
> To: olof nordling; chun@xxxxxxxxxxxxxx; gnso-idn-wg@xxxxxxxxx
> Subject: RE: [gnso-idn-wg] One comment on techno-policy details 
>
> Hi Olof,
>
> Basically, it looks to me as either you are in violent
> agreement with Chun (then there might be some changes to the
> para./statement he's referring to,) or you're missing the 
> point, (or again, I'm lost.) Your example refers to the "same
> script" (wether it's ASCII based or
> else) whereby "confusingly similar" makes some sense. The
> point Chun was making, as I've understood it, was that there 
> is a problem applying the notion "confusingly similar" in
> cross-script situation ("different languages", "different
> language script labels," etc.)
>
> In other words, does it make any sense to assume that a new 
> non-ASCII, IDN gTLD might be "confusingly similar" to an
> existing ASCII gTLD. If you (anyone) think so, then how and
> to whom? To the DNS server? to some users? If not, then are
> we talking about an IDN gTLD (string) in one script being 
> possibly "confusingly similar" to another IDN gTLD (string)
> in another script? And then again, at what
> level: root server? user visual experience? and in the
> latter, to what extent do we expect the users of different 
> languages and scripts to be the same, etc.? Ultimately, the
> rationale and relevance of this notion may need to be
> reconsidered and clarified in the IDN context, and if
> relevant, grounded in the IDN space as well as it may be the 
> case in the traditional TLD space.
>
> Or have I wholly missed the point?
>
> Thanks,
>
> Mawaki
>
> --- olof nordling <  <mailto:olof.nordling@xxxxxxxxx>
olof.nordling@xxxxxxxxx> wrote:
>
> > Dear Chun,
> >
> > Thanks for your comments! One clarifying point, hopefully, would be
> > that the "confusingly similar" test (as conceived in the new gTLD 
> > recommendations)
> > would be applicable to concurrent applications for gTLD strings.
> > Accordingly
> > (by way of example in ASCII), if an application for a
> string ".tuvw" 
> > is received and another application (in the same script) is
> received
> > for ".tuVw", where v and V symbolize variants (again for
> the sake of
> > example only), they would be considered "confusingly 
> similar" in the
> > string tests and be handled in accordance with a specific procedure
> > foreseen.
> > Hence the
> > statement you refer to.
> >
> > Very best regards 
> >
> > Olof
> >
> >
> >
> >   _____
> >
> > From: owner-gnso-idn-wg@xxxxxxxxx
> > [mailto: owner-gnso-idn-wg@xxxxxxxxx] On Behalf Of Chun Eung Hwi
> > Sent: den 6 februari 2007 19:59
> > To: gnso-idn-wg@xxxxxxxxx  <mailto:gnso-idn-wg@xxxxxxxxx> 
> > Subject: [gnso-idn-wg] One comment on techno-policy details
> >
> >
> >
> > Dear all,
> >
> >
> >
> > I couldn't catch up the recent debates, but I want to make quick 
> > comment on one issue of "limit confusion caused by
> variants", which I
> > could read from conference call 23 January overview - 2.2
> as follows;
> >
> >
> >
> > 2.2 Agreement to limit confusion and collisions due to variants.
> > Agreement
> > that this may be a stability and security issue and part of the
> > reserved name process. Agreement that variants of an IDN 
> gTLD string
> > be treated in analogy with current practice for IDN SLD
> labels, i.e.
> > variants are not available for registration by others.
> Agreement that
> > this approach implies certain "ex ante rights" with similarities to 
> > the "confusingly similar" test foreseen in the New gTLD
> > recommendations. Agreement that such "rights" must not be
> confounded
> > with IPR rights as such. Some support for enabling a choice 
> for an IDN
> > gTLD strings with variants to only block variants or to use
> variants
> > as aliasing.
> >
> >
> >
> > What I want to clarify here is the fact that variants come from the 
> > same language or the same language family. Therefore, the
> confusion or
> > collision happen in the same language or within the same language
> > family as well. We cannot use the term of variant in case when some 
> > translated or transliterated or phonetically same or similar words
> > (language script
> > labels) are to be taken into account. And obviously, in different
> > languages or in different language families, there is no longer 
> > confusion or collision even when those  in respective language are
> > similar or the same in graphics, semantics and sound
> because different
> > language scripts must be distinctive itself. So, in this case, 
> > "confusingly similar" test cannot be applied.
> > Accordingly, across different language script labels, there
> should not
> > be any "ex ante rights" of the existing TLD label, and so 
> any reserved
> > name policy would not necessarily be designed.
> >
> >
> >
> >
> >
> > regards,
> >
> >
> >
> > Chun 
> >
> > --
> > ---------------------
> > Chun Eung Hwi
> > General Secretary, PeaceNet Korea
> > chun@xxxxxxxxxxxxxx
> > pcs (+82) 19-259-2667 
> > fax (+82)  2-2649-2624
> >
> >
>