Re: [ssac-gnso-irdwg] Draft matrix

[James M Galvin <jgalvin@xxxxxxxxxxxx>]

I agree with everything that Jay says here with two clarifications 
indicated inline below.


-- On April 1, 2010 11:40:42 AM +1300 Jay Daley <jay@xxxxxxxxxxx> wrote 
regarding Re: [ssac-gnso-irdwg] Draft matrix  --
> Hi Steve
>
>
> Work pressures have prevented me from joining any calls recently so
> apologies if these comments seem to backtrack at all.
>
>
> 1. Models
>
>
> I am happy with models 1 and 2 as they both seem sensible
> possibilities.
>
>
> Model 3 however seems unfeasable - I cannot imagine a situation where
> model 3 will be the *only* model allowed, which means it effectively
> describes 'added value'.  Our focus (as with all policy groups)
> should be on baseline behaviour as there is an infinite spectrum of
> added value that we cannot predict.  I consequently won't be
> commenting on model 3.
>
>
> It is interesting to note there are no models that push the problem
> to registries.  Is this a deliberate omission?  I would be happy if
> it is but I think the reasoning needs to be recorded.
>
>
> 2.  Impact to registrars
> This is entirely gTLD specific.  Given the importance of ccTLDs in
> this particular arena I think that needs to be corrected.  That
> doesn't take too much work - just a note that 'most' ccTLDs also
> require ASCII in addition to the local script.
>
>
> The claimed cost issues are irrelevant and can go - registrars will
> ultimately choose what to support and will do so on a cost/benefit
> analysis that is out of scope for us to go into.
>
>
> 3.  Impact to registries
> For thin registries surely the box should say "no impact" - the whole
> definition of a thin registry is that they don't have registrant
> data?  I would say that cell B11 and B12 need to be combined into B12
> and C11 moved to C12, leaving "no impact" in C11 and C12.
I think what you mean is "no impact" in C11, at least that's what I 
would do.  If that's not what you mean then I'm confused since you can 
not both move C11 in to C12 and have "no impact" in C12.

> The "if a registry needs to get involved in administering a domain"
> in C11/C12 hides masks the big issue.
>
>
> The gTLD model and most ccTLD models are quite clear - the
> registrants are customers of the registrar not the registry and so
> all interaction with them is through the registrar.  Every day on
> mailing lists I see abusive domains reported and almost always the
> only interaction from the registry is to prompt the registrar to do
> something.
>
>
> I think this should say "The impact is that the registry will not be
> able to interpret the registrant information unless that have a
> service (internal or otherwise) that can translate the
> script/language used.  This will prevent them from engaging in
> administering a domain or from extracting detailed statistical
> information.  If may also hinder them when looking for similar data
> in different registrations for such purposes as abuse detection.
> Under the current model of gTLDs and most ccTLDs there is no official
> role for the registry to do any of these but if such a role were to
> develop then this would prevent it."
The important point here is that a registry is not obligated to 
understand all languages.  When registries get involved in 
administering a domain it is always through a registrar.  If it's not 
then the likely context is "abuse" and understanding the language is 
hardly necessary.  The action most likely to be executed is a "take 
down", but only after the registrar has not responded.  This action 
does not require understanding the language.

Jim




> 4.  Impact to registrant
> No comment.
>
>
> 5.  Impact to WHOIS user
> Need to add lines for "non-local user receiving WHOIS display in
> 'official' language" and "non-local user receiving WHOIS display in
> local language".
>
>
> The comment "Pose signficant challenges as Whois now in many
> languages that the local user would not understand" is not
> necessarily true - it depends on whether model 2 includes "from a
> restricted set that the TLD supports" or not.  If it does then this
> comment is not valid.
>
>
> kind regards
> Jay
>
>
>
>
>
>
> On 30/03/2010, at 4:36 AM, Steve Sheng wrote:
>
>
> Dear IRD-WG,
>
>   Attached please find the draft matrix requested by the WG in the
> last call. This matrix identifies three different models for
> registration data and the impact of each model on potential
> stakeholders.
>
>   Before using this matrix for further deliberation, we want to make
> sure of its correctness. So I feel it is important for WG members who
> are from (or knowledgeable of) registrar, registries operations to
> comment on the registrar and registry section;  for WG members who
> represent registrants’ right, or users of Whois to comment the
> respective sections as well.
>
>   Last but not least, if there are additional models that needs to be
> considered, please don’t hesitate to put them on the table.
>
>
> Warmly,
> Steve
>
> <matrix-draft-329.xls>