RE: [soac-mapo] On "universal resolvability" and useful questions that emerged yesterday

[Milton L Mueller <mueller@xxxxxxx>]

Bertrand
I don't think you have adequately answered Antony's paradox.

I am still trying to figure out what exactly is the added value of so-called 
"universal resolvability" when one purchases this "value" at the price of 
completely excluding hundreds, perhaps thousands of top level domain services 
that hundreds of thousands or millions of people would be willing to use. Can 
you or anyone from GAC explain to me exactly what public benefit is achieved by 
doing that? I just don't get it. I see a net loss in connectivity and service, 
not a gain.

Further, (as we discussed personally yesterday) I think the GAC approach to 
universal resolvability is based on a technical misunderstanding. If a TLD is 
in the root, and a single DNS root is universally accepted, and the root name 
servers and TLD servers are configured properly, then a TLD is "universally 
resolvable." That is, it is available to anyone who wants to resolve it. That 
is the best the root operator can do. If a local network chooses not to avail 
themselves of the resolution capability, that is their choice and their 
problem, not the root opertors - and it is a choice that happens all the time.

By the way, the distinction between the second level and top level is 
completely irrelevant from a technical point of view. If you block a SLD you 
are restricting the resolution of a domain name. Many SLDs have more domains 
and users than TLDs, so don't try to say the distinction is one of scale.

The only real concern about universal resolvability comes from the threat of a 
fragmented root - and this refers to a situation where people who WANT to 
resolve a TLD name CANNOT do so, regardless of how they configure their system, 
because there are two or more incompatible, disconnected DNS roots. That 
situation, most of us can agree, poses public interest problems because many 
people will be confused about how to resolve names. But a deliberate choice by 
one household, network operator or government to block a specific TLD does not 
pose any of those issues.

Based on my discussion of this with various government officials (including 
mainly in the US) I think the threat of national-level blocking is being 
confused with the threat of a fragmented root. Do you think GAC's position 
would change if this mistake were pointed out and explained?

--MM



From: owner-soac-mapo@xxxxxxxxx [mailto:owner-soac-mapo@xxxxxxxxx] On Behalf Of 
Bertrand de La Chapelle
Sent: Tuesday, August 31, 2010 10:44 AM
To: Antony Van Couvering
Cc: soac-mapo@xxxxxxxxx; Gomes, Chuck; Cheryl Langdon-Orr; Frank March; 
Heather.Dryden@xxxxxxxx
Subject: Re: [soac-mapo] On "universal resolvability" and useful questions that 
emerged yesterday

Anthony,

It's only a paradox if you interpret universal availability as a principle that 
should apply to any possible string proposition (as if there is a sort of 
"right to be in the root"). But the objective is universal availability of the 
strings that will be accepted in the root (those rejected will be for many 
reasons, ie : all the diverse objections listed in the DAG). The ambition is 
that strings that are in the global common root are as universally available as 
possible. But you are right in a certain way : there is potentially a sort of 
intellectual feedback loop there.

Interesting qualification of .xxx as "existing to be blocked" :-) Not sure this 
is what Stuart had or has in mind ...

Best

Bertrand



On Tue, Aug 31, 2010 at 4:24 PM, Antony Van Couvering 
<avc@xxxxxxxxxxxxxxxxxxxx<mailto:avc@xxxxxxxxxxxxxxxxxxxx>> wrote:
Bertrand,

Thank you for this resume and overview.  It is very helpful.

At the heart of your argument is an interesting paradox.  Shall we set up rules 
to block certain strings in order to preserve universal availability?  If a 
string is blocked, it is available to no-one at all, and hence is the opposite 
of universally available.  If it is blocked only partially, for example by 
certain governments, then it is not universally available, but it is 
nonetheless far more available than if it had never seen the light of day at 
all.

What are your thoughts then on .XXX, which exists to be blocked?

Antony


On Aug 31, 2010, at 4:56 AM, Bertrand de La Chapelle wrote:

> Dear all,
>
> Following Milton's request, I'm trying here to reformulate more clearly what 
> I said at the end of the discussion yesterday regarding universal 
> resolvability. This is an attempt to reframe the issue we are facing to make 
> it more addressable. I am not speaking on behalf of the whole GAC here, as 
> this has not been discussed in that level of detail yet.
>
> 1) It is true that there is no absolute universal resolvability today at the 
> Top Level. However, among the 270 TLDs or so, blocking of a whole TLD is an 
> extremely rare case and only done (from what I've heard) by a very limited 
> number of countries. Hence, we can consider that there is a general situation 
> of universal resolvability, with some rare exceptions. (universal 
> resolvability here is not understood in the pure technical sense of the term 
> but more as "universal availability"). This has clearly been a positive 
> situation for the Internet as a whole.
>
> 2) However, the desirable opening up of the domain name space is likely to 
> introduce more cases where the string may not be considered universally 
> objectionable (by whatever criteria or process), but nonetheless would be 
> sufficiently "sensitive" in some countries for them to decide to block it. 
> This is what the GAC alludes to (in its gTLD principles) when it says that 
> TLDs should respect sensitivities regarding terms of national, cultural, 
> geographic or religious significance.
>
> 3) To handle such sensitive cases, there are two extreme approaches : either 
> making no limitations whatsoever at the root level and potentially reducing 
> significantly the universal accessibility because many countries would block 
> many TLDs; or at the other extreme, giving a de facto veto right to every 
> individual government on what gets into the root. Both approaches seem 
> inappropriate, or at least, unlikely to gather consensus in the group.
>
> 4) In other terms, the expansion of the TLD space means that there will be 
> some strings that will be in the root and blocked at the Top Level in some 
> countries. This is regrettable but probably unavoidable.
>
> 5) As we finalize the new gTLD program, I believe there is a legitimate 
> common and public interest objective of having/keeping as much universal 
> resolvability/availability as possible and as little blocking of whole TLDs 
> at the national level as possible. Therefore, we should probably not speak of 
> a "principle of universal resolvability" but of an "objective of universal 
> availability, with limited exceptions".
>
> 6) Such national exceptions should, building on the mechanisms of the UDHR or 
> the Treaty of Paris (often used as reference to MaPO provisions), be made by 
> law and be based upon national norms of morality and public order (here MaPO 
> norms are at the national level and this is OK). Moreover, with respect to 
> the traditional principle of proportionality, any blocking should ideally be 
> conducted at the lowest granular level possible, which means that blocking of 
> a whole TLD should remain an extreme and exceptional measure.
>
> 7) Therefore, I believe the challenge we are trying to address is to find 
> ways, at the global level, to handle such cases in the most predictable and 
> objective manner, so that objections can be formulated, evaluated, and 
> ultimately measured with respect to the global public interest (ie : the 
> benefits of a new TLD outweigh the inconvenients).
>
> I hope this clarifies what I tried to convey yesterday.
>
> Finally, I would like to highlight some very interesting questions that came 
> up in the good discussion yesterday evening and could structure part of our 
> future interactions :
>
> - how early in the overall process should MaPO/public interest/sensitivity 
> objections be handled ?
> - are we talking "string only" or is the applicant also a relevant element ?
> - would a panel (however it is formed) provide "expert advice" or amount to 
> full "outsourcing" (Frank) ? in other terms, how binding would the 
> recommendations of such a panel be ?
> - does the Board have to make an explicit decision for every TLD (even if 
> it's a mere endorsement of the result of the process, like in the IDN ccTLD 
> FT) or does the final decision rest with the staff determination or the 
> different panels ?
> - how much flexibility and direct responsibility would/should the Board have 
> in the final decision, in particular in the case of sensitive strings ?
> - would it be useful to explore a mechanism of supermajority for the Board to 
> refuse a TLD and/or to overrule a negative recommendation by the panel or 
> objections by some governments ?
>
> Looking forward to further discussions on the list and conference calls.
>
> Best
>
> Bertrand
>
> --
> ____________________
> Bertrand de La Chapelle
> Délégué Spécial pour la Société de l'Information / Special Envoy for the 
> Information Society
> Ministère des Affaires Etrangères et Européennes/ French Ministry of Foreign 
> and European Affairs
> Tel : +33 (0)6 11 88 33 32
>
> "Le plus beau métier des hommes, c'est d'unir les hommes" Antoine de Saint 
> Exupéry
> ("there is no greater mission for humans than uniting humans")



--
____________________
Bertrand de La Chapelle
Délégué Spécial pour la Société de l'Information / Special Envoy for the 
Information Society
Ministère des Affaires Etrangères et Européennes/ French Ministry of Foreign 
and European Affairs
Tel : +33 (0)6 11 88 33 32

"Le plus beau métier des hommes, c'est d'unir les hommes" Antoine de Saint 
Exupéry
("there is no greater mission for humans than uniting humans")