[CCWG-ACCT] WP2 Issues from last night's call

[Eric Brunner-Williams]

I agree that the microprocessor language wasn't helpful.

But, I don't agree with the distinction below of "enable" from "use".

Neither registrars, nor registries, are necessary for services, whatever 
those may be, from being associated with unique identifiers. We never 
had to construe "competitive" to mean separate the registrar function 
from the registry function. Registrars are a consequence of a policy 
choice made decades after users started associating services with names 
that mapped, first via HOSTTABLES, and subsequently via the DNS. As for 
the non-necessity of registries, had our design choices been different 
(from munging the central point of failure of SRI-NIC as the HOSTTABLE 
publisher and the subsequent DNS zone file publisher), we might have 
made use of a mechanism to allocate unique strings -- aka "names" -- to 
addresses that does not require persistent state, let alone the rest of 
what we think of as a registry.

To respond to the example Malcom offered, many registrars register names 
that for some period of time are not associated with the end user's mail 
service. To attach semantics to when the user takes over substantial 
control of the name-to-address association is just asking for trouble.

I really recommend rethinking the absolutism and small c catholicism of 
"no service". Being this non-specific is going to cause people other 
than myself to ask "did they really mean that or are there some secret 
holes so that real work can be done?"

Eric



On 11/4/15 3:26 PM, Malcolm Hutty wrote:
> Greg,
>
> Thank you also for being willing to consider my suggestion.
>
> Some brief comments to your specific points:
>
> On 04/11/2015 22:47, Greg Shatan wrote:
>> I still have an issue with the phrase "service that uses the Internet's
>> unique identifiers."  I've looked at this phrase a hundred times, and I
>> would not have come up with the idea that it is limited
>> to *_hardware_ *(which are things, not services...).
> I agree, a service is not hardware. I assume you're replying to my mail,
> in particular where I referred to web servers and mail servers. These
> are not hardware, they are services. A web server is not a
> microprocessor, it is a data processing service that instantiated
> through software running on a microprocessor. But everyone (whether from
> the technology community or the law) calls it simply a "service".
>
>> Registries and registrars clearly are services
>> that use the Internet's unique identifiers;
> No, they really don't (or not in the relevant way). They enable the use
> of those identifiers. That's not the same thing.
>
> Imagine I go to GoDaddy and register hutty.com. Would you say that
> GoDaddy is using hutty.com? Surely not.
>
> When I point hutty.com at a mail server I operate, am I using hutty.com?
> Surely yes.
>
> Remember, this isn't going to be looked at by some sort of Cartesian
> "Evil Demon" that will be trying to twist the answer into precisely the
> opposite of what is reasonable, despite all evidence to the contrary.
> This will be looked at by the ICANN Board, quite possibly by the IRP,
> and - very unlikely, but with a faint possibility - by a court. The
> Board has every reason to interpret this how we would expect, and the
> IRP and the court are fair and impartial arbiter, not evil demons. So
> they will look at this phrase in the context of the rest of the Mission
> statement, and indeed the rest of the text. It really isn't as hard to
> interpret as you are making out.
>
>> To capture the meaning you attribute to the phrase in question, I would
>> suggest the following:
>>
>> "[microprocessor controlled] technologies that use the Internet's unique
>> identifiers to access the Internet"
> Honestly, I could accept that if really necessary, but it seems absurdly
> cumbersome.
>
>> In this way, we have more explicitly limited the "service" as you
>> suggest, and also limited the "use" as you suggest.  And we avoid the
>> ambiguity of the word "services," which clearly means very different
>> things to different people.
> You play upon being a lawyer, but from your reasoning you make me wonder
> if you are a lawyer with an intellectual property specialism not used to
> telecoms law. Referring to "services" as I have described is, I assure
> you, entirely common within telecoms law within the EU and the UK -
> which is where I have direct knowledge. I do not have professional
> experience of US federal telecoms law, but as someone with a layman's
> interest I'm pretty confident the term is also in common use there.
>
> Moreover, the term "service" is also used to describe web service, mail
> service etc, in the manner I have described, amongst the technical
> community (Andrew? Are you still here? Could you confirm please, if
> you're following this thread?).
>
> And, taken as a whole, I think any reasonable layman would also
> understand the intent.
>
> ICANN gets to regulate the DNS system, in the manner and to the extent
> set out in more detail in other clauses. It doesn't get to regulate the
> rest of us, with entirely unrelated businesses, merely because we've
> signed up for a domain name. Seriously, the concept really isn't that
> difficult.
>
> I also happen to think the Food and Drug Agency can get to regulate
> food, without being granted unlimited lawmaking power over anyone that
> wants to eat. But hey, maybe that's just me.
>