At 12:48 AM 7/14/97 EDT, Mark Hughes  mark.hughes§ccamatil.com wrote:
>Geoff wrote:
>>I must admit that after much thought about this I wonder about the
>>wisdom of using an external citation as the reference of 'validity'
>>of a DNS name. The problem is that there are so many different
>>citation points of reference and no clear uniformity or hierarchy
>>between them all, that a DNS system which attempts to reference
>>various external citation points still cannot resolve name clashes
>>cleanly and in my humble view does little to solve the problem.
>
>I agree - unless the external citation has already solved the
>uniqueness problem, (and there aren't many I can think of that will)
>then the problem of multiple organisations wanting the same name is
>not resolved by resorting to external citation as a criteria for
>allocation.


Historically there was never the need to solve the uniqueness problem
simply becuase there was never a technology which required such
a namespace which had the properties of large span and no
qualifiers. Hence the historically 'normal' way to solve the nameclash
problem was to limit the applicability of each administratively scoped
nameadministrations. (eg trademark classification categories,
state-based registry of business names, etc)


>FCFS is a very viable policy for avoiding unwanted complications -
>but it would have to be applied to conflicts WITHIN a 2LD as well as
>problems BETWEEN 2LDs.
>
>* Conflicts within a 2LD - eg: 'you gave him my name' type disputes.
> The issue here is how does the Domain Name Administrator protect
>themselves?  There are really only two strategies:
>1. First Come First Served - pretty simple and effective.

Here the DNA is not an active party - by applying for a name a
FCFS registry would simply allocate the name, and all subsequent
applicants are rejected. The recourse for such applicants lies
with the first applicant.

The downside  - hoarding and speculation (registring
the nameof every shire council and then attempting tosell it
back)


>2. Using a very precisely defined criteria, which may include an
>external citation of some sort.  eg, if we continue with the .tm.au
>example, the criteria might be:
>
>'We will first check if this is a valid trademark.  We will then
>check if you are an owner of this trademark.  We will then check if
>this domain name has already been allocated to someone else with the
>same trademark in a different category.  If no one else it using it,
>its yours.'

But the second last sentance of yours breaks down the whole thing,
and cannot really solve the dispute problem. You really have to
create a namespace which _precisely_ matches the external citation
structure if you want this to work.

Which leads me back to my original point - sometime, somehow, we have to
construct a structure which makes the DNS namespace a name citation
authority _of its own_,  ON PARITY with state business name registeries
and the Australian Securities COmmission's Company Name registry.

Well I can see here that there are two ways of achieving this: setting up
a structure and being prepared to defend DNA's decisions in the courts
time and time again (risky becuase you may win the first 50 cases, but if
you lose the next one you're out! - i.e. you are never _proved_ correct,
each case simply proves that you did not make a mistake that time
with the structure.)

The second way is to recognise the public nature of the DNS name space
and, like the ASC's company name registrar, work _WITHIN_ a
publically regulated space.

Depends on what you want to do - muck around in the courts,
and pay DNS registration fees to ultimately hand to lawyers and
an already massively overloaded legal system? Or work
towards an administrative structure which works in tune with
our existing social structures?

g