On Tue, Jul 13, 1999 at 04:44:17PM +1000, Geoff Huston wrote:
> "However this proposal, of increased levels of supply of SLDs is, in
> its present form, as functionally flawed as the current environment of
> processing bottlenecks within the current set of SLDs

Somebody please kick me for re-entering this debate.

The situation has changed since 1996.  We no longer have processing
bottlenecks in .au (except for org.au but that isn't important).
We have no competition in .au yet - but that's also beside the point.
The bottleneck now is availability of "good" names (i.e. all the good-fit
names have been taken, and what's left are poor-fit names).

I see this problem frequently with customers.  E.G. Their business name
is mostly generic - generic words in com.au not allowed; already taken
in .com|.net; .org not appropriate; acronyms all taken; concatenation
unwieldy. I have had to suggest customers register their domain name
in Tonga ... (didn't go down well with the customer either).

> in that neither 
> process can definitively ensure the continued value and utility of the
> .au Domain Name environment itself, at least in its current
> form.

I'd like to know how this value and utility thing is measured.  IMHO a
situation in which there are bottlenecks is an indicator of low value.
Integrity could also be measured from a purely functional point of view,
such as % of operational nameservers, incidence of lame delegations,
locational diversity of NS hosts.

Geoff's paper goes on to mention a proliferation of SLDs reducing utility
of the DNS due to "xyz.X.au" existing potentially for many values of X.
As Geoff well knows the DNS is not designed to be a directory of content
and IMHO its utility cannot be measured on how well it achieves this aim.
Users need external tools (i.e. search engines) to find the location
of content and/or organisations and this practice is well-established.
IMHO increased proliferation of SLDs will increase the gap between domain
name and content and will cause users to modify their searching behaviour
- i.e. be a good thing.

> This proposal therefore is not a solution to the expectation
> of name functionality that many users of the Internet implicitly
> expect with the name system, as the equation of the name system to a
> de facto Internet directory service is a strongly entrenched
> belief."

Indeed.  IMHO the problem at hand is not name functionality but name
availability.  Names should necessarily be less functional because on
the Internet they must be unique and there are N billion people on the
planet who might want a domain name.

> Is it too much to ask to stop recycling over paths we've trodden time
> and time and time again with this topic?

I'd sympathise, except that there's no common understanding of the issues
and the thrust of my article here is that the problem now is essentially
different from the situation we had in 1996.

*ouch* Somebody kicked me! That'll teach me for getting involved in DNS
muckraking again.

Nick.
-- 
Zeta Internet                     SP4   Fax: +61-2-9233-6545 Voice: 9231-9400
G.P.O. Box 3400, Sydney NSW 1043        http://www.zeta.org.au/