Re: [wg-c] registry contracts

[Mikki Barry <ooblick@netpolicy.com>]

At 10:46 AM -0500 11/15/99, Kevin J. Connolly wrote:
>If we do not recognize a distinction between for-profit and 
>not-for-profit, then, over
>time, there will be a tendency for the "good" TLDs to be captured by 
>for-profit registries.
>I think it would be bad for the future of the Internet were all of 
>the "good" TLDs to fall
>onto one side (or the other) of the for/non//profit divide.


That is not something for ICANN to decide.

>It's not a matter of taxation or wealth-distribution.  It's a matter 
>of responsibly allocating
>a fixed economic resource, i.e., slots in the authoritative root zone.


1) this is not a "fixed economic resource" nor should it be
2) responsible allocation should be limited to technical abilities only.

>If we are not to sit back and allow all of the TLDs fall into the 
>control of profit-
>maximizers, then ICANN needs to make some choices.  I have always thought
>that it was for this reason that the DNSO was called into existence.


Who is this "we" and what gives this "we" any power to manipulate 
markets and constrain trade in this way?

>We are also setting precedents for the near- and mid-term future of the DNS.
>It will become *harder,* not easier, to make these decisions as time goes on.


And there are some decisions which should not be made by ICANN.

>It can also view the "testbed" as a dry run of its allocation 
>policies.  (I enclose
>"testbed" in inverted commas because I do not think of 6-10 
>registries as a small
>enough rollout to constitute a true testbed.  Three registries 
>(e.g., .nom, .per and
>.mus) would be a true testbed; 6-10 feels more like a limited-scale 
>deployment,
>likely to have real economic significance.

This "real economic significance" would be greatly reduced if the 
"testbed" was of a much greater number of gTLDs.  30 would be a good 
number.  After all, in order for this to be a true "test" we need to 
stress the system in order to ensure that things work as we expect 
them to.  That is the cornerstone of software/hardware testing.