To:
Geva Patz <geva@bbn.com>
cc:
"'provreg List'" <ietf-provreg@cafax.se>
From:
Jaap Akkerhuis <jaap@sidn.nl>
Date:
Wed, 20 Dec 2000 16:36:04 +0100
In-reply-to:
Your message of Wed, 20 Dec 2000 10:15:31 -0500. <20001220101530.B37763@bruno.bbn.com>
Sender:
owner-ietf-provreg@cafax.se
Subject:
Re: domreg BOF Meeting Minutes
Geza,
On Wed, Dec 20, 2000 at 10:06:48AM -0500, J. William Semich wrote:
> Many (the majority?) of the ccTLDs require a minimum two-year initial
> registration period.
Precisely my point. Minimum and maximum initial periods and renewal
periods are all policy issues, not protocol issues. The protocol should
allow policy to be expressed, but shouldn't dictate policy through
design limitations. We shouldn't even mandate that registrations need
expire: although the contrary is true only in a small minority of cases,
we should nonetheless cater for these cases, particularly if we envisage
the protocol potentially being used to register other classes of objects
(AS registrations, for instance, don't expire). The protocol should
allow an expiry date to be specified, but shouldn't require it, and
certainly shouldn't constrain it to one-year resolution.
You beat me in time, but this is also my point. In the GRRP draft
are more places where were policy issues are specified as protocol
issues. Talking to the German ccTLD peole, they had the same problem
with the draft; policy issues are mixed with the protocol issues.
It might be natural for someone from NSI/Verisign to mix them in,
but if you have a different policy, one tends to notice this.
jaap