> The bottom line is that if a WG (which defines a protocol) explicitly states
> the rules (or conditions) for allocation of a parameter, is the ICANN (and
> by ICANN I mean the processes sponsored by the ICANN which may include the
> PSO) bound to follow? If it is, then I have no problem. If it isn't, then
> the autonomy of the standards process becomes at risk.
Good point, it's something that hadn't occurred to me. And I agree with
your concern.
It seems to me that its more than simply autonomy that is at risk -- the
integrity of a standard depends on the rules set forth by the keepers of
the standard (i.e. the WG or its designated experts via the IANA
guidelines section of new standards). In fact, I feel that the rules for
allocation are part of the standard itself and not only should be, but
must be, off limits to anybody else.
That said, in a lot of cases, the WG ought to feel empowered, and
encouraged, to relinquish control in those cases where an assignment is
really simply adding one in a space that isn't going to be exhausted (like
the MIB enterprise number space), and that the freedom implied by the
number isn't likely to reduce the value of the protocol, or, as in the
case of security, lead to what amounts to frustrating the expectations of
the users of the protocol.
The sticky area is that in the middle -- say the protocol number in the IP
packet -- there's not all that many possibilities and exhausting the set
is quite foreseeable -- and having another IP based protocol won't mislead
anyone -- yet having the IP protocol WG (do we have one?) sit on the space
like a jealous goose guarding a nest could lead to inter-standards body
food fights in which each side has positions of merit.
It's in this middle ground where ICANN with some set of neutral guidelines
or rules would be (to my mind anyway) a good and useful thing.
--karl--
