ietf-nntp WG Review of draft-dfncis-netnews-admin-sys-05.txt

[Charles Lindsey]

In <01KS4132RBJO002O3W at mauve.mrochek.com> ned+ietf-nntp at innosoft.com writes:

>(1) Great stuff, needs to be a proposed standard, OK to put on standards track
>    as-is.
>(2) Needs to be a proposed standard, WG needs to reach consensus so needs to
>    become a WG work item.
>(3) OK but not quite ready for prime time. Publish as experimental.
>(4) No good, don't publish at all.

Basically, I agree with Russ, except that I would pick (4), with at most (3).

Whilst there is a need to propagate this sort of information, and some of
their record structures for containing it may be OK, for sure it does not
justify a brand new protocol with a brand new port no.

Essentially, this needs a distributed database, and since Usenet is
already a distributed database, the obvious way to do it is to use Usenet
itself to propagate this stuff. Yes, it might need a few extra features
such as 'named articles' which have been discussed from time to time but,
with care, it could be introduced without those, or in a manner that would
upgrade smoothly to those. Another method which has been discussed, and
even implemented on a trial basis, is to use the DNS mechanisms for this
(though one might argue that the DNS mechanisms are too much overloaded
already).

Another problem with their scheme is that it seems to require a grand
"root" of the whole system, and that is just not the Usenet way of doing
things.

-- 
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131 Fax: +44 161 436 6133   Web: http://www.cs.man.ac.uk/~chl
Email: chl at clw.cs.man.ac.uk      Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9      Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5