ietf-nntp TLS version requirements

[Russ Allbery]

Jeffrey M Vinocur <jeff at litech.org> writes:

> At the moment I have the text below, which "strongly recommends" that
> virtualhosting support be provided.  Is this acceptable as is?

This text seems fine to me, at least on first read.

> Should a RFC 2119 keyword be used?

Hm.

We could say that servers that are known by multiple names, and all
clients, SHOULD implement TLS-EXT, since those are the ones directly
affected.  In fact, the more I think about it, the more that sounds like
the right approach to me, unless I'm missing some other issue.  (It's
increasingly common for news servers to be known by multiple names, what
with outsourcing, although I don't know how widely TLS would be used in
that situation.)

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>