ietf-nntp LIST EXTENSIONS

[Ken Murchison]

Clive D.W. Feather wrote:

> Ken Murchison said:
> 
>>>I think it is. LIST EXTENSIONS provides a way to ensure what is there and
>>>to know what options are available.
>>
>>Exactly, so why shouldn't we encourage clients to use it?  The trial and 
>>error method is archaic.
> 
> 
> Agreed.
> 
> I also don't have a problem with *encouraging* people to use it. I'm sure
> we can find working to satisfy that.
> 
> My problems are elsewhere.
> 
> More precisely:
> (1) Having reviewed the matter, I don't see why LIST EXTENSIONS can't be
>     pipelined.

I don't have a problem with pipeling it, but certain commands (e.g., 
plaintext auth) SHOULD NOT be used before receiving the LIST EXTENSIONS 
response.

> (2) Having reviewed the matter, I don't see why we're forbidding caching
>     of the results as opposed to warning people of the risks.

OK.  I could live with SHOULD NOT.

> (3) Saying SHOULD is rather stronger than just encouraging, because it has
>     conformance implications.

Per 2119 SHOULD and RECOMMEND are synonymous and I don't see where this 
causes a compliance problem.

>>>>We could also add the following sentence:
>>>>"Clients MUST issue this command before attempting to make use of any
>>>>security extension."
> 
> 
> looks very like "you MUST use LIST EXTENSIONS" to me.

MUST might be too strong here.  This goes back to the plaintext password 
issue.  SHOULD might be satisfactory, we'll find out in last call and 
IESG review.

-- 
Kenneth Murchison     Oceana Matrix Ltd.
Software Engineer     21 Princeton Place
716-662-8973 x26      Orchard Park, NY 14127
--PGP Public Key--    http://www.oceana.com/~ken/ksm.pgp