[NNTP] LIST EXTENSIONS and an NNTPv2 capability

[Ken Murchison]

Charles Lindsey wrote:

> In <87acurh9qz.fsf at windlord.stanford.edu> Russ Allbery <rra at stanford.edu> writes:
> 
> 
>>The proposed solution is this:  Specify in our base draft that any server
>>that fully complies with the new specification advertise the NNTPv2
>>capability as part of LIST EXTENSIONS.  Clients can then ignore or treat
>>with dubiousness any server that doesn't advertise that capability, thus
>>solving the Collabra problem.
> 
> 
> I am not particularly enamoured of this idea. Presumably the two
> implementations that currently do it wrong will shortly be mended to do it
> right. So the alternative of putting up with the situation for a hopefully
> shortish transitional period can also be considered.

As Mark pointed out offline (and I agree), simply doing nothing won't 
work because at what point can the client start trusting the output of 
LIST EXTENSIONS?  If there is no definitive way for a client to 
determine that LIST EXTENSIONS follows the updated NNTP spec, the 
command is essentially useless, as clients will *always* have to resort 
to trial and error.

LIST EXTENSIONS, or any other capability discovery mechanism, should not 
be designed or implemented (willingly or otherwise) to be purely 
advisory.  It should be guaranteed that its output is both complete and 
deterministic.

-- 
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