ietf-nntp Simplification of wildmat

[Russ Allbery]

Clive D W Feather <clive at demon.net> writes:
> Russ Allbery said:

>> Yup, I think the simplification there is definitely worth it, for as
>> frequently as character sets are used in practice.  The second is much
>> shorter and easier to read.

> Personally I prefer them to remain. However, it's not a big deal.

I'm not going to argue it strongly either way.  I'd love to hear from some
of the rest of the folks on the list.

>> I note that both of these support comma and negation, so I guess
>> there's also a third proposal, namely to remove that as well from the
>> wildmat description, ban ',' and '!', and describe the matching in the
>> text of the one command that historically supports it.

> There might be such a proposal, but I'm not writing it. I agree with
> those who say that this should be part of the wildmat concept itself,
> rather than having a hodgepodge where sometimes I can reject subsets and
> other times I can't.

Fair enough.  I'll go with the conensus here.

>> I also agree with Andrew that if we keep comma and negation, we need to
>> clearly note that this is a behavior change from the previous version
>> of NNTP.

> Are we doing that in the other places we're changing ?

If we aren't, we should, but I think we have been for the most part at
least where it differs from existing practice (differing from RFC 977 is
another matter).

>> I think this is close to the last major outstanding issue.

> Streaming ?

I think we have that one mostly resolved.

> Reorganising the draft in a more sensible order ?
> Changing the command descriptions to a consistent notation ?

> [Are any of those "major" ?]

I wouldn't call the latter two "major"; they change a lot of text, but not
much in the way of meaning, which is the hard thing to reach consensus
about.

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