[ietf-nntp] draft-hoffman-rfc1738bis-02.txt

[Charles Lindsey]

In <878ygnh1r0.fsf at windlord.stanford.edu> Russ Allbery <rra at stanford.edu> writes:

>Charles Lindsey <chl at clerew.man.ac.uk> writes:

>> Well we discussed this briefly when the previous draft of this came out,
>> and I think it was said then that the 'news' scheme and the 'nntp'
>> scheme were not/should not be identical.

>They were not originally identical, but I don't see any particular reason
>why they can't be.

>> Things wrong with that:

>Yeah, these are all bugs in the ABNF that I think you should report to the
>author.  These all strike me as simple cases of something being written
>other than what the person intended.

>> 4. If the refbygroup form is used, then the news-server MUST be present.

>Is this something that you're claiming (in which case, you're wrong), or
>is this something that the draft is claiming (in which case that portion
>wasn't included in your message)?

It is something I am claiming. Article numbers within a group are unique
to a server, but bear no resemblance to the article numbers for the same
article on a different server. Therefore, if you do not specify the server
in a URL using such numbers, then you are likely to get the wrong
articles. Certainly so if you expect to publish the URL widely (which is
the usual case with URLs). At the very least, it needs a warning that a
URL without a server will only work in a context where the 'default local
server' is the one for which the URL was written. But safer to insist on a
server name every time.

On looking further, I see that the 'nntp' scheme is indeed defined in RFC
1738 (and is so registered with IANA), and the server parameter (or rather
<host>:<port> as it is written there) in indeed obligatory, and the proper
warning is provided.

>> 5. I am not clear as to the present usage of "nntp" as a scheme name, but
>> for sure I have never encountered a URL of the form
>> <news://news.foo.example/comp.foo/1234-1238> (it is certainly not
>> envisaged by RFC 1738).

>What's wrong with it other than it not currently being implemented by much
>software?  Seems useful in some situations to me, and only moderately hard
>to implement.

Yes, I see now that RFC 1738 did define both forms, one for the 'news'
scheme and the other for the 'nntp' scheme. I can see some sense in
combining them, but it is a decision that should be taken by the news
community after proper consideration.

>You seemed to think that I should take some action about this, and I'm not
>understanding why you believe that's the case.

I think this WG is the appropriate place to discuss this (as opposed to
Usefor, which is bogged down in other things). Indeed, I seem to recall
that we had included it amongst out list of things to do. I would think
the proper procedure would be for us to agree on a text to feed into
Paul's draft, or to invite Paul to come and discuss it here. In which case
it is better to come from you as Chairman rather than from me. But if you
really want me to contact Paul, then I could do so.

-- 
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 clerew.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