ietf-nntp Normative references

[Russ Allbery]

In reading the latest draft, I started thinking about the issue of
normative references.  Ned, could you take a look at this as well and let
me know what the IETF position generally is?

As I understand it, if we want to release this on the standards track, our
document can't have normative references to informational documents like
RFC 1036.  That would mean it's not kosher to defer syntax definitions to
RFC 1036.  This has effects in various different places, including in the
definition of message-ids and in the specification for POST.

I'm not sure the best way to handle this.  One approach would be to define
the only article concepts we actually use (namely that an article consists
of some set of headers and a body, that the headers consist of key-value
pairs with case-insensitive keys represented by the header name followed
by a colon and a space, and that each article is associated with a unique
message-id which is also present in the article headers) and then note
that RFC 1036 articles are one such type of article that can be processed
by NNTP and RFC 2822 messages are another type.

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