For the record, I don't think
1.2.3/24
is or should be legal.
Thanks to the reference to RFC 3513, Section 2.2, this sort of thing is
clearly not legal for ip6 networks (see the ABNF production for
ip6-network in the protocol draft.) If someone can supply me an
equivalent reference for the ip4 notation, I'd be happy to include it.
I also don't think "ipv4:" should be a synonym for "ip4:". Note, that
since the standard is built on RFC 2234 for syntax definitions, all
keywords are case independent: "IP4:" is a synonym for "ip4:".
In general, I subscribe to the philosophy that the spec should be clear
as to what is legal, and implementations should be ruthless in
rejecting things that are malformed, misspelt, or otherwise broken.
Just because the underlying representation is text doesn't mean that it
isn't a protocol.
I also don't believe that the spec has to constantly say "We really
mean it here, in this case you MUST MUST MUST follow the syntax." That
sort of thing should be implicit for the whole spec and if you start
putting them in, then people feel free to play loosey-goosey with the
rest.
That said, Wayne, if there are areas of protocol that are actually too
liberal in what they accept, I would be happy to tighten them up.
- Mark