Re: Normative/Informative References

[Harald Tveit Alvestrand <harald@alvestrand.no>]

--On torsdag, januar 02, 2003 07:56:38 -0500 Thomas Narten 
<narten@us.ibm.com> wrote:

> > but other folk (including Braden) do not see the logic in
> > a normative reference in an informative document
> Interestingly enough, the current wording the RFC editor came up with
> doesn't seem to distinguish between standards and non-standards
> track. It seems to say all RFCs should do the split. That is not my
> recollection of where we ended up (though I think I'd prefer it on the
> whole, since so many people seem to just not want to do the split if
> its not *required*), so I'd welcome being pointed to more definitive
> words. I just looked at draft-rfc-editor-rfc2223bis-03.txt and
> http://www.rfc-editor.org/policy.html

I've objected before to the idea that informative documents cannot specify 
protocols - the decision to make something informational is (mainly) 
orthogonal to whether it's a protocol. And protocol descriptions need both 
the MUST/MAY machinery (if they want it) and the split references.

             Harald