[rfc-i] new BCP number when RFC is updated?
braden at ISI.EDU
Wed Nov 3 12:41:31 PST 2004
Peter Koch wrote:
*> I'm not sure whether this is an RFC Editor or an IETF process question, but
*> let's see. On Friday, RFC 3934 was announced, which is an update to RFC 2418.
*> RFC 2418 is BCP0025 and RFC 3934 is BCP0094. Now I'm wondering why
*> RFC 3934 was assigned its own BCP number instead of being added to the
*> BCP0025 set. RFC 2026 allows for 'a' BCP to consist of more than one RFC:
*> A specification, or group of specifications, that has, or have been
*> approved as a BCP is assigned a number in the BCP series while
*> retaining its RFC number(s).
*> Although they do exist for the STD series I could not find any precedent
*> (multi-RFC 'set') in the BCP list. However, IMHO it would be easier to
*> keep track if related documents would be ref'd by a single BCP number.
*> Could anyone point me to some (archived) reasoning?
A good question. First of all, the RFC Editor does not determine whether
a document should get a new BCP number or be assigned to an old BCP number.
The IESG does that (we sometimes have an opinion, though... ;-))
Secondly, we have never taken the step of allowing a BCP to label a set
of documents; only the STD subseries allows multiple assignments. This
is historical... BCPs have sort of grown in importance and stature over
the years. They are really quasi-standards, and we have gradually
acknowledge this in our rules. Yet, this case shows an example where
we have not fully internalized the standard-like quality of a BCP.
It would seem logical to allow BCPs to name document collections.
To do so would probably violate some coding details in our software,
but that could be fixed, certainly. We did think about this issue
when we developed a DTD for the XML version of the index. I don't
remember right now whether we ended up allowing multiple docs per
BCP, but hopefully we did.
Bob Braden for the RFC Editor
More information about the rfc-interest