On Thu, 3 Jun 2004, Bob Braden wrote:
If a paper is citing RFC 793 because it wants to cite the standard
spec for TCP, then citing STD3 is exactly the right thing to do. I
would submit that this is nearly always the case. It it relatively
rare that a TCP citation depends upon some particular wording or
content of the a particular (obsolete) RFC.
*> STD numbers have dynamic content - as you mention above. Which is why
*> the RFC number (which never changes) is what should be cited.
No, it is STDs that have static meaning; RFCs can change.
RFC's themselves are not changed. They are superceded and amended by other
RFCs. Only their status is changed, according to the process defined by
The set of RFCs that compose a STD could change, and has changed, and
certainly will change in the future.
I think you are confusing the status of an RFC with a STD. A STD is a
related set of of one or more documents at the "Standard" maturity level
Specifying an RFC or a STD could mean different things. An RFC might
change status, but its contents will not change. So the protocol specified
in an RFC is "frozen". If you want something to work according to an RFC,
say a military communications system, and you never want that protocol to
change or you need to be compatible with old equipment, then specify an
RFC. For example, if you expect to pull the AEGIS destroyer out of
mothballs in 50 years, and you need equipment to work with its old IPV4
communication system, then you better specify an RFC. If you are sending
a satelite into space that won't return to earth for 50 years, you better
specify the RFCs, because in 50 years, probably the STD's will have
If you want to specify something that you can expect contemporaneous
commercial operators to support, and that might be updated from time to
time, but will still be anticpated to be supported by commercial
operators, then specify a STD.
So for example, when you want to get bridge laptops for the old destroyer
rather than overhauling all its systems, specify the RFCs that were
specified for the battleship. When you want to interconnect that system to
the rest of the present commercial network, specify that you need to proxy
those RFCs to the current STDs.
This isn't a hypothetical problem. I once had to work with HPUX 6 for a
long, long time (they probably still have those machines) because the OS
version was specified in the Military contract. The machines couldn't be
upgraded, and the contract couldn't be changed for such a trivial detail.
So we had to make due. Specify things in government contracts with
care--they sometimes last for a long time. Some things should be frozen,
other things shouldn't be.
Ietf mailing list