[Servercert-wg] [Ext] Referencing RFCs

[Paul Hoffman]

On 10/24/19 7:59 AM, Tim Hollebeek via Servercert-wg wrote:
> There is, perhaps, a legitimate point to be made about whether approved IETF Errata are normative or not.  We haven’t run into that situation yet, but it may arise in the future.

I believe that the IETF has no official position on this, and nor does the RFC Editor. There have been examples of errata that have been accepted and later removed. Your best bet is to point to the RFC and refer to the erratum by URL, saying that the erratum is normative for the document you are publishing.
  
> Another point which has been discussed a few times is how to handle RFCs that update an existing RFC (this is true of RFC 5280, for example).  I’ve heard both opinions expressed over the years, but the BRs themselves are silent on the issue.

Are you speaking of "obsolete" or "update"?

If RFC $y obsoletes RFC $x, everything that relies on RFC $x should be updated to refer to RFC $y. There are sometimes reasons not to, but they often fall into the category of "that's too hard" which is kinda lame. IETF decisions to obsolete an old RFC are usually careful and have strong reasoning.

If RFC $n updates RFC $m, RFC $m is not affected. There are many reasons to update, usually tangential to the core of an RFC. If you want the value of RFC $n in addition to RFC $m, you need to reference them both for the reader to understand the complete picture.

--Paul Hoffman