[rfc-i] draft-kuehlewind-update-tag/

Mark Nottingham mnot at mnot.net
Wed Mar 25 16:00:49 PDT 2020

> On 26 Mar 2020, at 9:41 am, Martin Duke <martin.h.duke at gmail.com> wrote:
> What I was going to say in the queue:
> Like mnot, I think Updated should mean "Amended". It may be worth it to change the term just to create awareness to tighten the meaning.

To expand on discussion a bit -- one of the things that kept on coming up was that "updated" was never well-defined, causing this confusion. So one approach we could take would be to _define it_.  But having data about how bad the confusion is would help make the decision of whether to do that vs. mint a new term.

> But I dislike the idea of having "Extends" and "See Also". I foresee foundational documents (like RFC 793) with a few pages of RFC references before the text starts. That is useless. Plus the formal existence of these categories will encourage people to use them.


> If we would like better forward-tracing of standards evolution through time, I would prefer if the datatracker and rfc-editor pages simply listed the times the RFC was cited by other RFCs both normatively and informatively. I think that would be sufficient and automatable.
> TLDR, rename Updated to Amended, build the citation tool, and call it done.

I don't think we should call it done.

To me, this is further evidence that a linearly-numbered series of archival documents as the primary source of truth is damaging. 

We'd be better served if the documents that most people referred to were what's called "consolidated acts" in legislation. The source documents would still be available and immutable, but not primary references (unless there's a dispute about the consolidation). 

That implies that we'd need a new kind of WG output -- one that "patches" existing documents. Not hard, but requires some conventions and discipline by the authors. That patch could then be applied and the resulting consolidation published by the RFC Editor, or someone else. And both the original document and the patch would still be available separately, for when that matters.

The problem in all of this is that brand of "RFC numbers" has been built up over the year, and people are conditioned to use them as the primary reference. Giving the consolidated documents their own numbers is pointless, because they should be stable, not changing every time there's a new patch.

STD and friends seem to have failed. A tentative suggestion -- still publish RFC *numbers* for archival documents, but publish RFC *names* -- e.g., RFC-HTTP -- for consolidated documents that *can* be updated over time.


Mark Nottingham   https://www.mnot.net/

More information about the rfc-interest mailing list