[rfc-i] Summary: use case for 2119 markup

Andrew Sullivan ajs at anvilwalrusden.com
Fri Jun 20 14:17:54 PDT 2014

On Fri, Jun 20, 2014 at 04:37:48PM -0400, Ted Lemon wrote:
> We are accepting this process change because it's something relevant to the publication stream, and it doesn't change the meaning of the document.   That is, we are saying that the output of the editorial process means the same thing as the input, and therefore there is no problem.
> Markup of 2119 keywords is different, because we're actually adding a new nuance of meaning.

What is this new nuance of which you speak?  In my experience dealing
with implementers who are not close to the IETF, there is a widely
held belief that "MUST" is not exactly the same thing as "must".  If
your claim is that they absolutely are, always when RFC 2119 is
mentioned in the document, then I think that (1) you have an
interoperability problem with some other interpreters and (2) this is
in any case easily solved by saying that every instance of emm yoo ess
tee must be spelled <must/> or &must or whatever.  That makes this a
spelling issue, and one squarely within the ambit of the RSE.


Andrew Sullivan
ajs at anvilwalrusden.com

More information about the rfc-interest mailing list