Tony Finch wrote (2005-09-05)
On Sat, 3 Sep 2005, John C Klensin wrote:
The specific difficulty arises in numbered paragraph 6 of RFC
2119, which, to save chasing references, reads: [...]
RFC 2821 imposes a number of requirements that, by some
interpretations, are "not actually required for interoperation".
I think the other part of 2119's guidance - "to limit behavior
which has potential for causing harm" - allows the stricter
requirements of 2821.
In any case, paragraph 6 of 2119 is a meta-requirement about IETF
documentation standards, rather than about the contents of those
documents - it makes recommendations for writers not readers.
Stripping off the rationale doesn't change the definition of the
keywords.
In that thread Tony, Wayne, John Leslie, and me argued to use the
2119 keywords instead of 2181's idiosyncratic definitions, while
John Klensin argued that 2119 keywords can only be used "where it
is actually required for interoperation".
At the moment 2821bis-01 uses the common 2119 keywords without any
definition or 2119 reference.
Frank