My two cents on draft-leiba-rfc2119-update

["Adrian Farrel" <[e-mail]>]

The recent high-volume threads on this draft had the effect of causing me to
read it. Sorry for being sucked in.

I oppose the deprecation of the keywords in Section 3. More specifically: if I
choose to capitalise these words in a document that I write then I either have
to define the meaning afresh in each document or write a new 2119 replacement of
my own to reference.

In general, however, I don't see the motivation for the document. 2119 (despite
some clumsy language) has served us well enough and is well-enough understood to
have been used a few times.

Barry's most recent suggestion is...

   In many standards track documents several words are used to signify
   the requirements in the specification.  These words are often
   capitalized, as shown below, but they do not have to be.  This
   document defines how these words are interpreted in IETF documents
   when the words are capitalized and/or marked as <bcp14> in the

This appears to say (well, it does say) that "the words" are used to "signify"
the requirements and are not always capitalised when they do. It then says it
defines how they are interpreted when capitalised "and/or" marked with BCP14,
which implies that if I use lower case but reference BCP14 then the
interpretation provided by this document applies. The later attempt to disclaim
definition of "normal English meanings" runs counter to these two statements.

If you feel a strong need to pursue this document then I think Ted's suggestion
(Talmudic commentary) is the way to go: that is, "Report on the use of RFC 2119
and guidance for future use of normative language." This would not be an update
to 2119 and would not require (but could be used as a) normative reference from
other documents.

Cheers,
Adrian