John Levine wrote:
In numerous places the development and deployment guide makes use of RFC
2119 language that is vague in its meaning. For example:
In particular, great care MUST be taken when
releasing memory pages to the operating system to ensure that private
key information is not disclosed to other processes.
This actually tells the implementor very little. My recommendation
would be to change to "must".
Channeling Dave here, 2119 language is case independent, i.e., must
and MUST mean the same thing, and the capitalization just calls it
out.
John forgot to list my +1, so here it is.
Lots of existing BCPs include 2119 normative language, so that
argument appears already to have been decided, but I agree they're
overused here.
IMO, cConfusion about the unimportance of case, for reserved normative terms,
has made it essential to use those terms only for normative purposes, and to
use
other words for non-normative meanings.
Interestingly, the exercise of replacing normative terms for non-normative ones
often results in clearer and more precise meaning, IMO.
d/
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html