On 9/25/2014 1:30 PM, Brian E Carpenter wrote:
> I've given my views of this draft on the appropriate list, but...
>
> On 26/09/2014 04:39, Joe Touch wrote:
> ...
>> I see no utility to a BCP that makes operational recommendations without
>> qualifying them in RFC2119 language.
>
> Why? The word "should" is a perfectly clear word in the English language.
> RFC 2119 qualifies its meaning in a particular way, but there's nothing
> in the IETF process that requires us to use that qualification.
If this issued as a BCP, there should be recommendations in RFC2119
language. A BCP with no such recommendations doesn't explain current
practice sufficiently, IMO.
>> Further, this doc does not explicitly indicate the distinction between
>> upper and lowercase of the RFC2119 terms.
>
> Huh? It cites RFC 2119 in the prescribed words, which make it clear
> that "SHOULD" is a qualified version of "should". It's quite common for
> RFCs to use both. (The only word that can be problematic in that way
> is "may" - I have take to using "might" to avoid the ambiguity in
> "may".)
Once you cite RFC2119, you're saying that the key words used have
special meaning. There's nothing in RFC2119 that requires them to be
capitalized; the use of capitals is mentioned only in the abstract:
These words are often capitalized.
Often. Not exclusively. IMO, if you want to mix 2119 and non-2119 it is
necessary to be explicit in how to tell the difference.
>> IMO, except for use in
>> discussion prose, those terms need to be avoided at all cost except
>> where used in their RFC2119 sense.
>
> Why? RFC 2119 is perfectly clear about how it qualifies normal usage.
I found only text that defines their use as exclusively having special
meaning:
This document defines these words as they should be
interpreted in IETF documents.
Can you point to where you found the doc to say otherwise?
Joe
_______________________________________________
OPSEC mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/opsec
