RE: Further discussion about IANA Considerations for MIBs

["C. M. Heard" <heard@pobox.com>]

On Sun, 6 Jun 2004, C. M. Heard wrote:
> On Sat, 5 Jun 2004, Bob Braden wrote:
[ ... ]
> > Also, using a capitalized MUST here seems to me to be abusing the
> > MUST/MAY/SHOULD notation, which is defined for applicability of
> > technical specifications, not administrative policies of the IETF.
> > [I am sensitive about this, because I myself invented the MUST/SHOULD/MAY
> > terminology for RFC 1122, 1123, and I hate to see it used in unnatural
> > acts.]
> 
> I have asked for guidance from Bert on this matter, and he is
> consulting with the other IESG members .  The terminology did seem
> to fit our purposes, and it's used throughout the document in
> the following sense:
> 
>    MUST:  something that is mandatory for an author to do in order to
>    get the document approved.  A reviewer is not supposed to let
>    such things slide.
> 
>    SHOULD:  a guideline that an author had better have a good reason
>    for not following.  A reviewer is expected to ask for an explanation
>    if the guideline is not followed.
> 
> and so on.  But maybe this is not the right thing to do.
> 
> In any case, the next draft will comply with the guidance I get from
> the IESG.

The feedback that I got from the IESG (in a comment from Harald, to
be precise) was that RFC 2119 itself uses the MUST terminology in
this way.  Note in particular Section 6, "Guidance in the use of
these Imperatives" --

   Imperatives of the type defined in this memo must be used with care
   and sparingly.  In particular, they MUST only be used where it is
   actually required for interoperation or to limit behavior which has
   potential for causing harm (e.g., limiting retransmisssions)  For
   example, they must not be used to try to impose a particular method
   on implementors where the method is not required for
   interoperability.

Based on that feedback, I'm leaving the MIB review guidelines
document as it is.  That decision is, of course, subject to
challenge at last call time.

Thanks and regards,

Mike Heard