Re: MIB module numbering for TBD MIB roots

[Juergen Schoenwaelder <schoenw@ibr.cs.tu-bs.de>]

>>>>> Andrew Smith writes:

Andrew> That's not our problem (see earlier remark about
Andrew> foot-shooting).  There is a general "versioning" problem that,
Andrew> perhaps, some new version of SMI might help with. "Work in
Andrew> progress" continues throughout the standards' process, through
Andrew> PS, DS and S: allocating an OID at the PS stage is, at best,
Andrew> arbitrary.

The SMIv2 has very clear rules what you are allowed to change/update
once a MIB module has been published (where publication in the IETF
context means RFC and not ID). These rules guarantee that semantics of
fielded MIB implementations based on a published MIB module won't
change to guarantee interoperability. (Sure, there is a fuzzy border
line between what consitutes a semantic change in some special cases
but I think it is fine to let the responsible WG use their knowledge
about fielded implementations to make wise decisions.)

Do you think this approach is generally broken? What do you expect a
new SMI does differently? Perhaps you should write your ideas up and
send it to the SMIng WG list - otherwise it might not happen.

/js