Re: sming meeting minutes draft... Questions for the wg.

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

>>>>> Durham, David writes:

David> Are people still interested in perusing the hierarchical
David> instance naming using oids as proposed in the smi-ds document?

David> Are you interested in helping by being an editor on one or more
David> of the documents listed below?

I came to the conclusion during the last months that the NMRG SMIng
proposal was right on target. During the SMIng WG process, I tried to
compromise up to an extend that I have serious trouble to write things
down without feeling what I write is somehow wrong. This is why I
never got serious work done on SMIv3 specs.

Andy's proposal to introduce hierarchical OID naming is an interesting
idea and I welcome the time he has spend on it. However, I feel that
it really only shines if there is some protocol support. But if we
assume new protocol operations anyway, then there are probably simpler
ways to communicate instance identification fragments without having
to encode everything into a single OID value.

Regarding the original NMRG SMIng proposal, we were actually not in a
bad shape:

(1) We had real documents. There were some unresolved issues and some
    features that needed to be added (multivalued attributes, unions
    spring to mind), but I am sure that with some effort and
    enthusiasm this could have been solved relatively quickly.

(2) We have a prototype implementation from one of our students as an
    extension for libsmi (although very rough running code) and an
    SMIng specification of the DIFFSERV-MIB with protocol mappings.

(3) There is another independent implementation done at INRIA which
    has been used in a project. For more details, there is a paper
    coming up at the IM 2003 conference <http://www.im2003.org/>.

BTW, adding some XML support would have been no real problem with
SMIng since that basically boils down to adding another mapping in
SMIng without having to touch the data structure definitions.

The bottom line is that I decided for myself that the NMRG SMIng
proposal was indeed right on target and that my interest to actively
contribute to alternate SMI versions is too small to count on.

Please read this text as what it is meant to be: A description where I
stand wrt. SMIng. You do not have to agree with this message.

/js

-- 
Juergen Schoenwaelder    <http://www.informatik.uni-osnabrueck.de/schoenw/>