Some proposals

["Maria Greene" <maria@apparatenet.com>]

Hello, everyone. I recently returned to work after a lengthy time off to
pursue "other interests" and was excited to find that work is moving forward
on the SMI. I have read the drafts and the archives of the mailing list
(including the minutes of the first WG meeting) and I believe I'm up to
speed. I'm going to post a series of proposals in separate mail messages
which recommend some changes and additions, primarily to the SNMP mappings
document. Before I post the specifics, I wanted to give a little background
on them.

About five years ago I developed a set of ASN.1 macros while at Ascom Nexion
(now virtually defunct) which included a CLASS macro to add OO capability to
SMIv2. When I joined Xedia (now Lucent) I implemented the next generation of
that language. These extensions were used to generate agent implementation
classes in C++ and to data-drive the command line and web browser
interfaces. Now that I have joined Apparate Networks, I was preparing to
develop yet a third generation of the language when I discovered the work on
SMIng. Unlike my evolutionary approach of augmenting SMIv2 MIBs with class
mappings, this work is revolutionary and can dump the cruft of the SMI, like
ASN.1! What luxury.

Note that the two products that this language was used for (an ATM switch
and router with sophisticated VPN and QoS capability) were both fully SNMP
manageable and implemented many standard and proprietary MIBs. So, the
experience is practical and not just theoretical. It can be done and it has
enormous benefits. So, it is based on this experience that I make the
following proposals.

1. Improve alignment between SNMP mapping statements and core SMIng
statements.
2. Simplify class instance naming.
3. Add the snmp-method statement.

As I said. I will post these proposals in separate mail messages to better
break up the discussion. My "other interests" (two little girls) make my
work schedule a bit sporadic right now so I apologize if I do not respond to
email messages as fast as I should.

Regards,
	Maria