Ira,
It is unfortunate that you somehow misunderstand some of the major
objections I have with the way the Job Monitoring MIB has been
progressing. Hopefully this message will serve to improve that
understanding.
First, it is important for you to recall that you have never been
present at any of the JMP meetings. (And believe me, I wish you
had been there, at least for some of them.) If you had been to
at least *one* of the meetings, I'm confident you would have
interpreted my last mail message in a very different way.
One sentence you wrote is a bit disturbing:
> Why not ease up on Tom and the other members of the JMP?
I don't recall making any kind of statements about any of the JMP
members other than Tom. In fact, I think the other members are doing
an excellent job of persevering through the extraordinarily lengthy
process of getting the Job MIB to closure.
Another disturbing comment, for which I have no idea how you've reached
this opinion:
> I am puzzled by your objections to the IDEA of the IETF Job Mon MIB.
How did you ever reach this conclusion? Those JMP members who've been
attending the JMP meetings all along know very well that your statement
does not reflect my opinion about the Job MIB in any way. I am very
much in favor of a Job MIB...however, a SIMPLE Job MIB, and not a complex
one that attempts to solve all the world's problems in one fell swoop.
What *many* of us in the JMP group believe is that the approach taken
by Tom has resulted in an overly complicated approach to providing
print job-related info via a job MIB. Harry Lewis (IBM) has echoed
this position *countless* times over the last several months...yet Tom
always seems to totally ignore this serious concern, and continues to
propose an object model that does not map very well to existing printing
systems. (True, bits and pieces of the Job MIB map to most printing
system, but only in a "sparse" sense.)
> Within weeks of the first version of
> the Xerox Job Monitoring MIB (two years ago), our common client team
> built a live prototype.
Why didn't Xerox (via Tom or some other Xerox representative) simply
propose this Xerox-developed model as a starting point? If you review
the history of the JMP Job MIB, you can't help but believe that its
development was *not* spawned from real practice, but rather some sort
of "fundamental research" approach.
Please note that during a break at the JMP meeting in San Jose last
December, I specifically asked Tom about how Xerox's efforts were
going with prototyping the model he has been championing. His response
was that Xerox was not currently attempting to prototype the Job MIB.
Perhaps you can now understand my concern about this.
> After converging on how to manage a printer DEVICE (which will only
> happen when the various vendors upgrade to the Printer MIB V2.0
> and fix conformance problems found in the recent interop testing),
> isn't the next natural thing to standardize the management of
> printer JOBs?
It's interesting you've chosen to draw this sequence of events here,
namely "first the Printer MIB, then the Job MIB". Yes, I agree with
you, but only in basic premise.
You see, after years of developing the Printer MIB, there is a
significant number of PMP participants who now believe the Printer MIB
itself is too complicated. (Steve Waldbusser, our original IETF
advisor and "mentor") repeatedly warned us about going too far in the
specification, both in terms of scope and depth. We appeared to have
ignored his warnings, and now we're paying the price.
Several of us in the JMP group believe the Job MIB has already crossed
the line of excessive complexity (quite some time ago), and we're now
trying to reign in some of that complexity, using this basic approach:
"Start off with a SIMPLE spec to get product in the market
and gain experience, then augment that work in the future."
However, whenever the group tries to seriously pare down the scope
and model of the Job MIB, Tom always seems to ignore these concerns
and continues to push on with a "let's get it all in on the first
draft" approach. (How many times have we heard Tom say in a JMP
meeting "I'd like to revisit XXX..."?)
I'm hoping other JMP participants state their views on this topics.
This is not meant to be a "bash Xerox" or "bash Tom" effort, but
rather a serious plea to SIMPLIFY the spec and come to CLOSURE.
The true success of a standards effort is the IMPLEMENTATION of
that standard within products available in the marketplace. Many
of us in the JMP believe the current model is too complex, and
will result in a "sparse" implementation for the average product.
Interoperability is key, which usually translates to lots of
products offering the same capabilities. As it stands now, the
current Job MIB design does not promote this.
...jay