There's a serious problem with Harry's suggestion below:
"In the absence of a firm commitment to achieve this your
recommendation to mandate (for WIMS) only essential elements
that are known to be well supported is a good one."
There is NO common set of well supported objects in Printer MIB.
I've been looking at MIB walks of Printer MIB for the last decade.
Each manufacturer and each _product_ puts more or less effort into
getting the Printer MIB objects "right". So far, I've never
personally reviewed a MIB walk of a Printer MIB implementation that
did not contain at least one technical error (most have many).
The 80/20 rule could easily become the 20/1 rule (one percent
of objects consistently implemented). Counter and status objects
are particularly flaky in most products.
Nonetheless, the Printer MIB (in some form) is one of ONLY two widely
implemented standards in the printing industry (the other being the
IEEE 1284 spec).
Note that consistency of implementations is probably worse for IPP.
But since the last IPP bakeoff preceded almost all the extensions,
we don't really know.
Ira McDonald (Musician / Software Architect)
Blue Roof Music / High North Inc
PO Box 221 Grand Marais, MI 49839
email: imcdonald at sharplabs.com
From: owner-wims at pwg.org [mailto:owner-wims at pwg.org]On Behalf Of Harry Lewis
Sent: Friday, September 02, 2005 12:23 AM
To: wamwagner at comcast.net
Cc: 'wims at pwg.org'
Subject: Re: WIMS> Reflections on a (MIB) walk
Bill, I think ALL standards participants can stand a regular healthy
reminder about the 80/20 rule... how we should keep focused on standardizing
well the 20% of objects, elements or attributes that will provide 80% of
the key interoprable function. Your findings, although not specific, do act
as such a reminder.
I have been urging in past Plenary and Steering Committee meetings that one
of the most beneficial services the PWG can provide is further Printer MIB
interop testing. In the absence of a firm commitment to achieve this your
recommendation to mandate (for WIMS) only essential elements that are known
to be well supported is a good one.
Chairman - IEEE-ISTO Printer Working Group
IBM Printing Systems
wamwagner at comcast.net
Sent by: owner-wims at pwg.org
09/01/2005 06:37 PM To"'wims at pwg.org'" <wims at pwg.org>
SubjectWIMS> Reflections on a (MIB) walk
I just had occasion to MIB walk through a few implementations of a table or
two in the printer MIB, in several standard products including some from the
most predominant members of the industry. Remembering the discussion and
consideration that went into the objects, it is an eye opener to see with
what carelessness (or is it intentional sabotage) these things are
implemented. Granted, looking at the spec after all this time, there are
some ambiguities. But it often looks like someone just thru in arbitrary
values for some objects. "Mandatory" means nothing. Since most management
applications use private mibs, manufacturers appear not to put any effort
into validating the general MIB.Or else the objects are of so little use
that no one cares if the values are valid.
Aside from being discouraging, I think there is a lesson here that we may
apply to our on-going work. Keep things simple; don't expect that anyone
cares if you label an element mandatory; if they have a use for it, it
will be implemented. Let's not slavishly import things from the printer MIB
or other sources, or even strive for completeness. The criteria for
inclusion must be the clear need for the element as evidenced by existing
use or indisputable future requirement.
As for having a Printer MIB 2 cook-off (a non-proprietary bake-off), getting
consistent implementations may be important to some people, but apparently
not to manufacturers. I doubt anyone would come.