Are we still in agreement with Ron's strategy that we need to do the
mappings BEFORE we forward the MIB to the IESG to become a proposed
standard?
By "mapping" I mean how does a JMP agent respond to a Get for JMP object
xxx or JMP attribute yyy when instrumenting a device that supports the
particular job submission protocol, i.e., how does the agent map from one or
more job submission protocol attributes to return the value of the xxx JMP
object or yyy attribute when the agent receives a Get operation for the JMP
xxx object or yyy attribute?
I think that this is a very important strategy. The printer vendors who
will be putting the Job Monitoring MIB into printers have a problem with
changes, because we have customers who will have written monitoring code
that use the MIB. So we want to minimize the incompatible changes when the
document goes from proposed to draft (additions are fine and deletions
that no one implemented are fine). Doing the mapping is one way to reduce
the number of incompatible changes, because the mapping is really a
"paper bake off". This "paper bake off" will flush out ambiguities as
we all look at each others mappings and ask, "Why did you map to that submission
protocol attribute to that JMP object/attribute? I would have though that
you would have..."
I hope that we can use some of the time Thursday night or Friday to finish
some more mappings, if they are not forthcoming, and compare one anothers
so that we can forward this MIB right after this meeting.
Comments?
Tom