At 11:14 05/27/97 PDT, Harry Lewis wrote:
>Even printers that don't spool or queue will typically "buffer" some number of
But isn't the definition of "queuing" on page 13, line 364, what you mean
by "buffering" some number of jobs?
Given the possibility of single page jobs... then some jobs can be
>pending. I guess we should ask ourselves what it means to have a "mandatory"
>state. Just because a state is Mandatory does not mean an implementation must
>"force" jobs to enter that state. If that were so... I'd REALLY be in favor of
>eliminating the NeedsAttention state! ;->
We keep getting confused by whether a MIB is talking about the printer
or is talking about what the agent shall show about the printer.
I claim that MIBs are all about the latter and nothing to do with
See my early answer to Ron about the completed state. Most printers
won't have such a state, the printer throws the job away as soon as it
is printed. Its the agent that has to keep the job information in the
MIB tables for the time indicated in the jmGeneralJobPersistence object.
If the 'completed' state were conditionally mandatory, then an agent would
only have to show jobs in the 'completed' state, if the device implemented
the completed state. But we aren't going to allow agents off the hook.
The agent has to implement the 'completed' state. Same for 'aborted'.
'canceled', is conditionally mandatory, because level 1 IPP doesn't requiure
the CancelJob operation. So we aren't going to force agents to show
a canceled state, except for printers that implement a CancelJob operation.
For needsAttention, is is impossible to design a device that handles paper
that doesn't need attention some time (unless it plants, grows, harvests trees,
makes paper, and fixes all jams). So making needsAttention Mandatory forces
the agent to show jobs in the needsAttention state, if the printer jams.
>>Harry Lewis - IBM Printing Systems
>>>------- Forwarded by Harry Lewis/Boulder/IBM on 05/27/97 11:59 AM ------
>>jmp-owner at pwg.org> 05/27/97 11:51 AM
>Please respond to jmp-owner at pwg.org @ internet
>>>To: jkm at underscore.com @ internet, jmp at pwg.org @ internet
>cc: ipp at pwg.org @ internet
>Subject: Re: JMP> jmJobState and jmJobStateReasonsTC [ISSUE: Are ther
>>If a simple output device implements IPP and doesn't queue or spool, wouldn't
>the jobs in that output device never be in pending? Such a device would
>refuse acceptance of another job, while it was processing its current job.
>That is why pending is conditionally mandatory in IPP and therfore also in
>>Would it help to indicate that if an implementation queues or spools,
>that it shall implement the 'pending' state jobs that are queued or spooled
>but are not yet processing?