Ira McDonald wrote:
> Hi,
>> Per our discussion at yesterday's IPP WG telecon,
> I looked up these two operations, just now:
>> Restart-Job - section 3.3.7 in RFC 2911
> - MUST not change job-id - eek!
> - moves job from completed state to pending
> or pending-held - violates state diagram
> - BAD idea for accounting and security
Not really - accounting can keep track of the job being printed
multiple times and keep the counts/quotas/etc. in sync.
As for security, it is no more or less secure than any other IPP
operation.
> Reprocess-Job - section 4.1 of RFC 3998
> - MUST copy job attributes and create new Job
> with new job-id (returned in response)
> - fixes the awful Restart-Job bugs
>> But, as we discussed yesterday, Reprocess-Job
> means that the system (spooler or printer) must
> keep all job attributes AND all job data for the
> entire job history duration - costly and dubious.
Keep in mind that the environments we are looking to make
it required already have large resources for this, so it
is not unreasonable to store the data and attributes.
Also, this can be a configurable option - CUPS, for example,
offers an option to determine whether job data is preserved
and how many jobs are preserved. Future versions will offer
this as a per-printer and per-job option as well.
> Conclusions:
>> (1) We really should DEPRECATE Restart-Job
> in IPP/2.x Profiles and explain the rationale.
Given that there is no other way to retry a failed print job,
I would be against complete deprecation but OK with
clarification or partial deprecation (i.e. OK to restart a
canceled, stopped, or aborted job, not OK to restart a
completed job...)
> (2) Reprocess-Job is technically OK, but costly.
> Should remain OPTIONAL, even for IPP/2.2
> (Production Printing).
I would argue that production printing already requires large
resources to support things like document objects and the tight
integration with the print device, so Reprocess-Job should be
REQUIRED for IPP/2.2 and OPTIONAL for IPP/2.0 and IPP/2.1.
--
______________________________________________________________________
Michael R Sweet Senior Printing System Engineer