In discussing the host-to-device requirements, we came up with a requirement
that the printer be able to feed back information about whether it was
choked up with data or needed more data for the current job.
So we could have events like:
Slow down data transfer
Speed up data transfer
There is an even more important event for a device to send to a host:
Ready for another job
This event is useful for a number of configurations:
1. Such an event could anticipate the completion of the current job, so that
the devices could overlap jobs. Printers that completely interpret a
job or document before marking would want to indicate that they are ready
for the next job much sooner that printers that mark as they interpret.
Printers with a long output paper path may want to ask for the next job
while the output paper path is being emptied, so that the printer doesn't
slow down between jobs. A host that has a job would then be advised
that this device could accept it now. If the host did not have a job,
the host could still keep an indication that this device is a candidate
for a job when a new job is submitted to the host.
2. This event is especially useful for the IPP fan-out case of a server/host
that controls multiple devices represented by a single IPP Printer object.
3. This event is also useful for the simpler case where a device is
controlled by more than one host and the device wants to indicate that it
is ready for another job from any of the hosts (or from a particular one,
if the device is trying to be fair).
I just read the current Notification Requirements and it is focused on
events that end users needs, so the above event are ones that hosts
need, not end-users. But these events are NOT events that administrators
need. Usually in the IPP WG, we have been making the distinction between
end-users vs. system operators/adminstrators, not between end-users versus
servers/hosts that are submitting jobs to devices on behalf of end-users.
So are the above events in-scope for our IPP notification effort or
out of scope? I think the answer depends on whether the IPP WG is going
to tackle the host-to-device requirements for IPP.
Tom