So we've more or less settled that in the common case, there HAS
to be one link in the chain that can look at both ends of the job
at the same time, and understand enough about it to figure out
what its requirements are.
Print-by-URL is irrelevant to this discussion, since there
the distinction between parameters "before the URL" and "after the URL"
is pretty slim; they are both communicated to the printer before
the URL is dereferenced.
So - the spooler is there; just choose which side of the IPP-defined
interface you want it.
Harald A