The notion of a valid preemption model for *jobs* should
not depend on any task abstraction. Split the two notions.

This is a purely accidental dependency; purge it.

With proper re-exports everywhere, this is no longer necessary.

The two concepts are tightly coupled and it does not make sense
to mix and match them. Therefore, it's cleaner and easier to
understand if both aspects of the preemption model are instantiated
in the same module next to each other.

No need to keep the two separate as they are tightly coupled anyway,
and the joint file doesn't become too long.

The property is a noun: a model of task sequentiality.

All "kinds of schedules" and associated validity conditions are being
collected in model.schedule.

EDF is defined in terms of jobs, whether or not those jobs stem from
tasks. Make sure our definition is general enough to capture the
possibility of a set of jobs without associated tasks scheduled by
EDF.

The Prosa notion of priority-driven scheduling depends on the
preemption model. That is ok; any theory that doesn't want to deal
with non-preemptive sections can simply Require Export the fully
preemptive model and be done with it.

With this patch, the model module is finally completely independent of
any definitions or proofs in the analysis module.

Move proof of readiness model property to analysis module.

By our newly adopted convention, the model/ module should not depend on analysis facts.

Related issue: #37

Now that we have non-ambiguous dependencies, parallel builds should
work without issue.

This fixes all warnings about ambiguous module names and resolves #49.

It also highlights that we still need to cut down on superfluous
Require Import commands in recently ported files.

...and "Defined" can end one.

This can be useful for learning / teaching.

Add instantiations of aRTA for
(1) fully preemptive,
(2) fully non-preemptive,
(3) limited preemptions,
(4) and floating non-preemptive regions EDF models