Added the proof of correctness of the work-conservation transformation for an ideal uniprocessor schedule.

Fix a typo and note Pierre Roux's comment on disjuctions vs inductives.

See also: RT-PROOFS/rt-proofs#54 (comment 42229)

In anticipation of future reorganization efforts by Sophie, there's no
point in keeping the two kinds of definitions separate.

Move everything into the new namespace starting with 'prosa' rather
than the bland 'rt'.

The main restructuring thrust is nearing completion, so let's get rid
of the `restructuring` namespace.

improve comments, fix names, move some stuff around

spell checker: add 'supremum' to dictionary

...and define a job's total suspension time.

For improved parallelism and nicer documentation.