Make it possible to add/remove processors at runtime

The way processors are registered to Dagger is slightly ugly, and has the issue that we can't easily unregister processors later (because the callbacks in the list aren't uniquely identified). We should require identifiers be passed along with each callback so that it's trivial to remove them later (and we can unique on the name, which prevents double-registering the same processor).

Additionally, in the same vein as what #147 accomplished, we should make it possible to dynamically add and remove processors from a worker while the scheduler is executing. The user could then safely transition a computation across different groups of resources as their needs evolve.