multiplasms/multischedulers

[straszheim]

several here:

• two plasms that contain the same node N. N might be called by two different schedulers with two different sets of inputs.
• two plasms, one scheduler. maybe the capabilities here are scheduler-dependent. How important is it to keep a consistent interface across schedulers?

probably the same issues are here as do in 'black box' cases... strands? locks?

[ethanrublee]

Not really relevant right now. This should be switched to a scope context in the graph.