REARCH: sset_loop reveals a design problem with argument handling

[noblem]

See the comment made on 6/8 while closing out #29. Extra code was added to flow_start to safeguard against edge cases that some parameters may not exist in the args object (even though they have default values specified in the argparse boilerplate). This happens because the argparse parse_args() method is not being called for the respective CMD that sset_loop() is attempting to invoke.

It is not desirable that all CMDs (callable by sset_loop() etc) be bullet-proofed in this manner, because the issue can be avoided entirely by streamlining the way args are parsed & passed to CMDs. When Tim was still here I'd discussed with him my desire to simplify and generalize the arg parsing, so this issue is just a placeholder for that effort.