AutoRA documented use case for online experiments

[musslick]

Include a use case that describes the use of AutoRA just for running online experiments.

example ideas:

• task switching

benefits for the user:

• no server needed
• no jspsych knowledge needed
• just python

learning goals for the user:

• generating experiments with sweetbean
• counterbalancing with sweetpea (optional)
• running experiments with autora

Also consider re-framing autora functionality in main doc, to make closed-loop part of it but not the central things.