Those two look quite similar from the far....
Basic idea behind PlugIns is to extend functionality of the process model without touching the process model itself.
You could do so by process model callbacks, yet this doesn't scale as good for general changes.
PlugIns are somewhere in between using callbacks and changing the process model and can be configured individually for each machine if required.
In fact, TestStand reporting / database logging is just a plugin.
Does this help?
