As you noted, you cannot specify a serial component in PEP1. You must add the serial actions individually, couple them to an extender, and to a serial port.

IMO, the move to the component database of serial commands in PEP2 was not a very good idea as it forced users to add functions for each and every serial command they wanted to send in the database vs allowing ad-hoc testing via a single action. Yes, it does allow you to share components easily but the standardization of component function names is less likely on a serial component.