Non-Atomic components of Data Flow Diagrams: Stores, Persistent Flows, and Tests for Empty Flows

It has been shown in Symanzik and Baker (1996a) that a particular subclass of Formalized Data Flow Diagrams (FDFD's) is Turing equivalent. We call this Turing equivalent subclass of FDFD's persistent flow-free Reduced Data Flow Diagrams (PFF--RDFD's). PFF--RDFD's do not contain persistent flows, reference only values whose types have finite domains, and have enabling conditions that contain no tests for empty flows. In addition, FDFD's do not contain (direct) representations of stores. This raises the question whether any of these common features of traditional Data Flow Diagrams elevates the expressive power of FDFD's, or whether the various subclasses have the same expressive power as FDFD's with these features. This paper addresses this issue of whether persistent flows, arbitrary domains, tests for empty flows or stores are essential features with respect to the expressive power of Formalized Data Flow Diagrams.