Visual parallel programming via petri nets

The objective of this paper is to determine the feasibility and produceability of a Petri net based CASE tool, thus the majority of this paper is the implementation and explanation thereof. The requirements of this implementation are as follows: a. Graphical input and manipulation of a Petri net. b. The generation of a formal definition from a given graph. c. The generation of code from a given formal definition. d. A CASE library to support parallel structures. During the implementation of this thesis it became apparent that a static dictionary for the CASE library would not be sufficient and a means to allow for customizing was required. Providing this programmability was not the topic of this thesis, but has become a substantial facet and possibly is a topic in and of itself. I leave it to the reader/user to explore the extents of that realm. The Apple Macintosh, due to its graphic ability and availability was selected as the platform for this implementation. This version of the program requires a color Macintosh and does not support many of the Macintoshs standard features. The implementation presented in this document meets the above requirements, and has produced executable code. At present the program consists of two separate applications that take the user from Petri net to executing code. The first handles all graphical input and editing of the Petri net and culminates in the generation of the formal definition. The second program generates the code from the formal definition and the library language. The separation is required for the later to be easily transferrable to any computing platform while the graphical interface is machine specific. Due to the requirements of the supporting NASA project, the code generated is Occam, and the target processor is the Inmos Transputer