Donald Bren School of Information and Computer Sciences

Closed applicative languages (CALs) are a highly parallel and semantically appealing models of computation, but they do not address process and resource related issues in computing; i.e.:

- they do not express histories of computation.

- they cannot describe interprocess communication.

This dissertation defines a new model, CFFP, derived from a CAL, FFP, which addresses these issues. In CFFP, the semantics of FFP are relaxed to allow a computation to persist over a series of actions, and explicit cycles are added to allow bidirectional communication between loci of computation. CFFP retains the appealing characteristics of CALs, and addresses process and resource related computing issues.