Predicate Collection Classes

Collections are at the heart of every program. Modern programming languages have acknowledged this fact by including increasingly better expression mechanisms for manipulating collections of data and objects. When existing objects are selected as elements of collections, often there is an implicit intention that those objects, by means of having been selected, take on new roles. Such is the case, for example, with shared data in distributed simulations, which changes over time, and which distributed components may use for different purposes as the simulation unfolds. This paper presents the concept of Predicate Collection Classes (PCCs). PCCs are both classes and specifications of how to select and reclassify objects from existing collections. We describe the informal semantics of PCCs and demonstrate how they can be used to express several iterative algorithms that make heavy use of collections. Finally, we summarize our use of PCCs in a framework for distributed simulations.