3 Comparison

The ideas we have presented in this short paper will perhaps go some way towards convincing the reader that processes are complementary to, if not an alternative to, objects for the design and description of behaviour in distributed systems. The description technique given here is perhaps not as powerful as some object-oriented techniques (for example [Bre96]) but it does seem to have some advangages when it comes to description of concurrent behaviour. Perhaps, for distributed systems built from components, we need object-oriented techniques for the description of passive objects and techniques similar to those described here, based on process algebras, for description of active objects and concurrent combination of components.

An interesting question for formal methods remains: when we should use process algebras and when we should use state based methods of the Z, VDM, Object-Oriented kind to describe components and their composition into systems? For validation of system level descriptions it does seem that state-space search and model checking [Cla91,Ip96,Jac96] could play a major role. In particular as a practical way to remove design faults early in the system design process. We have shown how certain kinds of components can be described using pi-calculus and briefly described our method of validating these components. The trend in the component supply business seems to be towards versatile, dynamically-reconfigurable components such as applets, oblets, servlets and generally downloadable run-time add-ons.

This trend presents significant challenges for the formal methods community. The componentology provided by technologies such as Java (but also Corba, and OLE, and ActiveX) are state-of-the-art as far as realistically usable components go. In this short paper we have attempted to show that formal methods such as the pi-calculus are mature enough to describe such componentology. Perhaps we have succeeded in some small way. But, I am happy to conclude, there are still a significant number of major challenges ahead.