Towards a Reflective Implementation of Object Petri Nets

This paper presents the concept of reflection, where a program modifies its own interpretation. A number of applications of reflection are examined and it becomes apparent that the technique is particularly applicable to the implementation of object-oriented languages. The paper then considers how the technique can be applied to the implementation of Object Petri Nets (OPNs), a formalism for modelling concurrent, object-oriented systems. It turns out that reflection provides a simple and efficient implementation scheme for the compositionality constructs of OPNs. Further, it allows flexible scheduling strategies to be employed in the simulation of OPNs.