Hybrid aspects for weaving object-oriented functionality and rule-based knowledge

摘要

Software applications often consist of implicit knowledge for making decisions or giving advice in addition to object-oriented functionality. A rule-based system can be employed for representing and reasoning with this knowledge. Although several hybrid systems exist that combine object-oriented programming and rule-based reasoning, a survey we conducted reveals that both paradigms are not well integrated and programs are tightly coupled.We propose hybrid aspects for integrating object-oriented programming and rule-based reasoning. As expected, hybrid aspects specify join points where normal execution is interrupted and advice is executed. However, since two different languages are involved, we need join point models for both and advice that activates both. As such, we complement a simple join point model for object-oriented programming with a join point model for rule-based reasoning. Hybrid advice is independent of the interrupted language and supports sending messages as well as activating rules. It uses values of either language transparently.We present OReA, an implementation of hybrid aspects for weaving Smalltalk and a rule-based system. We discuss and illustrate two applications of hybrid aspects.