Metalevel decomposition in AL-1/D

Abstract

Research has shown that metalevel architectures and the concept of reflection are useful for modifying programming systems dynamically in a controlled way. To modify the system flexibly, it is necessary to represent various abstraction levels, from the programing language level to the OS level, as well as user's multiple views, such as the view where the distributed environment is transparent and the view where that is not transparent, in a programming system. In traditional reflective systems, it is, however, difficult to represent these aspects of the system because these systems are modeled by one metalevel. To overcome this problem, we have proposed a new reflection framework: Multi-Model Reflection Framework (MMRF), and implemented a programming system AL-1/D based on this framework. This paper gives a clearer definition of MMRF than in our previous paper. MMRF is a framework for decomposing a metalevel into multiple conceptual models according to the abstraction levels and user's views. These conceptual models may overlap each other in their functions and resources. The decomposed models should run concurrently because models represents system components running concurrently in a system. The definition of MMRF includes the conditions to enable models to run simultaneously without violenting consistently. The structure of a model includes information to decide whether or not these conditions are satisfied.