This paper presents a tool-supported methodological paradigm for object-oriented software development, called monitoring-oriented programming and abbreviated MOP, in which runtime monitoring is a basic software design principle. The general idea underlying MOP is that software developers insert specifications in their code via annotations. Actual monitoring code is automatically synthesized from these annotations before compilation and integrated at appropriate places in the program, according to user-defined configuration attributes. This way, the specification is checked at runtime against the implementation. Moreover, violations and/or validations of specifications can trigger user-defined code at any points in the program, in particular recovery code, outputting or sending messages, or raising exceptions. The MOP paradigm does not promote or enforce any specific formalism to specify requirements: it allows the users to plug-in their favorite or domain-specific specification formalisms via logic plug-in modules. There are two major technical challenges that MOP supporting tools unavoidably face: monitor synthesis and monitor integration. The former is heavily dependent on the specification formalism and comes as part of the corresponding logic plug-in, while the latter is uniform for all specification formalisms and depends only on the target programming language. An experimental prototype tool, called Java-MOP, is also discussed, which currently supports most but not all of the desired MOP features. MOP aims at reducing the gap between formal specification and implementation, by integrating the two and allowing them together to form a system.
展开▼
