A systematic judgement to automated programming contracts generation

Abstract

Contracts provide a pre-emptive approach in identifying programming errors at run-time using assertions or by formal Static analysis tool or Manual source code reviews. They describe the expected software behavior. Contracts written by developers have a greater error detection ability than the generic ones that are created automatically but may involve strenuous efforts for larger sized source codes. The intent of this paper is a concise study of prevalent approaches in the generation of contracts and to put forward an approach to derive programming rules for real-time concurrent Java source code automatically with reduced efforts. The proposed method extracts the scalar variables and computed constants from Static program code analysis, then identifies various dependencies dynamically and generates the declarative contracts automatically by Decision tree modeling of computed dependencies. These rules can then be utilized for software Verification.