This work aims to analyze imperative, object-oriented programs for type signature information, which can help automate several programming tasks that are currently performed manually. Type signatures summarize the behavior of a program's methods, functions, or procedures; while such information has been used to advantage in the functional programming language community, little has been done to investigate how popular, object-oriented languages could benefit from use of type signatures.
The research will improve the quality and efficiency of software navigation and reuse; software testing and demonstration; and construction of graphical and nongraphical user interfaces. Novel aspects include use of data flow analysis to refine and improve the precision of type signatures extracted from imperative languages and analysis of method call-sequences that accomplish or approximate a given type signature.