An emerging topic in software engineering and data mining, specification mining tackles software maintenance and reliability issues that cost economies billions of dollars each year. The first unified reference on the subject,
Mining Software Specifications: Methodologies and Applications
describes recent approaches for mining specifications of software systems. Experts in the field illustrate how to apply state-of-the-art data mining and machine learning techniques to address software engineering concerns.
In the first set of chapters, the book introduces a number of studies on mining finite state machines that employ techniques, such as grammar inference, partial order mining, source code model checking, abstract interpretation, and more. The remaining chapters present research on mining temporal rules/patterns, covering techniques that include path-aware static program analyses, lightweight rule/pattern mining, statistical analysis, and other interesting approaches. Throughout the book, the authors discuss how to employ dynamic analysis, static analysis, and combinations of both to mine software specifications.
According to the US National Institute of Standards and Technology in 2002, software bugs have cost the US economy 59.5 billion dollars a year. This volume shows how specification mining can help find bugs and improve program understanding, thereby reducing unnecessary financial losses. The book encourages the industry adoption of specification mining techniques and the assimilation of these techniques in standard integrated development environments (IDEs).