Empirical Evaluation of Multiparadigm Software Development

点击翻译按钮获取中文摘要

During software development, programmers employ a range of different program representations, including pseudocode, data flow diagrams, finite-state automata, control flow charts, and many others. Programmers may also create their own custom representations or languages to express their particular approach to a problem. Conventional development environments force programmers to translate these different paradigms into a single textual representation that the computer can understand, and to translate between this representation and their conceptualizations during coding, debugging, and maintenance. Multiparadigm development environments offer the opportunity for programmers to work directly with their conceptualizations, to work with a program using whatever representation is most natural; such environments eliminate the need to perform mental translations between programmer concepts and executable code. The proposed research is to test the validity of multiparadigm programming by designing experiments to answer the following kinds of questions: Does it make programs easier to design or code? Does it make programs or designs easier to understand? Does it simplify the process of enhancing existing programs to increase their functionality or their performance? Does it make it easier for programmers to track down and eliminate bugs? Do software developers prefer it?