Improving the implementation of high-level languages

点击翻译按钮获取中文摘要

The primary objective of my research consists in improving the implementation of high-level programming languages, and that of functional languages in particular. Programming has slowly been moving towards high-level languages, that is, towards simpler, more expressive, more powerful, more intuitive languages. Most of today's software projects would simply not be feasible if assembly languages were still the only languages available. Even if languages such as C and C++ are considered high-level compared to the assembly languages, they remain low-level compared to languages such as Haskell, ML, Scheme, and Java. While my primary objective relates to the implementation on high-level languages, I have other main objectives. I am interested in improving the implementation of languages, even those that are not high-level languages. I am also interested in increasing the reliability of software in general, not just the language-related tools. Even if I have a strong preference for high-level languages, lower-level languages are widely used nowadays and might continue to be used in certain applications for a very long time. Consequently, improving the implementation of these languages is important too. Regarding the reliability of software, it is obvious that it is not as high as it should. If cars, airplanes, buildings, and bridges were as (un)reliable as software, it would be catastrophic. Software production is clearly not as well established and as systematic as other production processes, e.g. in most engineering disciplines. Given my objectives, I intend to continue to conduct research on compiler technology, especially on analysis and optimization techniques, whether it is related to high-level languages or not. I also intend to contribute to the development of techniques that directly aim at improving the reliability and the security of software---like programming with contracts and static and dynamic formal software verification. Projects that I intend to realize include the static profiling of functional programs, the extension and improvement of adaptive static analysis, the development of a quick control-flow analysis, the application of our technique called Extended PCC to oracle-based PCC, the pursuit of the development of bit recycling.