Verifying Compiler Algorithms

9404646魔杖因为编译器几乎是每个计算机系统的可信组件，所以有可信的证据证明它们的正确性是很重要的。该项目将继续研究根据程序设计语言的正式规范开发和验证编译器。一个主要的活动领域是将语义学和编译器正确性证明扩展到用于并行计算的语言。还将努力研究面向对象编程的基本理论。有望与相关研究进行交叉受精。优化和具体化技术将被应用于并行设置，而并行设置将为优化研究提出新的问题。类似地，类型理论中的现有活动主要关注类型推理算法的复杂性。预计其中一些结果将在面向对象环境中应用，而面向对象程序设计的研究将在类型理论中提出新的理论问题。***

9404646 Wand Because compilers are trusted components of almost every computer system, it is important to have trustworthy proofs of their correctness. This project would continue research on the development and verification of compilers based on formal specifications of programming languages. A major area of activity would extend semantics and compiler correctness proofs to languages for parallel computation. Some effort would also be directed toward the study of fundamental theories for object- oriented programming. Cross-fertilization with related research is expected. Techniques for optimization and concretization would be applied in the parallel setting, and the parallel setting would raise new problems for studies of optimization. Similarly, existing activity in type theory is concerned primarily with complexity of type inference algorithms. It is expected that some of these results to be applied in the object-oriented setting, and that studies of OOP would raise new theoretical questions in type theory. ***