Semantics-Driven Compiler Synthesis

生成高质量编译器的一个关键因素是创建编译器和运行时机器使用的主要数据结构：分别是符号表和主存储。早期对lambda演算表达式的单线程和堆栈单线程属性的研究展示了如何检测、提取和实现这些数据结构。这项研究的主要部分将是将单线程技术集成到一个工作的编译器综合系统中。将研究数据流分析算法在改进合成编译器的静态语义处理和目标代码生成方面的作用。特别是，语言定义的流程分析应该产生将定义的语义代数转换为更有效的表示的信息。

A key factor in the generation of quality compilers is creation of the primary data structures that a compiler and run-time machine use: the symbol table and primary storage, respectively. Earlier research on single-threading and stack single-threading properties of lambda- calculus expressions shows how to detect, extract, and implement these data structures. A major portion of this research will be the integration of the single-threading techniques into a working compiler synthesis system. The role of data flow analysis algorithms in improvement of static semantics processing and target code generation in synthesized compilers will be investigated. In particular, flow analysis of a language's definition should yield information that directs transformation of the definition's semantic algebras into more efficient representations.