In December 1995, the Algebraic Methods in Language Processing conference (AMiLP'95) was held at the University of Twente, Enschede, the Netherlands. More details can be found on the AMiLP'95 website Below are copies of the abstracts regarding participation by Jim Jones.
Multi-axiom grammars have been introduced as an alternative to the single-axiom context free grammars and the all-axiom algebraic grammars for programming language specification. Multi-axiom grammars eliminate some of the limitations of the context-free grammars used for programming language specification and implementation. However, prior algebraic notions regarding multi-axiom grammars such as subgrammar, primitive subgrammar, quotient grammar, and grammar/language hierarchy, were based on limited use of non-axioms. In this talk, I will present some of these concepts and a naturally parallel algorithm for parsing programming languages specified by multi-axiom grammars. This algorithm is convenient for large program development because it can recognize any phrase in the language and thus facilitates incremental development of programs. The algorithm is efficient and can use the computation power of the parallel machines because it distributes the parsing task among a collection of smaller parsers associated with individual language layers. The algorithm is called a PHRASE parser, an acronym for its actions: Pass, Halt, Reduce, Accept, Shift, and Error. I will also demonstrate layered parsing on sample inputs using software that I have developed to study such activity, entitled MAGLAB (for Multi-Axiom Grammar LAB environment).
Multi-axiom grammars have been introduced as an alternative to the single-axiom context free grammars and the all-axiom algebraic grammars for programming language specification. Multi-axiom grammars eliminate some of the limitations of the context-free grammars used for programming language specification and implementation. However, prior algebraic notions regarding multi-axiom grammars such as subgrammar, primitive subgrammar, quotient grammar, and grammar/language hierarchy, were based on limited use of non-axioms. In this paper we use the mechanism of non-axioms to redefine some of these concepts and to develop a naturally parallel algorithm for parsing programming languages specified by multi-axiom grammars. This algorithm is convenient for large program development because it can recognize any phrase in the language and thus facilitates incremental development of programs. The algorithm is efficient and can use the computation power of the parallel machines because it distributes the parsing task among a collection of smaller parsers associated with individual language layers. The algorithm is called a PHRASE parser, an acronym for its actions: Pass, Halt, Reduce, Accept, Shift, and Error.