Affiliation:
1. Carnegie Mellon Univ., Pittsburgh, PA
Abstract
The Ergo Attribute System was designed to satisfy the requirements for attributes in a language-generic program derivation environment. It consists of three components: (1) an abstract data type of attributes that guarantees attribute consistency, (2) a Common Lisp implementation which combines demand-driven and incremental attribute evaluation in a novel way while allowing for attribute persistence over many generations of a program, and (3) an attribute-grammar compiler producing code based on this abstract data type from a high-level specification. Our experience with three major applications (one being the attribute-grammar compiler itself) confirms that the overhead in storing and accessing attributes incurred by our implementation scheme is more than offset by the gains from the demand-driven, incremental, and persistent nature of attribution.
Publisher
Association for Computing Machinery (ACM)
Cited by
1 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献