ATTRIBUTE GRAMMARS AND DATAFLOW COMPUTING

作     者：PAPAKONSTANTINOU, G TSANAKAS, P 

作者机构：Natl. Tech. Univ. Athens Dep. Electr. Eng. Comp. Sci. Div. 15773 Zographou Greece 

出 版 物：《INFORMATION AND SOFTWARE TECHNOLOGY》 (信息与软件技术)

年 卷 期：1988年第30卷第5期

页      面：306-313页

核心收录：

学科分类：08[工学] 0835[工学-软件工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

主　　题：attribute grammars data flow computing software engineering 

摘      要：Attribute grammers (AG), developed as a tool for formal specification of programming languages, now are used for applications in many other areas. Most recently, AGs have been used for knowledge representation, logic programming, and as a data-flow language. AGs and data-flow languages bear numerous similarities, which supports the claim that AGs can be efficiently used as data-flow languages. The primary advantages in the use of AGs as programming languages include: 1. AGs use a nonprocedural formalism. 2. AGs allow hierarchical development. 3. Several AG evaluators are available. 4. AGs formalism can be viewed as a higher level language. Using AGs as a nonprocedural programming language can overcome the problem with data-flow architectures needing to be programmed in special high-level languages. A new methodology for extracting the inherent parallelism from a computational task is presented, in which AGs are used to construct new forms of the computing problems suitable for execution on data-flow machines. The procedure has been automated and tested for several computing problems.