AN AUTOMATIC PROGRAMMING SYSTEM SPACE WITH HIGHLY VISUALIZED AND ABSTRACT PROGRAM SPECIFICATION

Citation
M. Harada et T. Yoshimizu, AN AUTOMATIC PROGRAMMING SYSTEM SPACE WITH HIGHLY VISUALIZED AND ABSTRACT PROGRAM SPECIFICATION, IEICE transactions on information and systems, E78D(4), 1995, pp. 403-419
Citations number
NO
Categorie Soggetti
Computer Science Information Systems
ISSN journal
09168532
Volume
E78D
Issue
4
Year of publication
1995
Pages
403 - 419
Database
ISI
SICI code
0916-8532(1995)E78D:4<403:AAPSSW>2.0.ZU;2-1
Abstract
In this paper, it is stated that visualization and abstraction of prog ram specifications can be highly integrated on the basis of decision t ables and condition expressions. In order to demonstrate this idea, we developed an automatic programming system called SPACE: SPecification Acquisition and Compiling Engine. SPACE is designed to ease the produ ction of business data processing program. SPACE has functions both to support the creation of visual program specifications and to generate COBOL programs according to the input program specifications. To visu alize program specification, SPACE design windows are comprised of two diagrams and four tables in a formal similar to the conventional deta iled design sheets. To represent module functions, in particular, a vi sualized computation model called a decision table is used. All the po ssible execution states of a module are represented by combining the s tate function called condition expressions. The condition expressions represent the typical file processing patterns in very familiar form t o actual business application designer. They do not simply give functi on values; each of them carries out implicit attached procedures accor ding to the characteristic I/O control logic for business data process ing. Hence users can describe program specifications concisely by desi gnating merely the condition expression instead of the detailed I/O co ntrol logic. This paper uses sample descriptions of stock control prob lems to explain how visualization of computation and abstraction of al gorithm can be integrated and formalized on a basis of a decision tabl e and a condition expression. Also the paper describes how to generate programs from visual specifications.