Project Title:

Semi-Automatic Data Structure Selection

06.04-6871
NAS2-13174
Semi-Automatic Data Structure Selection
Kestrel Development Corporation
	3260 Hillview Avenue
	Palo Alto, CA 94304
		Lee Blaine
(415-493-6871)

Abstract:

	This project addresses efficient prototyping of reliable software systems and deals
with the four intertwined issues of reliability, reusability, coping with changes
to the prototype, and generating efficient code from specifications. The company
has developed a language for describing implementations of data structures and a
data-type refinement environment (DTRE) for transforming to code the specifications
annotated with implementation directives. In order to generate efficient code from
specifications, the implementations of data structures must be carefully selected.
Since fully automatic selection of implementations is beyond current technology,
a semi-automatic system for data structure selection (DSS) will be developed. With
this system, the user will supply some implementation directives and the system will
determine the remainder. The DSS together with the DTRE will lead to critical, asymptotic
improvements in the efficiency of rapid prototypes based on executable specifications.
Just as critical is the increase in reliability that comes from automating the derivation
of efficient code from specifications.

Potential Commercial Application:

Potential Commercial Applications: DTRE is being considered for a commercial software
development environment based on executable specifications and program synthesis.
DSS combined with DTRE could be applied for generation of early, reasonably efficient
prototypes in requirements analysis and as a design and coding tool.