Summary of Technical Progress, Software Modeling Stuides.
Interim rept. 1 Jan-30 Jun 76,
POLYTECHNIC INST OF NEW YORK BROOKLYN DEPT OF ELECTRICAL ENGINEERING AND ELECTROPHYSICS
Pagination or Media Count:
This report presents the progress of this research. Subjects of investigation were error generation and seedingtagging models, measures for the evaluation of software, analytical data selection methods for program testing, modular programming techniques, methods for finding feasible program paths, statistical program testing and proving, and methods for automatically testing every program path. Work has been completed on development of error generationmanpower deployment models to describe the error correction process in terms of error generation and correction rates as well as the number of man-months spent in debugging. Work has also also been completed in the development of seedingtagging techniques to estimate the number of software errors and related statistical quantities measures for comparing programs, such as accessibility, testability, and testedness, based on defining a program as a set of executible modules and a method for selecting test data sets for a program based on determining the interrelationships among program variables. Work still in progress includes development of techniques to interactively write programs using stored library modules andor user supplied code, investigation of a satisfactory algorithm to estimate the number of feasible paths in a program, development of a statistical theory for program testing and proving based on using a strategy of both testing and journal proving using Blacks model, and implementation of a PL1 driver to automatically test every possible path and catch errors of any PL1 program, subject to some minor constraints.
- Computer Programming and Software