Abstract This paper describes a tool for real time software which can present debugging information with no particular debugging runs or a low degree of repetition of debugging runs. This tool is useful for locating problems in later test phases or after the delivery of large scale real time software which cannot be developed by a single individual and which performs dead line control. Many interface, timing, and performance errors often remain uncovered until later in the life cycles of large scale real time software. In order to determine these problems with conventional tools, it is necessary to repeat debugging runs almost in the same conditions as those when errors occur. Since such repetition of debugging runs is very expensive and because errors are sometimes covered again, it reduces productivity and reliability markedly. The tool proposed here avoids such shortcomings of conventional debugging tools. The tool concept is derived from experience with error analysis obtained during the development of train traffic control software, which is typical of large scale real time software. This tool is expected to considerably improve the productivity and reliability of large scale real time software.
[1]
Edsger W. Dijkstra,et al.
Notes on structured programming
,
1970
.
[2]
Ralph Grishman.
Criteria for a debugging language
,
1971
.
[3]
Edsger W. Dijkstra,et al.
The structure of the “THE”-multiprogramming system
,
1968,
CACM.
[4]
Robert L. Glass,et al.
Persistent Software Errors
,
1981,
IEEE Transactions on Software Engineering.
[5]
Brian Randell.
Reliable Computing Systems
,
1978,
Advanced Course: Operating Systems.
[6]
Per Brinch Hansen,et al.
Operating System Principles
,
1973
.