|
Software Requirements Analysis as Fault Predictor
|
|
Point of Contact
|
Dolores Wallace
dwallac@pop300.gsfc.nasa.gov
|
|
Dates
|
October 2002 - September 2003
|
|
Problem
|
Waiting until the integration and system test phase to discover errors leads to more costly rework than resolving those same errors earlier in the lifecycle. Costs increase even more significantly once a software system has become operational. We can a ssess the quality of system requirements, but do little to correlate this information either to system assurance activities or long-term reliability projections - both of which remain unclear and anecdotal.
|
|
Objective
|
Extending earlier work on requirements accomplished by the ARM tool, measuring requirements quality information against code complexity and test data for the same system may be used to predict specific software modules containing high impact or deeply em bedded faults now escaping into operational systems. Such knowledge would lead to more effective and efficient test programs. It may enable insight into whether a program should be maintained or started over.
|
|
Results
|
Paper & Recommendations Presenting Research Results.doc
Software Requirements Analysis_SAS.ppt
|
|
Keywords
|
ARM tool, software modeling, software defects, software development
|
|
Categories
|
Requirements Analysis
|
|
