Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Optimizing quality assurance activities by considering metrics and context factors

 
: Kremer, Stephan
: Rombach, H. Dieter; Liggesmeyer, Peter; Elberzhager, Frank

Kaiserslautern, 2012, X, 77 pp.
Kaiserslautern, TU, Master Thesis, 2012
English
Master Thesis
Fraunhofer IESE ()
quality assurance; defect prediction; quality metric

Abstract
Quality becomes more and more important in today's software projects. A considerable percentage of the total development effort is spent on quality assurance activities such as inspections and testing. However, synergy effects between the quality assurance techniques are often not exploited. Inspection results are not used to focus testing activities.
In this thesis, the Focused And Corrective Evaluation approach (FACE approach) - an integrated iterative defect prediction approach - is introduced. Through the use of inspection results, product and process metrics and consideration of the concrete project context, project files are ranked by their likelihood of containing defects. This prediction can be used to focus testing activities on the parts of the system that are most likely to contain defects. Once testing has been completed, the testing results are used to evaluate the performance of the prediction. Correlations between the prediction and the testing results are calculated and serve as a base for evaluation purposes. After the evaluation is complete, the results can be stored in an empirical knowledge database that serves as a reusable resource for future iterations of the FACE approach.
To demonstrate how the FACE approach can be applied, this thesis also introduces a software tool called FOQUS that implements the FACE approach and is designed to automate the prediction and evaluation steps. Visualizations enable intuitive representations of the mathematical concepts behind the FACE approach. Usage scenarios describe how FOQUS can be used to predict defects in project files and evaluate the prediction.

: http://publica.fraunhofer.de/documents/N-208421.html