ASQF Meeting Special Interest Group "Automotive Baden-Wuertemberg"
Statistical analyses play an important role in test processes: They support early detection of bugs in software systems. Many software systems are developed as configurable systems to meet various requirements on variant platforms. Therefore, developers often use preprocessors (e.g. #ifdefs) for software development. #ifdef annotated source code is configured with configuration options to produce tailored variants of software systems at one click. Existing configurable systems such as Linux Kernel offer up to thousands of configuration options to produce myriads of variants. This large diversity of variants creates new challenges for the application of statistical analyses: How to efficiently use statistical analyses for systems with up to 2^6000 variants? This presentation will introduce two approaches for efficient analyses: sampling measures and variability-aware analyses. Both approaches will be compared to each other, considering different criteria (performance, effort and bug coverage).
Presenter: Jörg Liebig is Senior Consultant in the Method Park Variant Management Team and consults his customers on topics such as variant management in software and systems engineering. He started in 2008 to concentrate on this topic and on challenges of software product lines. For his dissertation “analysis and transformation of configurable systems”, he received the software engineering award of the German Ernst-Denert Foundation. Jörg Liebig is author of numerous international publications and regularly gives lectures at conferences and professional group meetings. He also focuses on software and systems architecture and code complexity measuring, in software development. Jörg Liebig is iSAQB® Certified Professional for Software Architecture – Foundation Level and ISTQB® Certified Tester – Foundation Level.
February 23, 2017, 6 p.m.
Method Park Stuttgart