Please note: This PhD seminar will take place in DC 2310.
Rafael
Ferreira
Toledo,
PhD
candidate
David
R.
Cheriton
School
of
Computer
Science
Supervisor: Professor Jo Atlee
Analyses of a software product line (SPL) model typically report variable results annotated with logical expressions to indicate the set of products for which each result holds. These expressions can be complicated and difficult to reason about when the SPL has lots of features and derivable products. Previous work introduced a visualizer for graphical analysis results, that highlights the results that apply to specific products; but their work was weakly evaluated.
In this paper, we report on a controlled user study that evaluates the visualizer's effectiveness in helping the user (1) understand variable analysis results and (2) compare the analysis results of multiple products. Our findings indicate significant improvements in both efficiency (30%) and correctness (20%, but only on tasks where participants use the visualizer); significant reductions in cognitive load on tasks involving multiple products: and insignificant reductions in cognitive load on tasks involving a single product.