Für eine korrekte Darstellung dieser Seite benötigen Sie einen XHTML-standardkonformen Browser, der die Darstellung von CSS-Dateien zulässt.

. .

SPLC15 Accompanying Webpage


This page provides accompanying material for the paper

Fault-based Product-Line Testing: Effective Sample Generation based on Feature-Diagram Mutation
Dennis Reuling, Malte Lochau, Johannes Bürdek, Udo Kelter, Serge Rotärmel.

Evaluation Results

All data presented throughout the paper as well as additional results can be found here:


Feature Diagram Mutation Operators

All mutation operators for feature diagrams can be downloaded here:


They are grouped according to their fault type. Please note that the Henshin rules themselves are to be found in the folder editrules.

Sample Evaluation/Generation Framework

Our tool implementation currently supports:
  • Evaluating samples based on feature diagram mutation
  • Adapting a given sample to detect priorly undetected mutants
  • Generate a sample from scratch using mutation
Supported feature model formats:
  • SXFM - Simple XML Feature Model Format
  • SPLX (used by the FM Generator BeTTy)
  • Our own EMF-based implementation of feature diagrams
We currently support comma-separated value files (.csv) as sample input files
(Generated by the SPLCA Tool).


The Evaluation/Generation Framework of SPLC2015 can be obtained from the following Update Site:



The used data sets as well as already generated Samples are included in evaluationexample.zip.
It can be imported as Eclipse Project and be used as exemplary input.
Please do not use whitespaces in file paths as this may cause problems.

Installation Requirements

Our Framwork is based on the Eclipse Luna Modeling Tools, which deploys all needed core plugins.


Step 0: Starting
For evaluating or generating a sample, there are two possibilities:
  • Select the featuremodel and/or a sample at hand (in a Eclipse project, e.g. the example
    evaluation project), open the context-menu and select Evaluate/Generate Sample:

  • Open a Wizard through New->Other:

Afterwards the wizard will guide through the process in three steps:
Step 1: Input Configuration

Here the given input files can be either selected or generated from scratch.
If first possibility for starting has been used, the files at hand are already selected.
Otherwise the user is free to tailor the input files to his needs.
Step 2: Mutation Configuration

The mutation process can be configured in various ways, e.g. the mutation operators to be used.
The operators are grouped by their fault type they introduce. For sample evaluation/generation
the Undetectable Operatory are to be omitted. They are provided for completeness.
Step 3: Output Configuration

Finally the possible run scenarios as well as output parameters can be configured.
Additionally to printing the results in console as well as a Messagedialog after completion,
one can select to write the resulting logfile to a given text file.
The user can additionally select to optimize the given sample. If a empty sample
is selected in Step 1, a new one will be generated from scratch using our framework.


  • Please do not use whitespaces in file paths as this may cause problems.
  • Please note that the implementation is a proof of concept.
    Depending on the feature model as well as mutation configuration
    you max experience performance issues. In that case please try to increase
    the heap size of Java / Eclipse. If all else fails please try to restart Eclipse.