SkylineCompare is a Test Framework for Skyline-Algorithms. The Framework is expandable, several Test-Cases and Data-Generators are included.

No Skyline-Algorithms are included in this version. At least one Skyline-Algorithm must be integrated into the Framework first. Read the "New Algorithms" paragraph for further details.

The Start.java class in the starters package includes some basic tests. The unmodified version of this class runs two test-iterations on two algorithms (PSkyline and ParallelBBS). Two InputHandler are used here: light-anticorrelation and light-anticorrelation-gaussian. More InputHandler could be added. The values used by the tests could also be changed: The class-variables of the DifferentDimensionsBenchmark and DifferentDatasizeBenchmark and their respective super-classes could be overwritten.

The gradle run command runs the main method of the Start.java class.

Other test-examples are included in the StartExamples.java class in the starters package.

New TestCases must implement the abstract TestCase class or one of its subclasses. The run() and genOutput() methods could be overwritten. Run-methods should use the run-methods of their super-classes.

New algorithms must implement the abstract SkylineAlgorithm class. Each algorithm iteration needs an instance of the class TimeObject to save their runtimes. Specific checkpoints must be added to the algorithms by using the TimeObject methods. At least the following methods must be used within the new algorithm, to achieve functionality with this framework:

• setStartInput()
• setFinishInput()
• setStartCalc()
• setFinishCalc()