What is this?
Research in the field of Semantic Web technologies and development of reasoning systems requires rigorous evaluation methodologies, benchmarks and software frameworks. This is particularly crucial when designing new systems, or repurposing existing ones, for ubiquitous computing scenarios. The deployment of Semantic Web reasoners on mobile and embedded devices introduces concerns regarding the energy usage of such systems, which was pretty much a non-issue for desktop-oriented ones.
evOWLuator easens the assessment of inference tasks implemented by OWL reasoners, while being sufficiently flexible to support running reasoning tasks on remote devices, with Android and iOS supported out-of-the-box. Its plugin architecture makes it very extensible and configurable, especially with regards to integrating reasoning systems, supporing additional target platforms and interfacing with energy profilers and sensors.
evOWLuator currently supports the assessment of the ontology classification and consistency standard reasoning tasks, as well as the matchmaking non-standard reasoning task. Inference services can be tested under different evaluation modes:
Correctness: verifies that results of an inference task match those of a reference reasoner.
Performance: measures turnaround time and peak memory usage.
Energy footprint: estimates the energy usage of the reasoner.
The system is also able to generate summaries and graphical visualizations, providing human-understandable representations of the raw data collected during previous evaluations.
evOWLuator is distributed under the Eclipse Public License, Version 2.0.