We have just released the version 1.5 of TrOWL and the corresponding Protege Plug-in. They are new available for download in our website. This version further improves the performance with cardinality restrictions and the efficiency of ontology updates.
TrOWL v1.4 is now available for download. This version implements an updated ABox realisation algorithm to improve performance. It also offers better efficiency for retrieving super-classes, sub-classes and equivalent-classes. Several minor bugs have also been fixed.
In this year’s OWL Reasoner Evaluation competition, this TrOWL participated the classification and realisation tracks of OWL 2 DL and OWL 2 EL ontologies. It obtained the 2nd place in OWL 2 EL realisation, the 3rd place in OWL 2 DL realisation and 4th places in OWL 2 DL/EL classifications. Congratulations!
TrOWL v1.3 is now available for download. Besides numerous bug fixes, the new version includes the following major updates:
- improved memory consumption by simplifying certain data structures without affecting the performance and recall of reasoning;
- improved performance by optimising the reasoning algorithms, avoiding computing superfluous results;
- separation of approximation and reasoning, and separation of TBox and ABox reasoning. Now TrOWL will not automatically perform reasoning upon creation of reasoner. Instead, users can invoke TBox or ABox reasoning by using the precomputeInferences method of the reasoner. When in TBox reasoning, ABox results will not be computed unless they are relevant to TBox reasoning. When ABox reasoning is performed after TBox reasoning, TBox results will not be re-computed. Users can also directly use the normal get… methods of the reasoner to retrieve reasoning results. TBox or ABox reasoning will be automatically invoked depending on the type of reasoning needed, if they have not been precomputed;
- as a consequence of separation of TBox and ABox reasoning, TBox relevant results now can be computed much faster for ontologies with large ABox;
- improved recall by including more completion rules;
A Protege Plug-in for the new TrOWL v1.3 is also released and available for download.
New versions of TrOWL tools are now available to download. In this update, we have improved the recall of approximation and fixed a few bugs.
We are glad to announce the release of OWL-DBC v1.1, a new set of JAVA APIs that help TrOWL users to connect to the native inference capabilities of Oracle Spatial and Graph. OWL-DBC combines the TrOWL reasoning infrastructure with the scalable Oracle Spatial and Graph, on the one hand, allowing TrOWL to handle extreme large data sets and, on the other hand, enhancing Oracle Spatial and Graph’s support to complex OWL 2 ontologies. You can download OWL-DBC here.
TrOWL v1.1 is now available for download. Besides numerous bug fixes, this version includes the following major updates:
SPARQL is the W3C recommendation for query answering in the Semantic Web. TrOWL v1.0 treats SPARQL variables as distinguished variables that can be evaluated as individuals. TrOWL v1.0 supports the usual TrOWL API and the TrOWL-BGP API. The usual TrOWL API can be used alone in a same way as previous TrOWL releases. To support SPARQL query answering, both TrOWL and TrOWL-BGP are needed.
Stream reasoning is very useful for situations where ontologies are subject to change. The current version supports complete TBox stream reasoning of EL+ and approximate ABox stream reasoning of SHIQ. Reasoning is soundness-guaranteed and tractable as always. Additionally, an approximate justification is also maintained for each inference result. Retrieval of such jsutifications can be performed directly without further computation.