IBM Integrated Ontology Development Toolkit

1. What is ontology? How is it compared with ER model and object-oriented model?

An ontology is a model of entity and relationship in a specific domain or universe of discourse (UoD). An ontology can vary from a taxonomy (knowledge with minimal hierarchy or a parent/child structure) to a thesaurus (words and synonyms) to a conceptual model (with more complex knowledge) to a logical theory (with very rich, complex, consistent and meaningful Knowledge). The current focus of the industry is on conceptual model and logical theory (such as OWL). Ontology is distinguishable from other models (such as UML and ER): It has formal semantics; that is, it has a model theory for its language. It supports inference, especially most ontology languages; and it is decidable, and even tractable.

Back to top

2. What is the relationship between IODT and semantic Web standards?

EODM implements all the language constructs of RDFS/OWL. The EODM parser supports RDF/XML syntax input/output and have passed all W3C RDF test cases. The EODM RDFS/OWL inference engines support RDFS entailment and a subset of OWL DL language.

Back to top

3. What is ontology-driven development? What's the connection to model-driven architecture?

The principle of ontology-driven development is that ontologies can be developed by domain experts either from scratch or tailored from other domain models; the developed ontologies can be checked for consistency at design time and can be executed to derive new information providing data at run time. Other domain models can be transformed to ontology.

An ontology itself is a logic model for a specific domain. Specifically, ontology is suitable model of data sources on the Web that target for use by agent on the Web. Ontology-driven architecture can be part of model-driven architecture, but it applies these ideas in a much more extreme way: Domain models are not only used for code generation, but they are used as executable artifacts at run time.

The OMG's recent efforts in defining a mapping between OWL and MOF/UML (Ontology Definition Metamodel) is the cornerstone in bringing the fields and tools closer together. The core model of IODT, the EODM model, is derived from the proposed Ontology Definition Metamodel. A number of transformation engines are, or will be, added into EDOM to transform between OWL and other MDA models.

Back to top

4. What is EODM? Why use EMF?

EODM metamodel, the core metamodel of IODT, is derived from ODM, that is, RDFS/OWL mapped to OMG's Meta-Object Facility (MOF). EMF Ecore model uses a subset of the modeling concepts in MOF 1.4. EMF provides a modeling framework for model transformation and code generation, which is important in making ontology working with other models and tools.

Back to top

5. What's the benefit of using IODT?

Domain experts can develop ontologies and these experts have direct control over the behavior of the executing system and some aspects of the implementation. Feedback is available frequently since the domain models rapidly lead to executable systems. The support of RDFS/OWL directly implies the developed applications will be ready for integration in the semantic Web context.

Back to top

6. I always get a "Transaction Log is full" error when loading a large number of OWL documents into Minerva. What should I do?

Try to tune your DB2 database: Enlarge the LOG_SIZE.

Back to top

6. Why are the inference, real-time, and refresh buttons disabled?(Orient)

In the current release, the inference, real-time, and refresh buttons are enabled only when the ontology editor (that is, the useless-seeming view on the upper right corner of the Orient perspective) is selected as the active view. This problem will probably be solved in future releases.

Back to top

7. I deleted an item in the resource/class/property list, but it appears again when the views are refreshed. Why?(Orient)

The items listed in these views contain both the original items created by the user and the items created from entailment. But only the items created by the user can be deleted purposely. So, if you are trying to delete an item that is created from entailment, the delete operation will actually do nothing, and thus you will see the item again when the view is refreshed.

Back to top

7. Which DBMSes are currently supported by Minerva?

Currently, Minerva supports IBM DB2 and Derby as the back-end store. Note that Minerva on DB2 performs much better than Minerva on Derby because DB2 has some desirable optimizations. Sometimes the former is 10 to 20 times faster than the latter.

Back to top

8. Where can I find more information about IODT?

The documentation and Javadoc for IODT can be found in the doc directory in the released zip file.

Back to top

9. What was new in Version 1.1.1?

Version 1.1.1 contained the following improvements:

• a bug fix (file encoding is now preserved when saving in editors)
• a new function for importing OWL documents
• transaction support for incremental updates in Minerva
• upgrading of all components' version tags to 1.1.1
• updated documentation with a system architecture picture.

Back to top

9. What was new in Version 1.1.1?

Version 1.1.1 contained the following improvements:

• a bug fix (file encoding is now preserved when saving in editors)
• a new function for importing OWL documents
• transaction support for incremental updates in Minerva
• upgrading of all components' version tags to 1.1.1
• updated documentation with a system architecture picture.

Back to top

10. What was new in Version 1.1.2?

Version 1.1.2 contained the following enhancements:

• Bug fix: Transitive property deletion in Minerva""s incremental update mode will redo ABox inference.
• Minerva code namespace has been refactored to com.ibm.iodt.sor for future update.
• addition of Javadoc™ for documentation
• addition of OWL to Ecore and Ecore to OWL transformation example code
• addition of UOB (University Ontology Benchmark) data, which extends the LUMB (Lehigh University Benchmark).

Back to top

11. How do I specify a different RDF/RDFS namespace? (RStar)

In real RDF files, there may be different namespaces for RDF/RDFs, such as rdfs = http://www.w3.org/2000/01/rdf-schema# or rdfs= http://www.w3.org/TR/1999/PR-rdf-schema- 19990303#. In RStar, the default namespaces for RDF and RDFS are, respectively, rdf := http://www.w3.org/1999/02/22-rdf-syntax-ns# and rdfs := http://www.w3.org/2000/01/rdf-schema#. To specify a different namespace, please

• Set the jvm parameter:
  
  -Drdf_namespace=YOUR_RDF_NAMESPACE
  -Drdfs_namespace=YOUR_RDFS_NAMESPACE
  
  
• Use the static methods defined in com.ibm.crl.rstar.rdf.vocabulary.RDF/RDFS:
  
  RDF.setNamespace(YOUR_RDF_NAMESPACE);
  RDFS.setNamespace(YOUR_RDFS_NAMESPACE);
  

Back to top

12. Why did I get an error from the DB2 JDBC driver: encoding not supported? (RStar)

It seems that you are using DB2 Type4 driver in JRE 1.4. Try switching to JRE 1.3 and add some necessary "libs" mentioned in the Readme.txt file.

Back to top

13. How do I enable DB2 Net Search Extender (NSE) support? (RStar)

In RStar, fuzzy queries based on full-text search is supported for more efficient retrieval. However, a DB2 NSE must be installed. To properly use NSE, you should manually build the text index in field Name of tables Resources and Literals.

By default, NSE support is disabled. You can enable it as follows:

Trademarks