- From: Sean Bechhofer <seanb@cs.man.ac.uk>
- Date: Thu, 27 Nov 2003 10:43:55 +0000
- To: public-webont-comments@w3.org
The WonderWeb OWL API [1] is a collection of interfaces and classes intended to represent OWL ontologies. The structure of the data model closely follows that presented in the OWL Abstract Syntax. In addition to constructing an API, we built default implementations of the interfaces, an RDF parser, and species validator along with translators to other formats. In particular, a translation to FOL allowed us to use a FO prover for OWL reasoning (system known as Hoolet [2]). Handling OWL in RDF/XML was by far the hardest part of the task. I would estimate that the design and implementation of the codebase took around four months, with the bulk of that effort (>75%?) qbeing devoted to RDF parsing and validation. Throughout the implementation process I found the OWL Test Cases an invaluable resource. It is unlikely that I would have been able to implement species validation successfully without them. I did not find the OWL documents difficult to read and understand, although I had the advantage of prior experience in Description Logics, building tools to support OIL and DAML+OIL, and was personally involved in the Working Group. Some of the trickier aspects of validation were sometimes harder to find, but they *are* there in the documentation. Within the implementation we used the RDF parser from the KAON toolkit provided by the University of Karlsruhe. [1] http://sourceforge.net/projects/owlapi [2] http://wonderweb.semanticweb.org/owl/first-order.shtml -- Sean Bechhofer seanb@cs.man.ac.uk http://www.cs.man.ac.uk/~seanb
Received on Thursday, 27 November 2003 06:07:37 UTC