ImplementingOWLinSQL

• Main Authors: Chen, Yung-Cheng, 陳永錚
• Other Authors: Daniel. J. Buehrer
• Format: Others
• Language: en_US
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/87415047781366479778

碩士 === 國立中正大學 === 資訊工程研究所 === 91 === This paper is part of the project on the Cadabia database system, which is based on the class algebra. Cadabia, an inter-database system, and it can access data in traditional databases and add in an object-oriented IS-A hierarchy of classes along with traditional E/R binary relationships. Since Cadabia can get the data from relational databases, we translate OWL into SQL commands which can store the ontologies of OWL. Thus, Cadabia can seize the OWL ontologies defined by others on the internet to describe Cadabia’s backend data. The sematics of OWL is usually in the syntax of RDF/XML. We use XSLT to transform OWL into SQL commands which create tables to store the ontology information. Therefore, the DBServer at the backend of Cadabia can access these translated data in databases by using JDBC. OWL just facilitates machine-understanding of XML documents. It needs an additional reasoner to infer further meaning. After being translated in SQL tables, which are accessible to Cadabia, class algebra can be used to make inferences about the meta-data. The queries in Cadabia contain implicit relations which involve certain automatic inferences. Thus, deductions can also be used to prevent users from entering invalid data. The constraints of class definitions can involve complex predicates such as making sure that a relation remains symmetric or transitive. Since these intents are checked after each transaction, it is impossible, for instance, for the user to create a non-symmetric instance of such a symmetric relation.