Author

Shawn Bowers

Date

January 2004

Document Type

Dissertation

Degree Name

Ph.D.

Department

Dept. of Computer Science and Engineering

Institution

Oregon Health & Science University

Abstract

Information management systems (such as database and knowledge-based systems) are based on data models, which provide basic structures for organizing and storing data. A number of data models are in use, and each provides a slightly different set of structures. For instance, information is represented as tables in the relational data model, as ordered trees in semi-structured data models (such as XML), and as directed graphs in semantic networks (such as RDF). With several distinct data models available, developers can select the most convenient representation for their particular need. However, the use of multiple data models introduces structural heterogeneity, making it difficult to combine information and exploit generic tools (for example, for querying or browsing). This dissertation describes a new framework called the Uni-Level Description (ULD) that can accurately store and accommodate information from a broad range of data models. The ULD consists of a meta-data-model to describe the basic data structures used by a data model, and a uniform representation to store information within a source, including the data-model structures, schema (if present), and instance data. The ULD extends existing meta-data-models by allowing uniform access to schema and data, by permitting data models with non-traditional schema arrangements, and by providing a language for representing data-model constraints. The two primary motivations for our work are (1) to study the basic structural capabilities offered by data models and (2) to define languages for enabling interoperability among sources with structural heterogeneity. We use the ULD to describe a wide range of data models, including those that allow optional and partial schema. We present a query and transformation language (based on Datalog) for accessing and converting information among heterogeneous sources. We demonstrate the flexibility of the transformation language by defining a number of structural mappings. And finally, we use the ULD as the basis for generic navigation, where a single set of operators can be used to discover and browse information in structurally heterogeneous sources.

Identifier

doi:10.6083/M4JW8BS7

School

OGI School of Science and Engineering

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.