Investigation of the use of the object-oriented paradigm in the construction of a triple store based on dynamic hashing.

摘要

Any set of knowledge, no matter how complex, can be represented as a collection of triples, each of which consists of a subject identifier, relationship name, and object identifier. In order to support applications in natural language processing, it is advantageous if a triple can be retrieved using any field or a combination of fields as key. An appropriate data structure to support this type of access consists of multiple copies of the collection of triples, each of which is structured as a dynamic hash table. Such a data structure has a number of features which could cause problems in implementation. In particular, the multiple copies of the data could lead to inconsistencies on update. Also there is scope for code re-use owing to the fact that many of the operations on the multiple copies of the data are similar. The thesis investigated in this dissertation is that the object-oriented paradigm is well suited to the construction of a triple store based on dynamic hashing. The object-oriented paradigm provides a way to structure and manage complex relationships among a large number of system components. The object-oriented features or class, inheritance, and encapsulation facilitates construction, debugging, and maintenance of code in the triple-store system where there is a potential for reuse. In order to determine if these features are advantageous in this application, a triple-store was designed and implemented in the object-oriented language Eiffel and the design and code were analyzed and conclusions drawn. Paper copy at Leddy Library: Theses u26 Major Papers - Basement, West Bldg. / Call Number: Thesis1994 .S540. Source: Masters Abstracts International, Volume: 34-02, page: 0794. Adviser: Richard Frost. Thesis (M.Sc.)--University of Windsor (Canada), 1995.