A model and queries for databases managing structured documents with object links

Abstract

In this paper the authors propose a conceptual model and database internal representation for managing structured documents using databases. They describe physical storage methods for how to break down and manage documents and management methods using types other than character strings in a subdocument in an internal representation of a database of a structured document. In addition, the authors propose a database model for structured documents with links at the object level between text strings in the document and the semantic data (referred to herein as annotation data) in the database expressed by text strings as well as queries to such databases. The conceptual model for text with links at the object level is composed of two layers, the notation layer which holds the notation character strings and the reference layer which stores the semantic data that represent the partial text strings in the notation layer. This conceptual model introduces into a database a new abstract data type with an operational system that has practical descriptive capacity and a simple structure which is simple and allows for minimal expansion with respect to existing text models. The model also manages linked text. As a result, queries based not only character strings from documents but also on meaning are possible for a database of structured documents which manage the conceptual model proposed by the authors. © 2000 Scripta Technica, Syst Comp Jpn, 31(6): 29–44, 2000