Designing Good Semi-structured Databases

Semi-structured data has become prevalent with the growth of the Internet and other on-line information repositories. Many organizational databases are presented on the web as semi-structured data. Designing a \good" semi-structured database is increasingly crucial to prevent data redundancy, inconsistency and updating anomalies. In this paper, we deene a semi-structured schema graph and identify the various anomalies that may occur in the graph. A normal form for semi-structured schema graph, S3-NF, is proposed. We present two approaches to design S3-NF database, namely, restructuring by decomposition and the ER approach. The rst approach consists of a set of rules to decompose a semi-structured schema graph into S3-NF. The second approach uses the ER model to remove anomalies at the semantic level.