|
|
|
The main contribution of this cited paper is that it explores the use of relational system to process XML documents without any manual intervention. It then identifies the causes for the limitations of using relational database engines for processing XML documents and propose certain extensions to the relational model that would make it more appropriate for processing XML documents. The conversion techniques are evaluated by using real DTDs.
This paper assumes that all XML data has DTD. In reality, not all XML documents have DTDs and not all XML documents conform to their DTDs. The evaluation of various schemes in conversion techniques (Basic, Shared, and Hybrid) are based on the average number of SQL joins required to process path expressions of a certain length N. If the response time is also one of the evaluation metrics, then it is easier for the readers to get the sense of how well these techniques can process XML documents.
The testing queries in the evaluation process are not presented, so we do not really know whether the testing queries are the types of queries normally used in real world situations.The authors of the paper categorized 37 DTDs that they made an experiment into four groups depending on the comparison performance between Hybrid and Shared techniques. Nevertheless, what we could get more out of this categorization is to make the characteristics of the DTDs that different types of techniques work well with. We then may be able to apply an appropriate technique to an XML document according to its DTD.
Copyright © 2000 by the author(s). Review published with permission.