This paper is the first paper that I have read that presents and evaluates techniques for mapping XML documents with DTDs to relational databases for storage. It also presents and evaluates corresponding XML-to-SQL query translations. The paper closes with recommendations for relational DBMS improvements to make them better-suited for solving the XML problem.
My reaction on reading this paper was that it is a well-motivated and clearly explained piece of work on a timely and important problem -- namely, how to go about exploiting commercial DBMS technology to manage XML data (and improving commercial DBMS systems to support this better in the future). The paper has a nice discussion of the problem landscape as well as of the most relevant related work. The authors do a nice job of presenting and evaluating various algorithms for schema and query mapping, though the evaluation metrics are a bit simplistic. Overall, this was the best (and most exciting) paper that ended up in my to-review batch for the 1999 VLDB Conference. I hereby nominate this paper for the VLDB-2009 'ten-year paper' award. :-)
[FULL DISCLOSURE: I was a faculty member at the University of Wisconsin, where this work was done, for a number of years. However, this work was done after my time there, and I do not feel biased by that relationship with the Wisconsin authors.]
Copyright © 2000 by the author(s). Review published with permission.