Riccardo Torlone

The sixteenth edition of the ACM Symposium on Principles of Database Systems (PODS) was held in Tucson, Arizona, May 12-14, 1997. As it has been the case for each of the past six years, the PODS conference was held in conjunction with the ACM International Conference on Management of Data (SIGMOD), to foster close collaboration between the theoretical and applied database research communities. The two conferences had separate program committees and separate proceedings: theoretical papers contributed to PODS and applied ones to SIGMOD. However, there was only one registration process for the joint conference, and attendees received both proceedings and were free to attend sessions in both conferences. Moreover, some of the technical events, the lunches, and the outing were joint events. Overall, there were more than 500 registrants, and attendance in the sessions was always high, even when parallel events was occurring.

PODS'97 program featured one invited talk, three invited tutorials, and 23 contributed papers.

In his invited talk, Avi Silberschatz provided an overview on the support for multimedia data (images, video, audio) provided by next generation database systems, and highlighted the main challenges in devising new algorithms to manage this new type of data. He first discussed the need for multimedia support in database systems and the problem of specifying and evaluating content-based queries. Then, he presented different approaches related to the storage and retrieval of multimedia data.

All three tutorials featured challenging topics in the database research area.

In the first tutorial, Richard Hull discussed fundamental problems raised by semantic heterogeneity in databases, and presented a broad overview of theoretical frameworks that can provide solutions for them. He started by the assumption that "data modeling is an art'', and so "theoretical frameworks can help the humans, but not replace them''. He then examined three main issues: the problem of schema restructuring, with an emphasis on comparing the information capacity of different schemes; the problem of schema merging, illustrating, among others, an approach based on description logic; and the use of different paradigms for integrating data, according to a "closed world assumption'' or an "open world assumption''.

The tutorial of Peter Buneman covered a number of issues related to the management of "semistructured'' data, for which there is not a clear scheme describing them. The most immediate example of semistructured data is the information coming from the Web. He first illustrated a flexible model and a powerful query language for semistructured data, and then addressed, in this framework, the important topics of implementation and optimization.

In the third tutorial, Arie Shoshani compared the research done on statistical databases, with the emerging research area of On-Line Analytical Processing (OLAP) for the analysis of business data. He first pointed out the similarities between the two areas and the correspondence between terminologies. He then identified which research aspects are emphasized in each of these areas and illustrated important results in both of them. He concluded by arguing that the work done in one area can greatly benefit the other.

The 23 contributed papers were selected out of 118 submissions by a program committee chaired by Meral Ozsoyoglu. There were sessions on various topics: database queries, spatial databases, logic and databases, web and graph queries, replicated and continuous data, new databse applications, and query processing and optimization.

A rather significant number of papers considered various logic-based aspects: complexity of database queries (paper by Papadimitriou and Yannakakis), containment of queries (papers by Levy and Suciu, by Ibarra and Su, and by Albert and Ioannidis), query languages over interpreted structures (paper by Benedikt and Libkin), answering queries using views (papers by Beeri, Levy, and Rousset, and by Duschka and Genesereth) queries over paths (paper by Abiteboul and Vianu), queries over the Web (papers by Mendelzon and Milo, and by Atzeni and Mecca), and logic rules for referential actions (paper by Ludascher, May and Lausen).

Finally, the program committee granted two new PODS awards: the Best Paper Award, and the Best Newcomer Paper Award. The PODS'97 Best Paper Award was awarded to Christos Papadimitriou and Mihalis Yannakakis for their paper on the complexity of database queries, in which they revisit the issue of query complexity in the light of the recent parametric refinement of complexity theory. The PODS'97 Best Newcomer Paper Award was awarded to Oliver Duschka and Michael Genesereth for their paper on answering recursive queries using views, in which they extend previous results over conjunctive queries to general recursive queries.

Riccardo Torlone

Dipartimento di Informatica e Automazione

Universita di Roma Tre

Via della Vasca Navale 79

00146 Roma, Italy

E-mail: torlone@inf.uniroma3.it