What ' s Next ? Sequence

A large (and increasing) number of applications require the ability to manipulate sequences of data. While some of these applications use very specialized forms of sequences|e.g., video sequences|there is much to be gained by developing a general framework for managing and querying large sequence data sets. Important target domains include scientiic modelling, stock market and nancial analysis, analysis of trace data, temporal queries, and exploratory data analysis or`data mining'. Current database systems cannot address the problems in these domains satisfactorily, and developing suitable systems and techniques is an important and interesting research problem. In this paper, our goal is primarily to motivate the importance of sequence query processing, and to discuss why current systems are unsatisfactory. We illustrate some of the technical challenges in developing suitable systems through examples. We also outline some of the work that we are doing in this area in the SEQ project at Wisconsin. This includes a model of sequence data and a framework for the optimization of sequence queries based on several novel techniques, and a framework for visual querying of sequence data. These ideas are currently being implemented in the SEQ and DEVISE systems.