Coordinating Database and Programming Language Research∗

In this essay we examine the gap between database and programming language research and practice. Relational databases and object-oriented programming have been great success stories over the last 40 years. While the database community works hard to improve and extend database capabilities, it does not seem to pay much attention to how databases are actually used. We still use query languages that were designed for adhoc human queries via unique logins, while many queries are now automatically generated by enterprise applications with sophisticated security models. Many in the programming language community view relational databases as a necessary evil that should be papered over, or completely eliminated if possible. Object-oriented databases, orthogonal persistence, or just the file system, all have proponents. Industry experiences a constant churn of APIs and tools, with little guidance on what architectures really work. What is needed is more coordination between database and programming language researchers, to evaluate complete systems with realistic metrics, not just for performance but also for maintainability. This material is based upon work supported by the National Science Foundation under