UC San Diego Computer Science and Engineering Professor Victor Vianu will present his work at the ACM SIGMOD Principles of Database Systems (PODS) conference, which will be held June 20 to 25, as part of Gems of PODS series. These are similar to prestigious Test-of-Time awards but usually encompass a broader body of work.
This is not the first PODS honor for Vianu, who came to UC San Diego in 1984. In 2010 and 2015, he received ACM PODS Alberto O. Mendelzon Test-of-Time Awards. He also received a Test-of-Time Award in 2019 from ICDT, another major database conference.
At the upcoming PODS, Vianu’s talk, Datalog Unchained, will describe his work on database query languages inspired by logic programming, which started in 1988 and spanned almost a decade.
“Databases rely greatly on logic,” he said. “I've been working quite a bit on the theory of query languages and more recently on automatic verification of data-driven applications, which rely on databases.”
Based on the Prolog language, Datalog is actually a family of programming languages that Vianu and others have been refining over the years. In the process, Datalog’s use has expanded beyond databases, finding applications in areas as diverse as security and privacy protocols, program analysis, natural language processing, probabilistic inference, modular robotics, multiplayer games, telecom diagnosis, networking and distributed systems.
Vianu found that classical approaches to Datalog semantics, such as minimum model semantics, had disadvantages that limited the language’s usability and expressive power. A different approach, called forward chaining semantics, has proven more advantageous.
“We explored this in a sequence of six or seven papers starting in 1988,” said Vianu. “We showed that forward chaining semantics could overcome the expressiveness limitations of the previous approaches and could incorporate new, useful features, such as non-determinism and updates, which were not accommodated in the original design of the language.”