By Pat Hanrahan 2008/11/01

About ten years ago I was attending a meeting of the National Academy of Sciences with Jim Gray; one of our most brilliant minds in computer science, a Turing Award winner, and a legendary figure in the field of databases. Jim is a warm and open person, so I seized on the opportunity to get his input on some ideas I was exploring on how to improve people's tools for accessing and analyzing data.

First I asked him what he thought were the most widely used ways to interactively access databases. He immediately answered that there were two common ways: form-based interfaces (such as Query-By-Example which is used in Microsoft's Access) and Excel Pivot Tables. He agreed that people need better tools, and we then spent a good hour brainstorming about better ways to interactively analyze and visualize data. That conversation started Chris Stolte and me on the journey that lead to Tableau Software.
A couple of weeks later we had come up with two ideas. The first was that pivot tables could be generalized to produce a wide variety of graphical displays, not just tables of numbers. The second was that in the process of composing the display, the user could also formulate a query. Chris, Diane Tang and I developed a prototype which we called Polaris, and soon thereafter we published a paper on the ideas behind the system.
The original paper was published in 2001. Last week, the paper was republished in the Research Highlights section of the Communications of the ACM.
It is great honor to have our work selected to appear in CACM, but an even greater honor to have our paper recommended by Jim Gray, and to have the perspective on the paper written by Jim (with posthumous help from David Patterson of Berkeley). It is even a greater honor to appear in an issue dedicated to Jim. Many of you probably followed the story of his disappearance, the incredible search that followed, and the recent articles and symposium to honor him. We all loved Jim, and miss his encouragement and leadership every day. We probably would not have developed Tableau, if not for that fortunate meeting many years ago. Jim's influence and memory lives on.