Jim (James Nicholas) Gray was born January 12, 1944 and lost at sea off the coast of Northern California while sailing January 28, 2007. He was one of world’s most distinguished computer scientists. His numerous contributions to the field of database systems were recognized through memberships in the National Academy of Sciences, the National Academy of Engineering, the American Academy of Arts and Sciences, and the European Academy of Science. He was also a fellow of both ACM and IEEE. Jim was awarded the 1998 ACM A.M. Turing Award for his seminal contributions to our understanding of the concept of transactions and their implementation.
At a tribute event at the University of California, Berkeley, last May 31, 700 of Jim’s friends, family, and colleagues met to discuss both his professional accomplishments and the effect he had on their lives. Speaker after speaker discussed what he did in transaction processing and science applications, as well as the ways he had been a friend, mentor, and research collaborator to all.
Jim’s pioneering research on transactions at IBM in the 1970s is the foundation for today’s world of e-commerce. Every time we use an ATM, reserve a seat on an airplane, or purchase an item on the Web, we are relying on the mechanisms Jim first developed 30-odd years ago. These techniques ensure that the "right" thing always happens—even in the presence of software and hardware failure. While they seem second-nature to us today, when Jim conceived them they required very deep insight into the complexities of concurrently executing queries and updates against a shared database system.
Later in his career, Jim became interested in helping natural scientists with their work. He pioneered putting astronomy observation data into a database system. In this way scientists could query their data in SQL, rather than having to write custom programs in C++ or some other general-purpose language. Implementation of this idea for the Sloan Digital Sky Survey (www.sdss.org/) has resulted in more than 2,000 astronomy publications based on querying this data set through SQL.
Jim received his bachelor’s and Ph.D. degrees from the University of California, Berkeley, in 1966 and 1969, respectively. Soon after receiving his Ph.D. he joined the IBM San Jose Research Laboratory (now known as the IBM Almaden Research Center) where he helped lead the design and development of System R, one of the first database systems to use the relational data model. In 1988, System R (along with INGRES, for the INteractive Graphics REtrieval System, project at Berkeley) was honored with the ACM Software Systems Award for pioneering development of relational database systems. It was as part of the System R project that Jim first developed the notion of what it means for transactions to be "serializable"; that is, they produce the same outcome as the serial ordering of the transactions. He also developed the connection between serializability and database consistency and how a simple protocol known as "two-phase locking" could be used to ensure that two or more transactions are serializable with respect to each other without the user having to understand the semantics of the transactions.
From the time he left IBM in 1980 to his joining Microsoft in 1995, Jim worked for Tandem Computers (19801990) on the parallel relational database system Non-Stop SQL and at Digital Equipment Corporation (19901995). Over the course of his career Jim also made numerous technical contributions beyond his work on transactions, including database system architectures and algorithms, fault tolerance, input/output architectures, parallel database systems, database system performance evaluation and benchmarking, multidimensional data analysis, and e-science, including the TerraServer (www.terraserver-usa.com) and the Sloan Digital Sky Survey project. When he disappeared at sea in 2007, he held the title of Technical Fellow at Microsoft.
That disappearance spurred the computer science community to action, and a massive amateur search effort was pulled together to augment the professional one launched by the U.S. Coast Guard. This effort entailed retargeting satellites to sweep the region of interest and posting the imagery on the Amazon Mechanical Turk site (www.mturk.com) so the distributed community could examine it in parallel to look for his sailboat, Tenacious. Possible sightings were then examined by experts in image recognition. It is the hope of the community that this imagery workflow will be automated and performed in real time during future searches. Parallel efforts searched for wreckage along the entire length of the California coastline and posted flyers at every marina in California. No trace of Jim’s boat was ever found. An extensive underwater search was equally unsuccessful. Hence, it is likely that we will never know what happened to Tenacious, and the loss of Jim Gray will remain a mystery.
I (Michael) first met Jim while I was a struggling assistant professor at Berkeley in 1971. He was instrumental in helping me do the research that led to my first publication, which dealt with a simplification of Jay Forrester’s model of an urban area. I am forever grateful for his help motivating me in the publish-or-perish world of an assistant professor.
Jim was obviously brilliant, as anyone who talked to him quickly realized. However, he also read widely and knew a lot about a lot of things. In fact, he is one of the few people I have found to be intellectually intimidating. Moreover, he was always willing to read papers that other researchers sent him and offer insightful comments. I routinely sent him my work in draft form and was always amazed by the breadth of knowledge reflected in his comments. They usually took the form: "Have you looked at System XYZ?; the people behind it looked at the problem you are considering." XYZ would, of course, be an effort I had never heard of.
Jim was a mentor to many of the younger people in computer science and traveled widely to universities and research centers to interact with researchers. He was always willing to give service to the field. I remember vividly the creation in 2003 of the Conference on Innovative Data Systems Research (CIDR, www.cidrdb.org). We (Michael and David) became frustrated that SIGMOD routinely turned down our practical papers. Reaching the boiling point, we asked Jim to help start a new conference as a venue for such work, and Jim, as always, was willing to help. Moreover, the night before the opening session of the first CIDR conference in Asilomar, CA, we realized we did not have a data projector for showing PowerPoint slides. Rather than risk compromising the success of the conference, Jim made the five-hour round trip back to San Francisco to get a projector, returning to the Asilomar Conference Center at 3 A.M. That was Jim.
He was an unmanageable free spirit in the workplace who could write prodigious amounts of code and even more prodigious research reports.
Anecdotes reflecting his special character are legendary. He refused to conform to social norms; we never saw him wearing a coat and tie. He was an unmanageable free spirit in the workplace who could write prodigious amounts of code and even more prodigious research reports. It was reported at the Tribute that he had asked IBM to transfer him from its Thomas J. Watson Jr. Research Laboratory in Yorktown, NY, to its San Jose Research Laboratory in California to work on System R. When his boss refused, Jim quit on the spot and drove cross-country to be hired by the San Jose Lab. He loved to take people sailing on his boat, and it seems as if half the database community had this pleasure. Equally legendary are anecdotes of his backpacking and hiking trips in the Sierras.
Jim was a true scholar and friend. We will forever try to live up to the standard he set by his behavior. We can speak on behalf of the entire computer science community that we miss this mountain of a man every day. Our hearts and thoughts go out to his wife, Donna, his daughter, Heather, and his sister, Gail, who must deal with the ambiguous loss of Jim up close and personal.
Join the Discussion (0)
Become a Member or Sign In to Post a Comment