Codd’s computing career began in 1949, when he joined IBM in New York City as a programming mathematician, developing programs for the Selective Sequence Electronic Calculator (IBM’s first electronic—or at least electromechanical—computer, a huge and noisy vacuum tube machine). He also lived for a brief period in Washington DC, where he worked on IBM’s Card Programmed Electronic Calculator. In the early 1950s, he became involved in the design and development of IBM’s 701 computer. (The 701, originally known as the Defense Calculator, was IBM’s first commercially available computer for scientific processing; it was announced in 1952 and formally unveiled in 1953.)

In 1953, Codd left the United States (and IBM) in protest against Senator Joseph McCarthy’s witch-hunting and moved to Ottawa, Canada, where he ran the data processing department for Computing Devices of Canada Limited (which was involved in the development of the Canadian guided missile program). A chance meeting with his old IBM manager led to his return to the U.S. in 1957, when he rejoined IBM. Now based in Poughkeepsie, New York, he worked on the design of STRETCH (i.e., the IBM 7030, which subsequently led to IBM’s 7090 mainframe technology); in particular, he led the team that developed the world’s first multiprogramming system. (“Multiprogramming” refers to the ability of programs that have been developed independently of one another to execute concurrently. The basic idea is that while one program is waiting for some event to occur, such as the completion of a read or write operation, another program can be allowed to make use of the computer’s central processing unit. Multiprogramming is now standard on essentially all computer systems except for the smallest personal computers.) In 1961, on an IBM scholarship, he moved to Ann Arbor, Michigan, where he attended the University of Michigan and obtained an M.Sc. and Ph.D. in communication sciences (1965). His thesis—which was published by Academic Press in 1968 under the title Cellular Automata—represented a continuation and simplification of von Neumann’s work on self-reproducing automata; in it, Codd showed that the 29 states required by von Neumann’s scheme could be reduced to just eight.

In 1968 transferred to the IBM Research Laboratory in San Jose, California. Several database products did indeed exist at that time; however, they were without exception ad hoc, cumbersome, and difficult to use—they could really only be used by people having highly specialized technical skills—and they rested on no solid theoretical foundation. Codd recognized the need for such a foundation and, applying his knowledge of mathematical logic, was able to provide one by creating the invention with which his name will forever be associated: the relational model of data.

The relational model is widely recognized as one of the great technical achievements of the 20th century. It revolutionized the way databases were perceived; indeed, it transformed the entire database field—which previously consisted of little more than a collection of ad hoc products, proposals, and techniques—into a respectable scientific (and academic) discipline. More specifically, it provided a theoretical framework within which a variety of important database problems could be attacked in a scientific manner. As a consequence, it is no exaggeration to say that essentially all databases in use or under development today are based on Codd’s ideas. Whenever anyone uses an ATM machine, or purchases an airline ticket, or uses a credit card, he or she is effectively relying on Codd’s invention.

It should be noted, incidentally, that the relational model was in fact the very first abstract database model to be defined. Thus, Codd not only invented the relational model in particular, he actually invented the data model concept in general.

Throughout this time, Codd continued to be employed by IBM. Perhaps because it was heavily invested in its existing nonrelational database product IMS and was anxious to preserve the revenue from that product, however, IBM itself was initially quite unreceptive to Codd’s relational ideas. As a consequence, other vendors, including Relational Software Inc. (later renamed Oracle Corporation) and Relational Technology Inc. (later renamed Ingres Corporation), were able to steal a march on IBM and bring products to market well before IBM did. Seeing the way the winds were blowing, senior IBM management decided in the late 1970s that IBM should build a relational product of its own. That decision resulted in the announcement of SQL/DS for the VSE environment in 1981 and DB2 for the MVS environment in 1983. 

Codd resigned from IBM in 1984.

Indeed for the remainder of his professional life, he worked tirelessly to encourage vendors to develop fully relational products and to educate users, vendors, and standards organizations regarding the services such a product would provide and why users need such services. He was also interested in the possibility of extending his relational ideas to support complex data analysis, coining the term OLAP (On-Line Analytical Processing) as a convenient label for such activities. At the time of his death, he was investigating the possibility of applying his ideas to the problem of general business automation.

Codd died April 18th, 2003, in Williams Island, Florida. He is survived by his wife Sharon; his first wife Libby; a daughter, Katherine; three sons, Ronald, Frank, and David; and several grandchildren.

Papers

Videos