Who is Edgar F. Codd – Creator of the Relational Database Management System?
Edgar Codd is the creator of the relational databases model (RDBMS), an extremely influential general theory of data management, the foundation of RDBMS, used everywhere nowadays.
Early Life and Education
Edgar (Ted) Frank Codd was born on 23 August, 1923, in Fortuneswell, on the Isle of Portland, in England. He was the youngest of seven children of a leather manufacturer father and a schoolteacher mother. Edgar earned degrees in mathematics and chemistry at Exeter College, Oxford, before serving as a pilot in the Royal Air Force during World War II.
Career – What Did Edgar Codd Do?
Edgar Codd’s career has spanned many decades and can be broken up as follows.
The 1940s-1950s: The Move to the United States
In 1948, Codd moved to New York to work for IBM as a programmer for the Selective Sequence Electronic Calculator, IBM’s first electronic computer, an experimental machine with 12500 vacuum tubes. He then invented a novel “multiprogramming” method for the pioneering IBM 7040 STRETCH computer.
This method enabled STRETCH, the forerunner to modern mainframe computers, to run several programs at the same time. In 1953, disappointed by the USA policy, Codd moved to Ottawa, Canada. A decade later he returned to the USA and received his doctorate in computer science from the University of Michigan. Two years later he moved to San Jose, California, to work at IBM’s San Jose Research Laboratory.
The Late 1950s: The Start of Research
Codd made his mark with research contributions in programming languages and databases systems during the late 50’s/early 60’s. His application for US citizenship was postponed because he was working for an industrial contractor (IBM). However, it gave him time to publish several papers between 1955 and 1960 that laid out some principles about database design theory. These were present for decades to come both as inspiration and as a basis for many practical systems.
The Late 1950s -The Early 1960s: Proposals and Work on the OLTP System
He had submitted proposals for what was called “Procedure Oriented Languages” at IBM in 1957, which later received funding in 1960. The paper outlined a program written in PL/I, with several examples explaining how it could be used to solve business problems – such as payroll information systems or inventory control software.
This research was eventually published under the title “A Business Oriented Sublanguage Of PL/I.” This early development work led to the Business Data Processing Systems (BDP) project that Codd initiated while working at an IBM branch office in San Jose, California.
In April 1961 Codd and his team of engineers completed what is generally considered to be the first complete commercial implementation of the relational model and which contained all four major components (a Relational Data Sub-Language (RDSL), an interface language, a relational algebra manipulator subprograms and a terminal user interface).
This work was published in May 1962 by IBM as the San Jose Research Laboratory Technical Memorandum RM3420 – “RDBMS”. However it wasn’t marketed because IBM felt that it would compete with their existing data systems.
Edgar Codd’s Biggest Invention: The Relational Database Management System (RDBMS)
The first RDBMS, code-named “Elmer”, is generally considered to be the forerunner of IBM’s IMS (Information Management System) and was based on what Codd had written in his thesis. The new design produced several important features that would be incorporated into future database systems, including: a two-level storage system, a query processing method and an automatic index-building mechanism.
The key to the RDBMS was its design (which included the selection of an appropriate algorithm for query optimization) – in particular, it used hierarchical indexing so that only parts of records were fetched from disk when they were needed. This leveraged hardware resources at a time when memory was expensive and slow.
Codd, Frank W. Snyder and John G. Myers published a seminal paper outlining the concepts of “Second Generation Relational Systems”, systems designed to overcome the limitations of first generation relational systems.
By mid-1962 Codd had finalized his ideas for a network DBMS he called RM/360 (Relation Management System) before moving on to an IBM branch office in Bethesda Maryland where he initiated the development of the RDSL language implementation, while continuing research into how database query languages could be implemented more efficiently than was then available for other system architectures such as hierarchical and network models.
This work led him to postulate that it was possible to create an optimizer that would work out which operations were needed to retrieve the database’s information.
In 1963 Codd and Myers co-authored a paper outlining some of the theoretical work that was required to make an efficient query by example system possible. This paper also introduced the use of “delta keys” to represent the discriminator columns between similar tuples in any table, i.e. those columns which had different values for every row in one distinct table, depending on its specific key values – such as customer identification numbers or next transaction date/number fields in a sales order entry table.
The Late 1960s – Early 1970s: The DC System
During this time IBM began focusing on how systems could be networked together and how databases could be shared through these networks using their Remote Spooling Plan (RSP). This led to the development of IBM System/360 Mainframe Network Architecture System (SNA) in 1969.
Early in 1973, two years after he left IBM, Codd published his most influential paper titled “A Relational Model of Data for Large Shared Data Banks” which was accepted by Communications of the ACM in March 1973.
The Late 1970s and 1980s
In the late 1970s there was a lot of interest in computers that were not controlled by IBM and hence could be used to run alternative operating systems. This led to the development of Open Database Connectivity (ODBC) which uses an SQL driver interface for application programs connecting to relational databases.
In 1978, at a time when many computer vendors had implemented their own versions of SQL, Codd started work on what would become known as “version 3” or “SQL-3”. In this period he also wrote out all his thoughts regarding how future database implementations might evolve into what he called “The Fifth Generation Computer System” – a fully object oriented system with abstract data types forming its base fundamentals. The papers outlining these ideas are freely available for download and viewing.
In 1981 IBM released its first relational database product, SQL/DS. DB2, initially for large mainframe machines, was announced in 1983. IBM’s DB2 family of databases proved to be one of IBMs most successful software products and is incorporated in the operating systems of mainframe and middleware servers of IBM.
Still in IBM, Codd continued to develop and extend his relational model. As the relational model started to become fashionable in the early 1980s, Codd fought a sometimes bitter campaign to prevent the term being misused by database vendors who had merely added a relational veneer to older technology. As part of this campaign, he published his famous 12 rules to define what constituted a relational database.
In 1981 Codd was awarded the ACM Turing Award for his pioneering contribution to the relational model for database management and in 1990 he received a Golden Anniversary Award from IBM for “Outstanding Technical Achievement”.
Codd retired from IBM in 1984 at the age of 61, after a serious injury resulting from a fall.
Later he joined up with the British database guru Chris Date, whom Codd had introduced to San Jose in 1971, to form the Codd and Date Consulting Group. The company, which included Codd’s second wife Sharon Weinberg, made a good living from conducting seminars, writing books and advising major database vendors. Codd never became rich like the entrepreneurs like Larry Ellison, who exploited his ideas. He remained active as a consultant until 1999.
The 1990s and Beyond
In the 1990s, Codd made a number of appearances in popular press and scientific journals noting that object-oriented database systems, while very likely to have advantages in certain situations, lacked the persistence properties required for most business databases. He stated that “the relational model is an unbroken chain of reasoning from design concepts through application and on into future evolution” (Codd in Jim Melton’s column) and insisted on its suitability as a foundation for business DBMS systems.
In 2000 he was given the Computer History Museum’s Distinguished Accomplishment Award along with Larry Ellison and Ed Oates for his pioneering work towards developing the SQL language. In 2003 he received two more awards: one from The Federation of Enterprise Architecture Professional Organizations (FEAPO) for his work with the relational model and another from Time magazine as a member of their “Time 100 Scientists & Thinkers” for the same.
Edgar Codd: Marriage, Divorce, Death & Legacy
Edgar Frank Codd died in 2003, aged 79, at his home in Oceanside, California having had cancer for some time before he passed away. His contribution to what is now known as the relational model has cemented itself into the foundation of data processing systems across a variety of platforms and industries. He was also responsible for inspiring many thousands of people to become involved within this field by virtue of writing papers, giving lectures and taking part in public debates regarding these topics – both as an IBM employee and then after leaving them to teach computer science at the University of California.
His theories have formed the basis of database systems and business intelligence throughout much of the computer’s history and there is little doubt that he will be remembered as one of its greatest pioneers.
Edgar Codd: Awards and Achievements
Aside from this highly prestigious Turing Award, Codd received the following:
- In 2000, Edgar Codd received the Computer History Museum’s Distinguished Accomplishment Award (shared with Larry Ellison and Ed Oates).
- In 2002 Edgar received a Special Achievement Award from OOPSLA for contributions to database technology.
- In 2003, Codd was awarded the title of Fellow (for Contributions to Database Technology) by the Association for Computing Machinery and was one of the recipients of The Federation of Enterprise Architecture Professional Organizations (FEAPO) highest award, “The Silver Medal”, that year.
Published Works and Books
Codd, one of the most prolific authors in Computer Science publishing hundreds of papers over his lifetime, also managed to find time to author a number of books and booklets. Some of these included:
- The Relational Model for Database Management: Extensions for Integrity, Availability and Services
- The Relational Model: A Retrospective Review – IBM Systems Journal, Volume 18 Issue 1 January 1979
- How to Capture the Benefits of Object Oriented Programming Without its Costs (Presentation) – International Conference on Very Large Databases, Zurich, September 16–19, 1988
- Object/Relational Technology – Efficient Relational Database Management for the RDBMS Applications Server (Brochure) – IBM Corporation, January 1993
- Extending SQL: A Tutorial and Survey – VLDB Conference Proceedings Volume 12 Issue 2 August 1992
Papers and Publications
Edgar was also responsible for a vast number of papers and presentations during his time at IBM. Some of the most notable ones can be found below:
- Transmitting Data in Fixed Length Records Using Six Bit Characters – E. F. Codd, R. T. King, April 1965
- The Relational Calculus – E. F. Codd, August 1970
- Database Sublanguage: A Survey – E.F. Codd, JACM September 1972
- The relational model of data for large shared data banks – Communications of the ACM, V 15 N 6 June 1972 
- Query Languages for Managing Biomedical Information Systems – International Journal of Man-Machine Studies, Volume 10 Issue 1 (January 1977)