AgentIT SunWorld Network World LinuxWorld JavaWorld InfoWorld Computerworld CIO ITworld.com
Computerworld logo


news resources careers CW, Inc.


1959: The Creation of Cobol

By Mary Brandel
03/15/99 ----------------------------------- You know that little year 2000 problem? Well, it all began 40 years ago. On May 28, 1959, the Conference of Data Systems Languages (Codasyl) met for the first time, with the idea of developing a universal language for building business applications. That language was Cobol, short for "common business-oriented language." And it's Cobol's dramatic success that's at the heart of the millennium bug.

Influenced by Fortran, a programming language for the scientific community, and FlowMatic, an English-language compiler for business data processing built by Grace Hopper, the group recognized the growing needs of the business community.

"We thought, If the scientific programmers are going to get a single language, we could do the same for business," says Bob Bemer, who at the time was completing work on Fortran at IBM. With Hopper, Bemer served as an adviser to Codasyl. He is responsible for coining the term Cobol.

By April 1959, that undercurrent swelled into action. At an informal meeting at the University of Pennsylvania in Philadelphia, a small group of computer manufacturers, large users and academics asked the Department of Defense (DOD) to head the effort.

The next month, the DOD called the first meeting of Codasyl, which consisted of eight computer manufacturers and a few large users. The DOD broke Codasyl into several committees, and by June, the nine-member "short-range committee" was asked to undertake a six-month investigation into developing the language.

"We worked almost full time doing the language specification, even though we were all employed by different employers," says Howard Bromberg, who was a Codasyl member and an employee at RCA Corp.

In addition to machine-independence, one of the most important requirements of the language was simplicity. The committee wanted the language to be readable by laypeople, which led to the idea of using English.

But just because Cobol was designed to be easy to learn doesn't mean it was easy to build.

"In business, there are no scientific laws and no algebra, but there are different laws for the 50 states, different fiscal years and different reporting requirements," Bemer says.

In addition, computer manufacturers were trying to develop their commercial Cobol compilers while Cobol's specifications were being defined. All decisions had to be approved by Charlie Phillips, the DOD representative who directed Codasyl.

"I used to get frustrated — I had a group of people sitting there trying to build a Cobol compiler," Bromberg says.

That led to the famous "tombstone incident." Bromberg sent a granite tombstone to Phillips with the word Cobol inscribed on it. He figured it would get his point across about the fate of Cobol if things continued to move so slowly.

A complete specification was finished in just six months. That was in December 1959. By the following year, Cobol was commercially ready, and for the next 20 years, more programs were written in Cobol than any other language.

Unfortunately, it was the resulting tidal wave of Cobol programming that now has us anxiously checking our watches as they tick away toward 2000.

Although the Cobol creators played their part in the problem — specifying two-digit year fields for capturing and manipulating system dates — the blame falls just as squarely on the programmers, who could have used four-digit year fields, says Jerome Garfunkle, a year 2000 consultant who served on the American National Standards Institute's Cobol Committee for 20 years.

In 1974, Cobol officially changed to four-digit date fields, but that change obviously didn't catch on right away.

Brandel is a frequent contributor to Computerworld. Contact her at brandel@cwix.com.


Copyright © 2000 Computerworld, Inc. All rights reserved. Legal notices and trademark attributions