A half-century of COBOL

Fifty years after its initial specifications were laid out, COBOL continues to be the language that would not die. While new COBOL projects are nearly extinct, its resilience as a maintainable, simple, English-based language has kept COBOL programs going strong, long after their initial development has finished.
##CONTINUE##
For as long as there has been COBOL, there have been people saying that it was doomed. When Jan Stuart first learned the language in 1978, she was told, even then, that it was a dead language. Now retired in the UK, Stuart found herself working with the language right up until her retirement a few years ago.

“The COBOL programs are the ones that process large amounts of transactions, typically the overnight batch,” said Stuart, who wrote COBOL at a number of financial institutions. “And its usefulness is that it can process large numbers of transactions quickly in a way a lot of new languages struggle with. I think that's why a lot of financial organizations keep their old COBOL programs.”

Stuart said that one of the biggest benefits of using COBOL is its flexibility, which comes from its simplicity. “In COBOL, you can very easily map different data items in different formats. That's very useful. In the early part of the program, where you specify your variables, you can have overlapping formats so you can easily convert alpha to numeric without having to do any moves.”

She enjoyed working with COBOL for 31 years, because “it's very easy to learn," she said. "I suppose the clever part of it is the way COBOL compiles into assembler code. Some people might find that quite difficult. But it breaks right down into machine code."

Premature burial
Indeed, even before COBOL was complete, some were calling it a dead language, so much so that in the early 1960s, one angry developer bought the language a tombstone.

Howard Bromberg worked with Hopper in the early days of the language. When he began working at RCA on the implementation of COBOL, he found that the work was wildly difficult to coordinate with the central COBOL specification committee. As RCA's representative on that committee, Bromberg found himself and his team rushing to create a commercialized version of the language, often skipping ahead of the specification committee. He felt that this was an untenable position, compounded by RCA's business desires and the COBOL committee's lack of speed.

So one day, while driving home, Bromberg found himself passing a monument company by the freeway. He went in and commissioned a tombstone for COBOL. After a few weeks of harassment from his neighbors, he crated the tombstone up and mailed it to Charlie Phillips, one of the original architects of COBOL, who was then working at the Pentagon.

Phillips later asked why Bromberg had done this. Bromberg never admitted his motivations and, indeed, never publicly admitted his involvement with the affair until the 25th anniversary event.

More recently, in 2008, the state of California went through a rather rigorous round of layoffs to save money. This included the forced retirement and laying-off of all government-employed COBOL programmers. Unfortunately, when a later decision was made to cut state employee pay across the board, the state found that it would have to modify its check payment applications. Naturally, those applications were all written in COBOL.

Youth served by COBOL
Reports of the death of COBOL have been greatly exaggerated. Even today, companies like Fujitsu and Micro Focus offer IDEs for writing COBOL code. Fujitsu has actually done a great deal of work with COBOL, and it has brought it into both Eclipse and Visual Studio. It also has ported the language into HP/UX, Linux, Solaris and Windows.

Mickey Rosen, a sales representative with legacy modernization company Alchemy Solutions, has been writing COBOL since 1968. He said that the real reason for the language's longevity is its readability.

“I used to manage COBOL development for a number of companies," Rosen said. "I always found the most expensive thing was not creating and getting it running, it was the ongoing maintenance and support of those applications. When you compare the ease of maintenance and understandability to other languages, that was a big selling point for COBOL."

Stuart wound up being quite useful to the companies she worked for because she was often one of only two or three developers in those organizations that could comprehend the legacy COBOL code they were tasked with bringing into a new language. But she predicts that, despite her retirement, COBOL will still have a long life ahead of it.

“I'm in my sixties now. My generation grew up with COBOL," she said. "Then along came another generation of programmers who did stuff in Visual Basic and C and C++, and that generation thought of the mainframe and COBOL as old-fashioned. I've got colleagues in that age bracket who almost shake when you say you want to log into the mainframe.

"I have colleagues in their twenties, the generation that's grown up with laptops had a PC in their bedroom. Nothing holds fear for them. They want to learn COBOL. They recognize it as a useful language. They recognize it as the language that runs the engines of many businesses. It's frustrating because there are fewer and fewer experts who can teach them. What I have seen is the young ones coming up want to learn it."

The COBOL's English
In 1959, legendary programming pioneer Grace Hopper decided it was time to push programming further. She had previously worked on Flow-Matic, thought by many to be the first English-language-based programming language. By 1958, Flow-Matic was in general use, but Hopper felt that there was a need for an even more accessible method of programming.

“When I started, I just went ahead with the idea,” said Hopper at the 25th anniversary of COBOL celebration in 1985. “I have later learned that it is much easier to apologize than to get permission.

"In the case of Flow-Matic, we discovered that a lot of people hated symbols, even though the mathematicians and engineers loved them. These people used words. We proposed that we should write programs in English statements, providing a compiler that would translate to machine code. I was told that this couldn't happen because computers don't understand words. I said that they didn't have to; they just had to compare bit patterns. 'Add' has just as many bit patterns as a plus sign does. But I was getting nowhere. So we acted on the motto: 'Just go ahead and do it.' The lesson that we learned from COBOL is that you must go ahead and do it and make it work, and then get out and sell it.”

Thus Hopper laid out the first specification for the language that would come to be known as the Common Business-Oriented Language. Over the next 30 years, COBOL would be the workhorse of choice for financial institutions, military operations and anyone who ran a mainframe, cementing its long life.

-----------------------------
BY Alex Handy
Source：SDTimes

Copyright © 1999-2009 BZ Media LLC, all rights reserved.