I know from experience that the best way to hear from readers is to write something wrong. And as you can guess, I hear from a lot of readers a lot of the time. My most recent gaffe was in last Wednesday's Dr. Dobb's Report, where I reported that there's an "estimated 200 million lines of COBOL code in use." What I meant to say, of course, is that there's an "estimated 200 billion lines of COBOL code in use." A million here, a million there; pretty soon we're talking real programming.

Thank goodness readers are smarter than editors. It took Gilbert Herrera all of about 30 seconds to remind me that "There are NOT 200 Million lines of COBOL code in the world. Try 20 Billion--or even 100 billion." He went to say that "I worked on a small system--1995-2006--that had about one-and-half million lines of COBOL code. The other (much larger) systems in my shop probably had a combined total of 30-40 million LOC."

David Black didn't take me to task over LOC, but he did share some interesting facts about COBOL in this day and age. "I have been involved with COBOL since its first days. COBOL is still the best language for business applications. It is available for all brands of hardware and all operating systems from mainframes to Linux to Windows and any others you may think of. There are many commercial versions of COBOL and an Open Source version named 'Open COBOL'.

"May 28-29 1959 was the creation date of COBOL at the DOD. Next year it will be 50 years old. 200 billion lines of COBOL are in use today [see he knew] (65 percent of the total software). This represents a $2 trillion dollar total investment (Gartner Group). Five billion new lines of COBOL are being added each year (IBM). 85 percent of all computer programs are written in COBOL. Your life is influenced around the clock by COBOL applications. They vary from banks, stock markets, grocery store businesses, hospital administration, steel manufacturing, automotive production, insurance companies, ...

"The language continues to be updated. Many new features have been added including Object Orientated features including polymorphism, overloading. COBOL can call other languages such as C and Java. The COBOL Task Group is PL22.4 of the International Committee for Information Technology Standards (incits). They meet several times each year. Their next meeting is July 21 thru July 25 in Ontario, CA."

Scott McMaham sees COBOL as a "glue" language: "If you want to understand modern COBOL, it's a glue language. IBM has a 'stack' of technologies for data processing like CICS, DB2, IMS, VSAM, ISPF, etc which are glued together using COBOL because it's the best fit and well supported (DB2 generates COBOL record layouts, etc.). So modern COBOL programmers are specialists in the IBM stack of technologies, not just vanilla COBOL programmers. MicroFocus has never gotten much traction with their non-mainframe product because porting code that uses CICS, etc is not possible because there are no non-mainframe equivalents. It would be easier to write a JSP front-end to a CICS application through WebSphere than try to rewrite a COBOL program that uses CICS transactions. Any non-mainframe COBOL product is in this sort of position where it's tough to port. The idea of compiling to the JVM is a good one, but the sorts of COBOL programs that would work in this environment are extremely limited. MicroFocus has always had a few case studies, but never much traction. As I got deeper into the mainframe world, I understood why. Plus, no one in their right mind would start a new Linux project using COBOL--they're all using Java, the new COBOL. Even Mike Murach, the mainframe publisher, makes Java books now. And Java will continue to put food on the table for many decades to come (I hope). Also...this is why C# and .NET aren't getting much traction, either, outside the PC desktop world. If you write your Java program, you'll probably have to run it on an actual mainframe in WebSphere. I haven't used Windows in years, and even that was as a terminal emulator for several years."

But back to our millions vs. billions... Samuel Ramos was gracious enough to say that "you and I both know it is 200 Billion and I would say that figure is a bit conservative. My old partner Clark Christine wrote that many in a single program."

Leon Stevens, who teaches computer science at Lincoln University of Missouri, wrote that "a number of business organizations, including the Missouri State Government, report that they are looking for COBOL programmers. Hint for those who didn't go for COBOL: Find out what your parent's didn't tell you: THERE IS a **STRONG** market for COBOL programmers."

-- Jonathan Erickson
This e-mail address is being protected from spam bots, you need JavaScript enabled to view it