Thursday 29 March 2007

Bug Fix

We have corrected a bug in the www.cobol.tv buy page so you can now go to the Fujitsu Cobol support page correctly. We apologise for this error. If you find any other errors on the site, please email these to the address shown on the buy page.

Monday 26 March 2007

A Personal Cobol

I first met Cobol in 1967. Already a programmer in ICT 1300 machine code and punched card systems I thought this new language was verbose and clumsy. But even then Cobol had a major advantage in being (to some extent) portable across computer ranges from various manufacturers. At that time, programming was a laborious process. Having set out your task you would write your code on a specially designed coding form, being careful to be very precise and legible with your printing. Once complete, the pad of coding forms would be submitted to an operator who would accurately (you hoped) key in your code and create a pack of punched cards, one line to one card. If the punch room was busy, you would have to wait some hours to try your program.

It was important to start each line of code with a sequential line number. Why? Well, it was all to easy to drop your precious card pack and watch with horror as your 500 card program exploded across the computer room floor. By using line numbers punched in the cards, you could use a card sorter to rapidly restore your program to its original sequence. Then you would load your program into the computer card reader and type your compile command into the console. After five to ten minutes, the compile would finish and you would have a program listing on the printer. Inevitably, many errors would be found by the compiler and your would have to recode, re-punch and recompile your program, repeating the cycle until no errors remained.

Then your program was ready for testing. With no compile errors I was supremely confident that my first program would work. Wrong! A working program was still days away after many testing and debugging runs had been carried out.

A few years later, mini computers were developed and these could also run Cobol. I remember my first "personal" computer, a Burroughs B80 the size of two large desks shoehorned into our utility room at home. This "visible record computer" had a keyboard, printer, magnetic stripe ledger card processor and a tiny screen able to hold just a few short lines of data. I was immediately grateful to my wife, not just for allowing such a monstrosity into our house but for teaching me to type properly. Now I was able to type my own programs directly into the computer and modify them "online". But "mini" computer also meant mini power and compiling a Cobol program could take between 60 and 90 minutes.

Only the more expensive mini computers could run Cobol and the prices were still beyond the reach of many organisations. So for a while I programmed Wang and Dec computers in Basic and Fortran, returning to Cobol on ICL 2900 mainframes in the early 1980s. Now we had "terminals". Screens and keyboards connected to the mainframe enabling a number of programmers to simultaneously compile and test software.

I bought my first microcomputer. Sirius One. An early 16 bit micro with 2 floppy drives and Cobol. Using CIS Cobol (a "lite" Cobol) development could at last be done at a reasonable price. Today, with microcomputers many times the power of the old mainframes and the availability of superb construction tools, Cobol is in a golden age.

Thursday 1 March 2007

Fujitsu Site Change

Thanks for letting us know that Fujitsu have changed their site from www.adtools.com to www.netcobol.com. We have now changed the links on the site to the new URL so they should work OK. Please let me know if there are any more problems.