IBM 1130 (16758008839)

Remember Robert, the student who ruined his class curve back in the 1960s? Well, proving the old adage that the guy who graduates last from medical school is still a doctor, he managed to find another part-time job at a small hospital, earning just enough to pay his continued tuition.

Industry standard in those days was the IBM System/360 series, but it was out of the price range of this hospital. Instead, they had an IBM 1130, which was designed to be used in laboratories and small scientific research facilities. It used FORTRAN, which was pretty inappropriate for business use, but a set of subroutines offered by IBM contained routines for dealing with currency values and formatting. The hospital captured charges on punch cards and those were used as input to a billing program.

The printer was a monstrous beast, spinning a drum of characters and firing hammers to print characters as they went by. In order to print in specific boxes on the billing forms, it was necessary to advance the paper to a specific point on the page. This was done using a loop of paper tape that had 12 channels in its width. A hole was punched at the line in the tape where the printer needed to stop. Wire brushes above the tape would hit the hole, making contact with the metal drum inside the loop and stopping the paper feed.

There was one box in the billing form that was used infrequently, only every few days. When the program issued the code to skip to that channel, paper would begin spewing for a few seconds, and then the printer would shut down with a fault. This required stopping, removing the paper, typing the necessary data into the partially-printed bill, and then restarting the job from the point of failure.

IBM Field Engineering was called, but was unable to find a reason for the problem. Their considered opinion was that it was a software fault. After dealing with the problem on a fairly regular basis, things escalated. The IBM Systems Engineer assigned to the site was brought in.

Robert's boss, the author of the billing software, had relied on an IDEAL subroutine package provided by IBM—technically unsupported, but written by IBM employees, so generally one would assume it was safe to use. The Systems Engineer spent a while looking over that package, but eventually declared it innocent and moved on. He checked over the code Robert's boss had written, but ultimately that, too, failed to provide any answers.

"Then it must be the machine," Robert's boss stated.

This was the wrong thing to say. "It couldn't be the machine!" The Engineer, a prideful young woman, bristled at the insinuation. "These machines are checked. Everything's checked before it leaves the factory!"

Tempers flared, voices on the edge of shouting. Robert ducked back into the room with the computer, followed rapidly by the Field Engineer who had come along earlier in the day to do his own checks. Trying to pretend they couldn't hear the argument, the pair began another once-over on the machine, looking for any sign of mechanical fault.

"Hey, a question," said Robert, holding the thick cable that connected the printer to the computer. "Could it be a problem with the cable?"

The Field Engineer unplugged the cable and examined it. "The pin for that channel doesn't look seated," he admitted sheepishly. "Let's replace it and see what happens."

That day Robert learned two valuable lessons in debugging. Number one: when in doubt, go over each piece of the machine, no matter how unlikely. Number two: never tell an IBM Engineer that the problem is on their end.

[Advertisement] BuildMaster allows you to create a self-service release management platform that allows different teams to manage their applications. Explore how!