August 19, 2010 at 23:28Favorite Iron Bug
- Transfer
Larry Osterman: My favorite bug is the one we found on the ICL PWS-400. It was a new platform from ICL (a British company, bought by Fujitsu in 2002 ), and our task was to port MS-DOS 4.1 to this platform. Five were involved in the project - two from Microsoft and three from ICL.
The hardware of the PWS-400 was rather unusual: for example, real-mode programs could switch memory banks one page at a time — due to this application they could run in the background without interfering with one another.
There were five of us, we had no testers; therefore, we tested the new system with everything that comes to hand. My favorite “testing tool” was a game that Valori ( wife) brought from school. I don’t remember exactly what the game was; but every time I played it and reached a certain place, the car suddenly rebooted.
We connected the ICE to the machine (in-circuit emulator - a device that allows you to find out what is happening inside and outside the processor), and found that the processor receives a reset signal from the outside. So at least the bug was not in the code.
The guys who assembled the car took that game from me and began to sort it out.
A few days later they returned the game and said that they had found a malfunction. It turned out that the track to the speaker passed on the motherboard too close to the reset track. When a certain type of signal was supplied to the speaker, the electromagnetic radiation from the first track induced a sufficiently strong voltage on the second track for the processor to recognize it as a signal to restart.
Matt Williams: I remember a couple of bugs that drove me crazy.
The first story happened in New York City. A mouse was connected to one of the computers, which randomly jumped, as soon as she was guided along a certain part of the desktop. But this did not always work. After some time, we found a pattern: the mouse went crazy only at a certain time of the day. And the place on the table in which she went crazy also moved with time. We spent a lot of time before we discovered the cause of what was happening. But first, I’ll tell you about the second bug: they are related to one another, although they have been separated for about 5 years.
By that time, I moved to California and purchased a used car. Before giving it to me, the previous owners talked about a strange feature of the radio: for no reason spit out the inserted disk. They contacted the workshop several times about this, but they could not find a malfunction, and the mechanics simply replaced the radio with a new one, of the same model. This did not help: the fault remained.
Having traveled by this car for several months, I discovered a pattern: the radio tape recorder spit out the disk only when I was driving in a certain direction ... and, moreover, only at a certain time of the day.
So what was the matter? Both times the sun was to blame. The mouse had a small gap between the buttons, and when the sun shone at it at a certain angle, the light fell inside the photocouples and caused a reaction, like from a ball rotating in all directions. In the same way, when the sun fell through the skylight into the loading slot of the radio, the sensor of the inserted disk worked, and the radio spewed out.