The SCREAM Test has been around for a long, long time. It is a critical tool for any computer programmer or system administrator.
Don’t know what a particular server does? Turn it off.
Don’t know what some code does? Comment it out.
Then sit back and wait for the screams.
Not exactly sure where I first heard of “The SCREAM Test”, but I believe it was some time in the late 1990s while working at Hewlett-Packard. My job was to travel around and fix UNIX systems at big companies who had support contracts with HP.
I recall a time when I was working on-site at Boeing, in the engineering department for some of their airplanes. Back in those days Boeing used a wide variety of UNIX workstations — both old and new. There was one, rather dusty UNIX box — perhaps an HP-UX machine, or one of the older Apollo/Domain systems… it was one or the other — that was sitting there making a gnarly grinding noise. Clearly about to die. Or explode. Or dig a hole to China. Or something. Regardless, I was there to fix it.
There was a stickie note on the monitor that clearly read: “DO NOT TURN OFF.”
When I inquired as to whether or not the machine could be taken Off-Line for a while to be repaired… nobody seemed to know. In fact, nobody had any clue what that machine was actually used for. Nor who might have written that particular note, on that particular stickie.
At that point, one of the Boeing engineers — an older gentleman, who possessed one of the most spectacular neck beards I had ever seen — suggested we just turn it off and run a SCREAM test.
Long story short: The test worked. We found out what that system was used for. Rather quickly.
Because The SCREAM Test works.