History Software Engineering

Manchester SSEM – 1948 AD

Return to Timeline of the History of Computers


Manchester SSEM

Frederic Calland Williams (1911–1977), Tom Kilburn (1921–2001)

“The defining characteristic of the digital computer is that it stores both program and data in a single memory bank. In a modern computer, this arrangement lets one program load a second program into memory and execute it. On the limited-memory machines of the 1950s, intermixing programs and code made it possible to squeeze out more functionality by writing programs that literally modified themselves, now called self-modifying code. Modern computers use this ability to load code into the computer’s memory and execute it—the fundamental capability that makes a computer a general-purpose machine. But none of the machines built before the Manchester Small-Scale Experimental Machine (SSEM) were actually digital computers, at least not in the modern sense. Either they were hardwired to perform a particular calculation, like the Atanasoff-Berry Computer, they read their instructions from some kind of punched tape, like the Konrad Zuse machines, or the program was set on wires and switches, like ENIAC. They were really calculators, not computers.

The SSEM, nicknamed Baby by its creators at the University of Manchester, was built for testing and demonstrating the storage tube that Frederic Williams had designedWilliams Tube Random Access Memory (RAM) – 1946 A.D. in 1946. Baby filled a 20-foot-square room and consisted of eight racks of equipment, the Williams storage tube, many radio tubes, and meters that reported voltages. Each tube had 1,024 bits. As the program ran and changed what was stored in its memory, the arrangement of dots on the storage tube changed.

Because the program was stored in memory, and relied on self-modifying code, it was easy for Kilburn to make changes. The first program that Baby successfully ran, written by Kilburn, was designed to find the highest factor of 218 (262,144). The program ran in 52 minutes and found the right answer: 217 (131,072), averaging 1.5 milliseconds per instruction. The original program was just 17 instructions long.

Arriving at the correct answer was no easy feat. As Williams reportedly stated, “The spots on the display tube entered a mad dance. In early trials, it was a dance of death leading to no useful result . . . But one day it stopped, and there, shining brightly in the expected place, was the expected answer.””

SEE ALSO Z3 Computer (1941), Atanasoff-Berry Computer (1942), Williams Tube (1946)

“Recreation of the Manchester Small-Scale Experimental Machine (a.k.a., the Manchester “Baby”) at the Museum of Science and Industry in Manchester, UK.”

Fair Use Source: B07C2NQSPV