What it is and the problem it solves
It is a stored-program computer built at the Victoria University of Manchester from the Manchester Baby. It solved the problem of executing long, deterministic computations reliably on hardware that had no prior track record of sustained operation.
How it works
It used a 40-bit word length, holding either one 40-bit number or two 20-bit instructions. It stored data in Williams tubes backed by a magnetic drum using Manchester coding. It implemented index registers to enable sequential memory access. Its instruction set included hardware multiplication and grew to 26 operations.
What works
A Mersenne prime search ran error-free for nine hours on 16/17 June 1949. Index registers worked as intended for sequential memory access. Hardware multiplication executed correctly. The Williams tube–drum memory hierarchy held state across program runs.
What does not
It did not scale beyond its university lab. It had no operating system, no high-level language, no I/O abstraction, and no error recovery beyond manual intervention. Its memory was volatile, slow, and prone to drift.
What it changes
It changed how programs could traverse data: index registers enabled array iteration without self-modifying code. It changed industrial computing by becoming the direct prototype for the Ferranti Mark 1—the first commercially available general-purpose computer.
Is it worth your time
Yes—if you work on early computer architecture, instruction set design, or the material foundations of programmability. Its operational rigour and documented failure-free runtime make it a rare anchor point in computing’s pre-commercial phase.
