History of Assembly

Before Assembly: Switches, Cards, and Raw Numbers

beginnerHISTORYLesson 1 of 2

Long before anyone typed mov or addi, programming meant being the compiler — in your head, on paper, with numbers.

Programming the hard way

The earliest stored-program computers had no language at all. To run a program you fed the machine opcodes directly: numeric codes the CPU was wired to recognize. On machines like the EDSAC (1949) or the IAS machine, a programmer would:

1. Work out the algorithm by hand.
2. Translate each step into the machine’s numeric instruction format.
3. Enter those numbers — via punched paper tape, punch cards, or literally toggling switches on a front panel, one bit at a time.

A single mistake meant hunting through columns of octal or decimal by eye.

The machine never understood “add these two numbers.” It understood “the bit pattern that, when decoded, drives the adder.” Everything human about the program lived only in the programmer’s notes.

Why this couldn’t last

Two problems made raw machine code untenable as programs grew:

• Addresses were absolute. Insert one instruction near the top and every jump target below it shifted — you re-computed addresses by hand.
• Numbers carry no meaning. 00110010 tells you nothing; you had to remember that it meant “load.” Code was unreadable an hour after writing it.

The fix was almost obvious in hindsight: let the human write names, and let the machine do the bookkeeping of turning names into numbers. That idea — a program that assembles your symbolic instructions into machine code — is where assembly language begins.

What to take away

• Machine code is the CPU’s native tongue: pure numbers.
• Assembly is a thin, human-readable naming layer over that — not a different language so much as a notation.
• The motivation was never speed of execution; it was speed and safety of authoring.

In the next lesson we’ll meet the first assemblers and the mnemonics that are still recognizable in today’s code.