How Steve Wozniak Wrote Apple I BASIC on Paper in 1975

A copy of Steve Wozniak’s handwritten notebook has surfaced, showing how he designed the Apple I BASIC interpreter without a cross-compiler and with only 4 KB of RAM.
In 1975, Apple I did not yet have a working motherboard, and there was no cross-compiler for the MOS 6502 processor. So how did Steve Wozniak write BASIC for Apple I?
Recently, a copy of his original handwritten notebook from that time appeared online: handwritten notebook copy.
At the time, Wozniak could not afford time-sharing mainframe access, so he built the parser’s state table directly from the Wang computer manual.
Floating-point calculations consumed too much of the precious 4 KB of RAM. As a result, he abandoned them and created his own BASIC interpreter based on 16-bit integers.
The most impressive part is the stack of handwritten squared-paper sheets. On the right are 6502 assembly mnemonics, while on the left are the corresponding hexadecimal machine codes, or opcodes, and two’s-complement offsets for conditional jumps and jumps by address.
Wozniak found all of this in the instruction-set reference and calculated it manually.
If he changed just one line of code in the middle of the program, he had to recalculate all the offsets for subsequent jumps.
To fit both the interpreter and user code into 4 KB of memory, he used the 6502’s most valuable 256-byte memory region—the Zero Page—as a set of fast registers.
Right on paper, he managed to design a full virtual machine with two stacks: an operand stack and an operator stack.
Why it matters
AnalysisThe notebook shows how early software was hand-built with deep hardware awareness, and how severe memory limits shaped language design and architecture.
Discuss in community
Share your questions and insights with developers
Related in Ecosystem
Practical guides, core terminology, and agent skills
Codex: Basic Onboarding, Setup, and First Tasks
1. Three Approaches to AI Coding and Where Codex Fits In...
Codex: Basic Onboarding, Setup, and First Tasks
There are three ways to work with AI when writing code - and Codex covers the most ambitious of them.
Code Authorship Crisis: Who Wrote This Code?
A psychological state for developers resembling a new form of impostor syndrome, arising when the majority of the codebase is generated by language models, leading to a loss of emotional connection with the project, diminished professional pride, and doubts about one's engineering value.
Prompt Chaining Basics for Beginners
A methodology for breaking down complex tasks into a sequence of focused prompts (Prompt Chaining). The output of the first step feeds into the next, ensuring high quality for each segment.