For funders and press
The honest brief.
Code Like It's 198x teaches people how computers really workby building complete, finishable games on vintage machines. One person, an enormous ambition, built entirely in the open. Here's what's real today, what the roadmap holds, and where the line between the two sits — drawn honestly, because that's the whole point.
What it is
A curriculum that teaches the machine, not the framework
Most learning-to-code resources teach a language or a library. This one teaches the computer underneath — memory maps, cycle timing, sprites, sound chips — by having you build real games on the machines that made those ideas legible. The games are the sugar; the subject is computational thinking that outlasts any framework.
The deeper aim is preservation. The people who built these systems are retiring, and the why behind their work is evaporating faster than the code. The curriculum archives the mental models — and the Vault sets every machine, person and game against the world that produced it.
Scale and ambition
Catalogued wide, taught deep
The ambition spans the 8-bit, 16-bit and 32-bit eras, across home computers and consoles, and across computing scenes worldwide — not just the Anglosphere.
153 machines catalogued
The catalogue maps 153 machines — researched, classified, and queued for the curriculum. Catalogued is not the same as teachable. It's the map of where the work is headed, not a claim that all of it exists yet.
Four platforms teachable today
Complete on-ramps you can learn from now exist for four machines — theCommodore 64, Sinclair ZX Spectrum, Nintendo Entertainment System, Commodore Amiga. Each has a "Meet the Machine" primer and at least one finishable game.
A family of tools
The curriculum sits inside the wider 198x family — emulators (Emu198x), assemblers (Asm198x), asset cataloguing (Cat198x) and build tools (Build198x). Open-source siblings that share an origin, each with its own roadmap.
Original and enhanced
The curriculum teaches each system as it was, within its original constraints — and, where modern enhanced hardware exists (MEGA65, Spectrum Next, Vampire), how to use the extra room without losing the foundations.
How it's built
In the open, with nothing hidden
Free and openly licensed
Every page is free to read. The prose is Creative Commons (CC-BY); the code is MIT, public on GitHub. No ads, no paywall, no per-seat cost — and if the project ever stops, the content survives. One honest caveat: a couple of platforms need their original system ROMs to run, which aren't ours to give away.
One person, AI as scaffolding
This is built by Steve Hill, a software engineer of nearly four decades, alongside a day job. Large language models help draft lessons and accelerate the work — declared plainly, because a project about seeing all the way down shouldn't hide its own workings. The aim is to lean on AI less over time as real experts contribute.
The honest line
What's real today, and what isn't yet
Real today
- Four platforms with complete, finishable on-ramps
- BASIC and assembly paths where each machine had them
- The Vault, the Pattern Library, the Timeline and the Foundations
- Docker toolchains that build every code sample on the site
- Everything public, open source, free to read and fork
Roadmap, not yet built
- The wider catalogue of machines — mapped, not yet written
- A polished, drop-in pack for classrooms
- The deferred siblings (a workbench, a media player)
There are no fixed unit or game counts to hit, and the curriculum is not currently verified on real hardware — that verification is suspended. The bar is on technique and quality, not quantity, and we'd rather show the gap than fabricate a destination.
At a glance
Links and contact
Where to look, who to reach
The fullest statement of intent is the About page; the wider toolchain is on the family page. Everything happens in public:
For anything not covered here — interviews, permissions, partnerships —the contribution flow explains how the work happens, and Discord reaches a real person.