Forty playable exhibits, eight eras — and the machines between them.

This timeline connects every playable exhibit while also naming the major systems that belong to the same historical path but do not yet have an emulator. The point is the history of personal computing, not a chronology artificially limited by the current codebase.

How to read this timeline

The dates are organising periods, not claims that an architecture suddenly began or ended at a boundary. Many machines overlapped for years, and regional markets differed considerably. Each era groups the collection around the engineering and commercial argument that becomes easiest to see at that point.

1974–1979

Foundations: the microprocessor becomes a computer

The first decisive change was economic. A useful CPU no longer needed a board full of small-scale logic: the Intel 8080 and MOS Technology 6502 put general-purpose processing into affordable single chips. That created two complementary routes into personal computing. One route was the enthusiast board, represented here by the KIM-1, where the owner interacted almost directly with the processor through a monitor program, hexadecimal keypad and a few hundred bytes of RAM. The other was a portable software layer: CP/M separated applications from the details of one particular 8080 or Z80 machine by concentrating hardware dependence in a relatively small BIOS.

The Atari VCS shows a third idea arriving almost immediately: instead of treating the CPU as the star, build the product around specialised video hardware and make the processor feed it just in time. Its tiny RAM and timing-sensitive Television Interface Adapter make it almost the opposite of a general-purpose CP/M computer. Those three exhibits establish recurring themes for the rest of the museum: software portability, low-cost CPUs and custom silicon.

Also in the wider story: MITS Altair 8800, IMSAI 8080, Apple I, Sol-20 and early S-100 systems established the hobbyist and kit-computer world around the 8080/6502 generation. The interactive Apple-1 architecture lab lets you slow that complete 6502/RAM/ROM/I/O structure down to individual teaching steps.

1980–1981

The computer becomes an appliance — and then an industry standard

PET and VIC-20 show Commodore turning the computer into a complete consumer product. The PET integrated display, keyboard, storage expectations and computer into one recognisable appliance; the VIC-20 pushed the price downward and the retail reach outward. Sinclair took that logic further with the ZX81, accepting severe display and usability compromises to make a computer radically cheaper. The BBC Micro went in the other direction: more expensive, robust, expandable and strongly tied to education.

Texas Instruments took yet another route with the TI-99/4A. Its TMS9900 was a 16-bit processor derived from TI’s minicomputer architecture, paired with TMS9918A-family video and GROM-based firmware. The result looked advanced on a specification sheet but was constrained by the surrounding memory and I/O architecture. Its later price war with Commodore became an unusually clear demonstration that silicon width alone could not decide the home-computer market.

IBM's 5150 PC changed the axis of competition. It was not simply another technically distinctive home computer. Its use of commercially available parts, published expansion architecture and a software environment that could be cloned created a platform larger than IBM itself. The later XT hardened that model with built-in hard-disk expectations. From this point onward the museum contains two simultaneous histories: distinctive proprietary machines fighting to offer a better experience, and compatible machines benefiting from a growing common ecosystem.

At the same time, Sharp was shrinking a different idea. The PC-1500 of 1981 put BASIC, a QWERTY keyboard, numeric keypad, expandable memory and a graphics-addressable LCD into a calculator-sized pocket computer; the later PC-1350 made the display much more capable. These machines are important because they show portable personal computing emerging from the programmable-calculator world in parallel with desktop micros.

At the calculator end of the same portability story, Hewlett-Packard's Voyager series arrived in 1981–82. The HP-10C, 11C, 12C, 15C and 16C deliberately remained specialist RPN calculators rather than pocket BASIC computers, proving that a persistent keyboard-driven expert platform could be as durable as a miniature general-purpose computer.

Also in the wider story: Apple II/II Plus, TRS-80 Model I/III, Sinclair ZX80, Atari 400/800 and Acorn Atom show how many incompatible home-computer ecosystems were already competing before the 1982 boom.

1982

The 8-bit explosion: price, personality and portability

In 1982 the market fragmented dramatically. The Commodore 64 paired a cheap 6510-class CPU with unusually capable VIC-II graphics and SID sound, allowing Commodore's manufacturing scale to turn custom silicon into an aggressive price weapon. The ZX Spectrum offered a very different compromise: minimal hardware, low price and a huge software culture. Neither machine won because of CPU performance alone; graphics architecture, sound, memory cost, retail distribution and software mattered more.

The same year also produced machines that do not fit the conventional home-computer story. Vectrex put a vector display inside the console itself, creating an image fundamentally different from raster television graphics. Epson's HX-20 compressed a real general-purpose computer, keyboard, LCD, storage and battery operation into a portable format. Together they make 1982 more than “C64 versus Spectrum”: it is also the point where display technology and physical form factor begin to diverge sharply.

Also in the wider story: Dragon 32, Oric-1, ColecoVision, Tandy/Radio Shack Model 100, HP-75C/71B, TI CC-40, Psion Organiser and portable/business systems such as Osborne, Kaypro and Epson PX-8 make the early-1980s field much broader than the machines currently playable here.

1983

Standards, longevity and the console reset

Apple's IIe demonstrates the value of continuity: rather than discard a successful architecture, Apple refined an ecosystem whose roots reached back to 1977. Atari's XL family represents the opposite kind of continuity, keeping the ANTIC/GTIA/POKEY custom-chip architecture competitive through revised packaging and memory. MSX attempted something more ambitious — a multi-vendor home-computer standard around a common Z80/video/sound architecture, seeking software compatibility without requiring one manufacturer to own the whole market.

Nintendo's Famicom/NES arrived as the North American console business was collapsing under poor quality control and a flooded cartridge market. Nintendo treated the software ecosystem itself as part of the hardware platform, combining custom PPU graphics, strict cartridge licensing and controlled distribution. The result matters because it shows that “platform architecture” includes business rules and developer access as well as CPUs and buses.

Also in the wider story: Apple Lisa, Acorn Electron and IBM PCjr illustrate three very different attempts to redefine graphical, educational and home-oriented personal computing.

1984–1985

The mature 8-bit machine meets the 16-bit future

Amstrad's CPC 464 packaged computer, monitor and cassette into a coherent ready-to-use system and competed on convenience as much as architecture. Commodore's C16 tried to cover a lower-cost segment with the TED chip, but sacrificed the compatibility and software momentum that made the C64 so strong. The C128 then took the opposite route: it explicitly preserved C64 compatibility while adding a faster 8502 mode, a second Z80 CPU for CP/M and the independent 8563 VDC for 80-column display.

At the same moment the Atari ST made the Motorola 68000 and a graphical desktop available at a comparatively low price. The address-space and register model of the 68000 removed many of the contortions that defined 8-bit programming, while higher-resolution displays, mice and disk-based GUI software changed the expected interaction model. This era is therefore not a clean handover; it is a period in which vendors tried compatibility, segmentation and architectural replacement simultaneously.

Also in the wider story: Sinclair QL, Apple Macintosh 128K, Apple IIc, IBM PC/AT, Amstrad PCW and Commodore Plus/4 show the market splitting into GUI, compatibility, portable and productivity branches.

1985–1987

Custom chips versus the general-purpose graphical computer

Sega's Master System kept the Z80 generation alive in consoles with a more capable video subsystem and cartridge hardware. In personal computers, however, the 68000 became a common CPU beneath radically different philosophies. Macintosh paired it with a tightly controlled graphical environment and a strong emphasis on the mouse and desktop metaphor. Atari ST aimed for affordability and relatively conventional system integration.

The Amiga 500 made custom co-processors the centre of the design. Agnus, Denise and Paula handled display generation, blitting, DMA and multi-channel audio in ways that let the 68000 orchestrate rather than perform every operation itself. Comparing Macintosh, ST and Amiga is therefore especially instructive: similar CPU family, very different division of labour between processor, memory subsystem and specialised silicon.

A quieter display revolution was happening in the calculator aisle. Casio introduced the fx-7000G in 1985 as the first graphing scientific calculator; the fx-8000G followed in the same early generation with more memory and I/O. A 96×64 LCD turned the scientific calculator into a visual programming and graphing environment years before handheld devices had general-purpose graphical user interfaces.

Also in the wider story: Amiga 1000, Apple IIGS, Acorn Master and Archimedes, Atari STE and later 68000-family workstations show how quickly the 16/32-bit landscape diversified.

1989–1994

Compatibility becomes the mainstream — while portable machines choose different compromises

By the beginning of the 1990s, the IBM-compatible PC was no longer merely one architecture among many. The 486 generation combined a much faster x86 processor with VGA/SVGA graphics, increasingly capable hard disks and a rapidly standardising ecosystem of sound cards, buses and DOS software. Compatibility let buyers change manufacturers without abandoning the software investment that mattered to them.

Portability did not follow one blueprint. Atari’s Portfolio compressed DOS-like compatibility into a palmtop, while Nintendo’s Game Boy deliberately chose modest processing and display hardware in exchange for low cost, long battery life and a huge software library. Casio continued the pocket-computer branch into more capable programmable machines, and HP’s 48 series turned a calculator into a persistent RPL engineering environment whose interface mattered as much as its processor.

This period is important because the desktop and the pocket begin to optimise for different constraints. The desktop rewards compatibility, expansion and multimedia peripherals; handhelds reward battery life, instant access, durable interfaces and software that fits the physical form.

Also in the wider story: 386/486 clone makers, Windows 3.x, Sound Blaster, Super VGA, Macintosh II and Quadra, PowerBook, Amiga 1200/4000, Atari Falcon, Acorn Archimedes/Risc PC, NeXT and Psion Series 3 show how broad the early-1990s landscape still was.

1995–2006

Multimedia, connectivity and mobile computing redefine the personal machine

From the mid-1990s, the compatible PC absorbed capabilities that had once distinguished specialist machines: high-colour graphics, sampled audio, CD-ROM, networking, 3D acceleration and increasingly graphical operating systems. The platform was becoming less a particular IBM-derived box than a continuously evolving compatibility contract between processors, firmware, operating systems, expansion hardware and applications.

The most revealing changes happened away from the desk. Palm III and IIIc prioritised instant-on response, stylus input and personal information management rather than desktop-style generality. Game Boy Advance brought ARM into a mass-market handheld. Texas Instruments kept two calculator lineages alive in parallel: the Z80-based school graphing family and the Motorola 68000 symbolic line from TI-92 through TI-89 and Voyage 200, while HP’s 50g preserved the RPL/Saturn programming model even after the physical processor underneath had moved to ARM.

By 2006, personal computing had therefore converged and fragmented at the same time. The desktop had largely standardised around compatible commodity hardware, while phones, PDAs, consoles and calculators succeeded by keeping specialised interaction models. The history does not end with one winning CPU: it moves into a world where software ecosystems and interfaces can outlive the silicon that first created them.

Also in the wider story: Windows 95 and later PC generations, PowerPC Macintosh, PlayStation, Saturn and Dreamcast, Psion Series 5, Newton, early smartphones, the later TI-Nspire and HP Prime lines extend these threads beyond the current emulator set.

Research note

Dates, specifications and market claims are cross-checked against original manuals/service documentation and established archival histories where possible. Sales figures and “first” claims are treated cautiously because definitions and regional figures often differ. See research method and source collections →

The recurring pattern

Every apparent winner solved a different constraint.

CP/M solved software portability across incompatible hardware. Commodore solved cost. Nintendo solved software quality and platform control. IBM compatibility solved vendor independence and ecosystem scale. Game Boy solved battery life and price. Palm solved friction in everyday pocket use. The technical history becomes much clearer when those constraints are kept visible beside the processor specifications.