March 11, 2020
Jef Raskin’s Macintosh proposal was, at its foundation, a software and interaction-design argument: a machine built around a specific set of beliefs about how non-expert users should be able to interact with a computer. But a design philosophy has to run on something, and the hardware Raskin’s team built it on was not a neutral, interchangeable substrate. It set real limits — on cost, on speed, on what the software could be asked to do — and those limits shaped the project as much as any of Raskin’s own design principles did. The engineer who built that hardware was Burrell Smith, and his part of the story is usually reduced to a supporting-cast credit in a narrative about Raskin and, later, Steve Jobs.
Who Smith Was, and How He Joined the Project
Smith was a self-taught hardware engineer working in Apple’s service department when he came to the attention of the small team Raskin was assembling in 1979. Accounts of his recruitment describe him being brought onto the Macintosh project by Raskin and Bill Atkinson, drawn in part from his informal reputation for elegant, minimal circuit design — an aesthetic that had already shaped Apple’s own machines and that Smith had absorbed and extended on his own initiative. He became one of the project’s first full-time engineers and, over a single Christmas holiday in 1979, built a working prototype essentially from scavenged Apple II components, matched to specifications Raskin had laid out.
That first prototype ran on a Motorola 6809E processor with 64 kilobytes of RAM and a 256-by-256-pixel frame buffer — a modest, low-cost design consistent with Raskin’s founding insistence that the machine be genuinely inexpensive, not merely marketed as such. The 6809E was itself a constraint dressed up as a design choice: it was a processor Raskin’s cost targets could plausibly support, not necessarily the processor best suited to the graphical, bitmap-driven interface Raskin’s own writing described.
The Processor Change That Reshaped the Machine
The single hardware decision with the largest downstream effect on what the Macintosh could become was the move away from that original 6809E design toward the Motorola 68000 — the more powerful processor Apple’s separate, more expensively funded Lisa team was already building around. Accounts of the switch describe Apple engineer Bud Tribble recommending the change to Smith around the end of 1980, reasoning that the 68000’s greater headroom would let the Macintosh support a substantially richer interface than the 6809E-based design could manage.
Smith’s contribution here went beyond simply swapping chips. He is credited with an unusually compact, cost-conscious implementation that folded peripheral control functions directly into the processor design, running the 68000 at roughly 7.83 megahertz with a 16-bit external data bus, wired to squeeze real capability out of a machine still bound by Raskin’s original insistence on an affordable price point. The tension in that sentence is the whole story: more powerful hardware, still constrained by a cost target that predated the more powerful hardware.
The 128K Constraint That Outlived the Processor Change
The processor upgrade solved one bottleneck and left another firmly in place: memory. The Macintosh that shipped in January 1984 carried 128 kilobytes of RAM — a figure driven directly by the cost of memory chips at the time, not by any engineering judgment about what the software actually needed. It was not enough. Contemporary reviewers, including the widely read BYTE magazine assessment published the same year, flagged the memory ceiling as a genuine limitation on how capable a machine the Macintosh could be, particularly given how much RAM the graphical interface itself consumed before an application ever loaded — accounts of the machine’s memory map describe roughly 21 kilobytes reserved for video display alone, leaving well under 107 kilobytes for the operating system and whatever program the user was running. Apple’s response came quickly by industry standards: a 512-kilobyte version shipped later the same year, and the original was retroactively renamed the “Macintosh 128K” to distinguish the two.
This detail belongs in the same discussion as the processor upgrade because it illustrates the same underlying dynamic from the opposite direction. Where the 68000 upgrade expanded what Raskin’s original cost-constrained machine could do graphically, the 128K memory ceiling constrained it, for the identical underlying reason — component cost, not design philosophy, setting the boundary. Any account of “how faithfully did the shipped Macintosh follow Raskin’s original vision” has to reckon with the fact that both the expansions and the limitations of the shipped hardware trace back to the same category of decision: what Smith’s team could build within a cost target, adjusted as component prices and competitive pressure changed between 1979 and 1984.
What This Means for the “Raskin’s Vision” Question
It is tempting to treat the hardware as a separate track from the “real” story of Raskin’s design philosophy and Jobs’s later takeover of the project’s direction. That separation doesn’t hold up. The processor upgrade happened in late 1980 — while Raskin was still nominally leading the project, before Jobs’s confrontation with him in early 1981 — and it materially expanded what the eventual, Jobs-era Macintosh could do graphically. Any comparison between Raskin’s 1979 memos and the shipped 1984 machine has to account for the fact that the machine’s graphical capability grew substantially during the period Raskin was still involved, on hardware decisions that were, at least at that stage, still made inside the framework Raskin had set.
Put differently: the widening gap between Raskin’s original conception and the shipped Macintosh is often narrated purely as a story about Jobs redirecting the software and interface philosophy. Some of that gap opened earlier and for a more mundane reason — a hardware engineer and his collaborators concluded the original chip couldn’t support the interface ambitions the project already had, and upgraded it, months before the personnel fight that gets most of the historical attention.
Smith’s Design Values, and Their Kinship With Raskin’s
There is a real affinity between Smith’s engineering values and Raskin’s design values, even though they operated in different domains. Raskin’s obsession, across his career, was eliminating anything that asked more of the user than the task genuinely required — no unnecessary modes, no unnecessary configuration, no unnecessary complexity standing between the person and the work. Smith’s reputation as a hardware engineer rested on a parallel obsession: minimizing chip count, minimizing cost, achieving a given capability with the fewest components that could reliably deliver it. Neither man was building a maximalist machine. Both were building toward a specific kind of restraint, in their own respective layers of the system.
That kinship is part of why the recruitment makes sense in retrospect — Raskin was not simply hiring an available engineer, he was hiring one whose instincts, translated from circuit design into cost discipline, aligned reasonably well with the low-cost, no-frills machine Raskin had proposed. The instinct held even as the specific chip changed under them.
What the Hardware Story Adds to the Record
None of this displaces the software and interface-philosophy history that is this site’s main subject. But treating the hardware as a footnote understates how much of what became possible — and how much of what Raskin’s original, more austere 1979 machine had to give up — was decided at the board level, by an engineer working from Raskin’s cost constraints toward a processor Raskin himself had not originally specified. The Macintosh that shipped in 1984 was shaped by memos, by a leadership fight, and by a circuit board redesigned in late 1980 for reasons that had little to do with either.
Frequently Asked Questions
Did Jef Raskin recruit Burrell Smith personally?
Accounts of the Macintosh project’s early staffing describe Smith being brought onto the team by Raskin together with Bill Atkinson, drawing on Smith’s informal reputation within Apple for minimal, elegant circuit design before he had any formal role on the project.
What processor did the first Macintosh prototype use?
The earliest working prototype, built by Smith over the 1979 Christmas holiday, used a Motorola 6809E processor with 64 kilobytes of RAM — a deliberately modest specification consistent with Raskin’s insistence on a genuinely low-cost machine.
When did the Macintosh switch to the Motorola 68000 processor?
Accounts place the recommendation to move to the more powerful 68000 — the processor already used by Apple’s separate Lisa project — around the end of 1980, prior to Steve Jobs’s takeover of the Macintosh project’s leadership in early 1981.
Did the hardware change happen because of Steve Jobs?
The documented sequence places the processor upgrade before Jobs’s confrontation with Raskin in early 1981, suggesting the decision was driven by engineering judgment about what the project’s existing interface ambitions required, not by Jobs’s later redirection of the project.
Was Burrell Smith’s design philosophy similar to Raskin’s?
There is a documented parallel: Raskin’s interface philosophy prized eliminating unnecessary complexity for the user, and Smith’s hardware engineering reputation rested on a similar discipline — achieving a given capability with the fewest, most efficiently used components.
Related: