A while back, Ars Technica published a thoughtful piece about Jef Raskin, tracing his long pursuit of the “humane computer” and the cul-de-sacs where that pursuit ended. It’s a generous, well-told account of the designer who wanted to make machines simpler, kinder, and more aligned with the way people actually think.

But part of what makes Raskin interesting is that his story isn’t just Apple’s story. He came out of the same cultural current John Markoff chronicled in What the Dormouse Said—the Bay Area tradition that treated computers not as office appliances but as tools for thought, instruments of liberation. Read that way, the Canon Cat and Raskin’s other projects aren’t just an eccentric side quest from a frustrated Apple veteran. It’s evidence of how far the humane ideal could stretch, and how quickly it ran up against the limits of commercial computing.

Apple couldn’t deliver Raskin’s vision then, and it can’t deliver it now. Neither can any other big platform company. If we want to understand why, and what Raskin still tells us about humane computing, we have to put him back in the longer lineage he belonged to, and look at how his version of the dream carried that vision but also narrowed it.

Prophets and participants

What the Dormouse Said documents how the Bay Area counterculture shaped early personal computing. LSD, communes, systems theory, amorphous defense research contracts, and Engelbart’s “augmentation” experiments all swirled together in a weird scene that accidentally (or maybe not so accidentally) created much of the modern world.

The story usually gets told with a neat list: Engelbart’s demo, Nelson’s Xanadu hypertext, Kay’s Dynabook, Brand’s Whole Earth. Xerox PARC, Steve Jobs, the World Wide Web. The familiar pantheon. But that version turns a messy, improvisational moment into a plaque. Engelbart’s system needed a whole research staff just to operate; Nelson’s Xanadu was (and is) more sermon than software; Kay’s Dynabook lived mostly on paper; Brand mostly supplied vocabulary and vibe. What bound them together wasn’t working code so much as the conviction that computers could be more than appliances and calculators, even if no one agreed on what “more” meant.

Ultimately all these weird white guys had a futurist vision: computers could be liberation machines. They weren’t just for business automation or scientific number-crunching; they could be deployed to expand consciousness and reshape how people thought and worked.

Raskin belonged to this current. Before Apple, he was an artist and a musician. He brought a humanist’s suspicion of machine logic into the design lab. He argued for humane interfaces: modeless, predictable, low-friction, focused on the human first. He wasn’t a prophet on his own crying in the wilderness so much as another strand of the same weave.

That said, his role was different than that of some of these other figures. He tried to pull those ideals out of the lab and into machines ordinary people might actually use. The Macintosh began under his hand, though what shipped was less a tool for thought than a polished derivative—what you might call a “popular religion” of computing, stripped of the harder doctrines.

The Canon Cat and its predecessors were Raskin’s counterargument: humane, text-first systems that tried to carry the spirit of the Dormouse tradition into the commercial world without sanding off everything that made it strange. It sort of worked, but only sort of.

Raskin’s Humane vision

Raskin’s principles are laid out most clearly in 2000’s The Humane Interface (2000), but he’d been developing them since the late 1970s:

• Modelessness: eliminate modes generally, and especially when they confuse users or are hard to reason about.
• Quasimodes: short-lived states (like holding a key down) that don’t trap the user.
• Humane defaults: undo everywhere, consistent commands, predictable behavior.
• Low cognitive load: interfaces designed around human memory and perception limits. These ideas are recognizably part of the “Tools for Thought” tradition. Like Engelbart and the others, he wanted to reduce friction between thought and machine. Like Nelson, he believed in fluidity and extension.

But there’s a subtle difference. For Engelbart, augmentation meant complexity: bootstrapping a system so wild it demanded co-evolution between user and tool. For Nelson, it meant endless layers of possibility. For Raskin, it often meant protection or constraint. Humane computing wasn’t only about empowerment… often it was about shielding users from mistakes, overload, and confusion.

That protective impulse would shape the systems he built.

Raskin’s first clear articulation of his humane ideals wasn’t hardware at all but The Macintosh Papers, his internal proposal at Apple for a low-cost, appliance-like computer that would boot straight into a simple, modeless interface. The Mac project that followed eventually diverged—under Steve Jobs it became a graphical machine aimed at competing with the Lisa, for the reasons we’ve all read about—but Raskin’s vision was considerably more radical. He imagined a computer that behaved less like a business workstation and more like a humane, everyday tool.

In tone, the Macintosh Papers have more in common with Ted Nelson’s Computer Lib than with any corporate white paper. They read like a manifesto: plainspoken, insistent, arguing that ordinary people deserved machines that bent to them rather than the other way around. Where Nelson declared that “you can and must understand computers now,” Raskin’s papers laid out what such a computer should look like if you started from human needs instead of technical conventions. Both belong to that peculiar genre of the 1970s and early ’80s: the computing manifesto as cultural text, half engineering and half tract.1

You could argue that the Swyft, built a few years later by his company Information Appliance Inc., was “the real Macintosh” in that sense. Compact and text-first, it booted instantly, eliminated modes, and introduced the Leap keys for fluid navigation. It was Raskin’s manifesto rendered in hardware. But the Swyft never made it to market; without a manufacturer to back it, Information Appliance pivoted to the SwyftCard, a fallback product that brought the same interface into the Apple II while IA waited to find a dance partner.

That partner came briefly in 1987, when Canon released the Canon Cat, the only mass-produced computer to carry Raskin’s humane vision into the world. The Cat retained the Swyft’s defining ideas: instant boot into a blank page, consistent commands, Leap-based navigation. Marketed as a word processor, it was framed as an appliance for the office rather than an exploratory tool for thought.

After its failure, Raskin returned to the same design principles in the 1990s with Archy, an unfinished software environment that tried once again to realize his humane interface on contemporary hardware. Archy never reached a finished state, but it shows how Raskin’s ideas kept circling back to the same point: computing stripped down to words, presented as simply and predictably as possible.

I’ve always had a real fondness for the Swyft/Cat lineage, and it’s certainly influenced what I think a computer can be. Each one of these attempts embodied humane design: a blank screen for writing, consistent commands, no modes to trip over. The Cat in particular was radical in its way—a computer designed to feel less like a computer and more like a natural extension of the mind. It truly could have changed everything about how we use our computers had it succeeded.

Unfortunately for all of us, by 1987, the market for dedicated word processors was already fading. Canon didn’t seem to know what to do with the Cat—whether to sell it as an office appliance, a PC competitor, or something stranger—and the result was that it fit nowhere. Raskin’s design pushed toward humane simplicity, but Canon’s marketing treated it like just another machine for typing memos.

It isn’t surprising that it failed, though it’s hard not to wonder how it might have landed a few years earlier, when the ground was more open. As it is, the Cat survives less as a commercial product than as an idea in hardware—a glimpse of what a computer could look like if the whole thing were rebuilt around text, consistency, and genuine care for the user.

The Paradox of Openness

The Cat also embodies why Raskin’s philosophy was not necessarily on the same wavelength as some of those other visionary systems. On the surface, the Canon Cat looked open. It booted to a blank screen. Everything was text. You could jump anywhere, edit fluidly, undo anything. Compared to the modal labyrinth of DOS or early Mac software, it felt like freedom.

But look closer and you see the narrowing. The Cat gave you fewer ways to improvise. Its humane design was also constraining design. It reduced your options in order to keep you safe.

The real irony is that the Cat wasn’t even truly closed in the way a smartphone or Chromebook might be considered so today. Underneath, it ran on a Forth environment. You could, if you knew how, drop into Forth and even program directly in 68k assembler. In principle, it was as open as any hacker could want, at least from a software perspective.

The catch was cultural, not technical. From the Ars piece:

IAI’s back door to Forth quietly shipped in every Cat, and the clue was a curious omission in the online help: USE FRONT-ANSWER. This otherwise unexplained and unused key combination was the gateway. If you entered the string Enable Forth Language, highlighted it, and evaluated it with USE FRONT-ANSWER (not CALC; usually Control-Backspace in MAME), you’d get a Forth ok prompt, and the system was now yours. Reset the Cat or type re to return to the editor.

Canon didn’t provide documentation that would have made that power accessible, and Raskin’s design philosophy treated it as outside the normal use case. Extensibility was there if you knew where to look for it, but it wasn’t encouraged. The humane interface was meant to keep most users away from the hood, even though what was under the hood was remarkably open.

That makes the Cat’s paradox sharper: it was a genuinely extensible software environment (up to a point) presented as a sealed appliance. The hardware mostly was a sealed appliance. Contrast this with Emacs or Smalltalk, where openness is the posture of the environment itself. You are expected to extend and reshape as you go, building your tools out of themselves. The Cat offered the same possibility–Forth is a remarkably flexible language, especially for microcomputers–but it discouraged you from taking it.

Humane computing, in Raskin’s hands, edged toward hermetic computing. He built openness in, but sealed it away behind an interface designed to keep it out of sight.

Cul-de-sacs vs. Branches

All of this, to me, is why calling Raskin’s systems thinking a “cul-de-sac” misses the point, and is the wrong way to think about his legacy.

If “cul-de-sac” means “product that didn’t sell,” then sure, the Cat and SwyftCard qualify. They were total dead-ends. But by that same measure, Engelbart’s NLS, Nelson’s Xanadu, Kay’s Smalltalk, or even Lotus Agenda are dead-ends, too. By that measure, most of the “Tools for Thought” tradition didn’t lead anywhere.

The reality is different. These systems were branches. They were rhizomes, in the Deleuze and Guattari sense. They didn’t reach the mainstream, but they seeded ideas that echoed elsewhere, connecting threads that run throughout the history of computing. Hypertext, graphical interfaces, undo, modeless editing—all of these survived in one form or another.

Raskin’s branch is no exception. His machines exposed a fundamental tension inside the tradition: how far do you go in protecting the user from complexity? At what point does “humane” become “hermetic”? Those questions didn’t vanish with the Cat. They’re still with us every time a productivity app promises “simplicity” at the cost of agency.

Raskin’s humane ideals live on in obvious ways, to the benefit of anyone using a graphical computer today—undo everywhere, discoverability, and consistent commands and shortcuts are now interface common sense. But the deeper thread, the ethos that inspired him and others in the tradition of computers as tools for thought, survived mostly outside the mainstream. It persists in systems that never had to sell millions of units or satisfy quarterly targets, that never had to justify their existence to the mass of people using PCs—tools that could afford to remain strange, open, and humane on their own terms. Emacs, Oberon, and Smalltalk belong here, but so do newer experiments like Uxn and 9front.

The Cat failed partly because it tried to straddle two worlds: commercial appliance and humane machine, whereas something like Emacs survives precisely because it never had to. It’s as complex as you want it to be.

This is the sharper point: radical, humane, exploratory computing never survives in the mainstream. The mainstream is built for profit and predictability. Even Engelbart’s work was DARPA-funded, not venture-backed. When you put humane ideals through commercial constraints, they collapse into simplistic appliances, the “For Dummies” version of the original intent. That doesn’t mean the tradition is dead. But it does mean you have to look off to the side, away from the market’s center, to see it alive.

The dilemma(s)

Raskin’s story sharpens two dilemmas that haven’t gone away.

The first is practical: make a system too open, and it risks being overwhelming. Make it too humane, and it risks narrowing into something sealed and hermetic, and not useful enough. The Cat, while also a victim of other factors, tried to balance the two and ended up fitting nowhere.

The lesson isn’t that humane computing is impossible. It’s that humane computing can’t just mean protective computing. It has to mean trusting users with both simplicity and openness. That’s why Org mode and even Mac System 7 endure and the Cat does not.

The deeper implication is harder, but maybe truer: the true Tools for Thought we still wish existed will never come from Apple, Microsoft, Google, OpenAI, or any other large player in the software or hardware space. They can’t. These companies’ scale and incentives point elsewhere—toward lock-in, surveillance, and products that are safe enough to sell but never open enough to empower. The logic of scale makes them constitutionally incapable of building systems that are truly humane and open. The next humane systems, if they arrive, will have to come from outside those walls, as they always have: from margins, from hobbyists, from research labs, and from stubborn communities of practice. But as those platform companies make it more and more difficult to experiment, how do we keep pushing these philosophies forward?

Jef Raskin’s philosophy isn’t a cul-de-sac in computing history. He’s responsible for a branch of the “Tools for Thought” tradition—a branch that shows both the promise and the peril of humane design. His machines make clear how far you can go when you put the human first, and how easily that ideal can collapse into constraint once it’s pushed through commercial channels and turned into walled gardens.

The humane thread survives, but only outside the center—in the tools that don’t have to answer to quarterly earnings, in projects that refuse to die just because they don’t fit the market. The Dormouse lineage isn’t gone. It just doesn’t live where the money is, because it can’t. If you want your computer to be humane in the deeper sense—not an appliance, but an instrument for thought—you have to look to the margins. That’s where it has always been, and where it still is today. If it survives, that’s where it’ll still be.

---

That genre, “photocopied computer manifesto,” is very much the reason this blog exists. ↩︎