In December 1979 a team from Apple Computer arrived at the Xerox Research Center in Palo Alto not as spies, but as buyers of stock. Xerox had invested one million dollars in Apple's upcoming IPO, and the price included two demonstrations of technologies that PARC researchers had been working on for nearly ten years. What happened over those two days became one of Silicon Valley's most mythologized stories—the story of how Steve Jobs "stole the future" from a corporation too blind to see it. But reality turned out to be more complex: it was a meeting of two engineering cultures, each solving its own problem, and neither knowing that ten years later a federal judge would be sorting out who owned the right to click a mouse on a trash icon.
The Xerox Research Center opened in 1970 with the task of building the "office of the future"—a system where documents are created, edited, and printed within a unified electronic environment. The corporation that had made its fortune on copiers understood: paper would sooner or later give way to digital workstations, and they needed to invent those workstations before competitors did. By the end of the seventies PARC had assembled a collection of technologies, each of which looked like a fragment from another era.
Xerox Alto—a personal computer the size of a refrigerator, which appeared in 1973. A black-and-white raster display with 606×808 pixel resolution, oriented vertically like a sheet of paper. Three mouse buttons. Overlapping windows that could be moved, resized, minimized. File and folder icons. The WYSIWYG editor Bravo, where text on screen looked exactly like the printout. All of this worked not in command-line mode, but through direct manipulation of objects—click an icon, the document opens.
Ethernet—a local network invented by Bob Metcalfe in 1973, linking Altos into a unified system. Computers exchanged files, sent print jobs, shared resources. Speed 2.94 megabits per second—by the standards of an era when modems worked at 300 baud, this was a highway.
Dover laser printer—a device that printed graphics with the same quality as text. Not matrix dots, not thermal printing—full layout at 384 dots per inch resolution. A document created in Bravo came out of the printer ready for publication.
Smalltalk—an object-oriented programming environment where the entire interface was built from objects that exchanged messages. Windows, menus, buttons, text fields—all objects. The programmer doesn't write separate functions for rendering and event handling—they create an object that knows how to draw itself and respond to mouse clicks.
All of this worked not in a lab on mockups, but in a real office. PARC engineers used Alto every day, edited documents, sent mail, programmed in Smalltalk. But beyond the building almost no one saw these technologies. Alto wasn't sold—it was a research prototype, produced in a run of about two thousand machines, most of which ended up in universities and inside Xerox itself.
The PARC team didn't create the graphical interface from a blank slate. The mouse was invented by Douglas Engelbart at the Stanford Research Institute in 1964, and he first showed it in 1968 at the legendary "Mother of All Demos." Engelbart presented the oN-Line System (NLS)—a hypertext editor with windows, collaborative editing, and videoconferencing. It was controlled by a mouse and chord keyboard: one hand moved the cursor, the other entered commands through key combinations. This was an attempt to radically accelerate work with information, but NLS required retraining—users had to memorize dozens of key combinations and keep the tree structure of documents in their head.
PARC engineers, many of whom had seen Engelbart's demonstration, went in a different direction. Instead of an "intelligence amplifier" for trained operators, they were building a system for ordinary office workers that should be understandable without training. They simplified the mouse to three buttons. They threw out the chord keyboard. They hid commands in menus that opened with a click. Instead of text file descriptors, icons appeared—pictures recognizable at first glance.
Object-oriented programming didn't come from PARC either. Its foundations were laid by Kristen Nygaard and Ole-Johan Dahl in the language Simula 67, created in Norway in 1967. Simula described a program as a set of objects with their own data and methods—a concept that Alan Kay and his team developed in Smalltalk. Kay wanted to create a language in which children could program as naturally as they draw—building worlds from objects that communicate with each other. Smalltalk became not just a language, but an environment: code, compiler, debugger, interface—everything existed inside one object system where each element could be opened and rewritten.
PARC assembled these lines into a unified construction. Mouse, windows, icons, object-orientation, local network, laser printing—technologies born in different places came together into an ecosystem where a document traveled from concept to typographic print within a single machine. But assembling this in a lab turned out to be easier than turning it into a product.
Xerox didn't ignore its research center's inventions. In 1981 the company released Xerox Star (8010 Information System)—the first commercial workstation with a full graphical interface. Star wasn't a prototype, but a product: a complete system with windows, icons, mouse, integrated applications for text, graphics, email, and spreadsheets. A document created in Star looked on screen exactly as it did on a printout from the laser printer. Multiple stations were linked through Ethernet into an office network where files and printers were shared resources.
The price of one station—$16,595 ($59,000 in today's dollars). This didn't include the cost of file servers, network printers, and service. A full office setup for a small company ran into hundreds of thousands of dollars. Star was positioned not as a personal computer, but as a corporate solution for departments where secretaries, managers, and engineers worked with documents inside a closed Xerox ecosystem.
The problem wasn't in the technologies, but in the business model. Xerox sold not hardware, but "office systems"—comprehensive solutions with training, integration, service maintenance. Star was sold through Xerox's corporate channels, where managers were used to buying copiers and printers, not computers. The company didn't know how to compete with IBM in the computing market and didn't want to compete with Apple in the personal computer market.
Moreover, Star's architecture was closed. You couldn't write your own application or connect a printer from another manufacturer. The entire ecosystem was controlled by Xerox—a reasonable strategy for a corporate product, but fatal for the mass market, where the winner is the platform with the widest circle of developers and compatible devices. Throughout its production run—until discontinuation in 1985—Xerox sold about 25,000 Star stations—a number that looks like success for a niche corporate product, but failure for a technology claiming the role of industry standard.
Star didn't disappear without a trace. Many of its engineering solutions—file management through icons, drag-and-drop, universal Copy/Paste commands, network printing—became industry standards. But the platform itself was displaced by cheaper and more open systems that were built on PARC ideas but solved a different problem: not "office of the future" for corporations, but a personal computer for everyone.
Apple's first visit to PARC took place in November 1979, the second in December. The terms were tough: Xerox invests one million dollars in Apple's upcoming IPO at $10 per share (at IPO the shares went out at $22), and in return gets two demonstrations of research projects. PARC engineers didn't want to show their developments to potential competitors, but Xerox management saw Apple as a growing company that could become an ally or client. Corporate managers didn't believe that a handful of hippies from Cupertino could turn expensive lab prototypes into a commercial product.
The first demonstration disappointed Jobs—researchers showed projects far from completion. Jobs insisted on a second visit with deeper immersion. In December PARC engineers demonstrated Smalltalk and Alto in full power: overlapping windows, menus, icons, WYSIWYG editor, object-oriented programming. Larry Tesler, one of Smalltalk's lead developers, spent several hours explaining the system architecture. Bill Atkinson, an Apple engineer, later recalled that what he saw upended his understanding of what a computer should be.
But Apple didn't simply copy Alto. Jobs's team had been working on the Lisa and Macintosh projects since 1978, before the PARC visit. Lisa was conceived as a powerful workstation with a graphical interface, Macintosh as an affordable computer for the mass market. PARC gave Apple engineers not ready-made solutions, but confirmation of the right direction and a set of ideas that needed to be reworked for different constraints.
Atkinson created his own algorithm for drawing overlapping windows—faster and more efficient than Alto's. Instead of Xerox's three-button mouse, Apple chose a single-button one—cheaper to manufacture and simpler to use. Menus moved from the window body to the top line of the screen, which remained always visible. The Trash icon appeared—a metaphor that wasn't in Star. Double-clicking to open files, dragging icons, selecting text by dragging the mouse—all of this was developed at Apple as solutions to specific interface problems, not copying others' work.
Lisa came out in 1983 at a price of $9,995—cheaper than Star, but still too expensive for the mass market. Sales flopped. Macintosh, released in 1984 for $2,495, became the first affordable computer with GUI—not because Apple stole Xerox technologies, but because Cupertino engineers managed to fit a graphical interface into a machine with 128 kilobytes of memory and a Motorola 68000 processor at 7.8 megahertz. Alto required several times more resources. Macintosh worked at the edge of hardware capabilities, using every byte of memory and every processor cycle, but it worked and cost six times less than Star.
In 1989 Xerox sued Apple, accusing the company of copyright infringement on the graphical interface. By that time Apple itself was suing Microsoft and Hewlett-Packard, claiming that Windows and NewWave copied the Macintosh interface. Xerox demanded recognition that its technologies had been stolen without compensation, despite the 1979 deal.
Federal court dismissed the suit—without even delving into who took what from whom. In March 1990 the district judge ruled that Xerox hadn't shown "reasonable grounds" to fear counter-claims from Apple, and therefore had no right to demand annulment of its copyright registrations; canceling others' copyrights by law is generally impossible for anyone. "Federal law gives no copyright holder remedies for commercial failure," the judge summarized, also reproaching the plaintiff for seeking "verdict first, evidence later." And if Apple did copy what Xerox created, he concluded, the road lies in an ordinary copyright infringement suit—which the corporation never filed.
The principle often attributed to this case was actually formulated in the neighboring Apple v. Microsoft litigation: interface ideas aren't protected by copyright the way code or design are. You can't patent the concept of "clicking a mouse on an icon to open a file"—what's protected is the specific implementation, not the metaphor itself. While Xerox tried through court to cancel others' registrations, Microsoft calmly continued developing Windows.
Xerox's lawsuit became the corporation's last attempt to claim rights to the graphical interface. By that time Star had been discontinued, and Xerox itself had returned to its core business—copiers and printers, where it earned stable billions. The graphical interface became common property of the industry, and this lawsuit's failure clearly showed: no company can monopolize the way humans interact with computers.
Counterfactual history is a dangerous game, but here we can outline the boundaries of the probable. The graphical interface wasn't born at the moment of Apple's visit to PARC. It was developing in parallel in several places: MIT, SRI, PARC, Bell Labs. By the early eighties the idea of controlling a computer through windows and icons had ceased to be exotic—it was in the air of research centers. Xerox Star proved that GUI could be brought to a commercial product. Apollo Computer and Sun Microsystems were building graphical workstations for engineers. Commodore Amiga and Atari ST brought GUI to home computers. The question was not whether the windowed interface would appear, but when it would become the mass standard.
Without the PARC visit, the Lisa and Macintosh projects would have developed more slowly. Apple engineers were moving toward a graphical interface, but the Alto demonstration gave them a ready reference model—not for copying, but for understanding how a completed system should look. Perhaps Lisa would have come out in 1984 instead of 1983, and Macintosh in 1985 or 1986. Perhaps Apple would have spent more time experimenting with the interface, making mistakes that were avoided thanks to PARC's experience.
But other companies were moving in the same direction. Microsoft began developing Windows in 1983, guided by Xerox Star, Apple Lisa, and its own research. The first version of Windows came out in 1985—primitive, slow, but working. By the early nineties Windows became the dominant platform not because Microsoft stole the interface from Apple, but because Windows worked on cheap IBM-compatible computers, of which there were tens of times more than Macintoshes.
If Apple hadn't seen the PARC demonstration in 1979, the graphical interface would have become the standard anyway—but perhaps a year or two later and with a different set of basic metaphors. Perhaps the model with multiple mouse buttons, like Xerox's, would have won. Perhaps menus would have remained tied to windows rather than the top line of the screen. Perhaps the trash icon would have looked different. But the idea of controlling a computer through direct manipulation of objects on screen was too powerful to stop. PARC created a working prototype, Apple turned it into a product, Microsoft made it a standard. Each company solved its own problem, and none could claim sole authorship. In engineering there's rarely one inventor—more often it's a chain of improvements, where each link is critically important, but none is self-sufficient.