The Pocket Computer, Fifteen Years Early

The world it was born into

In the late 1980s personal computers sat on desks. Laptops were heavy and expensive. Mobile phones were large, used for voice and owned mostly by business people. Wireless data networks for consumers barely existed. Yet several teams, many with roots at Apple, believed that computing would soon move into the pocket, and that the pocket computer would be connected, personal and operated by a pen or a finger rather than a keyboard.

The bet

Two efforts carried the bet most clearly. Inside Apple, a project begun in 1987 became the Newton. Apple’s chief executive introduced the phrase “personal digital assistant” in 1992, and the first Newton MessagePad went on sale in August 1993 for about $700. It promised to recognise natural handwriting, organise notes, contacts and calendars, and eventually communicate wirelessly.

In 1990 a group of former Apple engineers founded General Magic, with backing and partnerships from Sony, Motorola, Philips, AT&T, Matsushita and others. General Magic built Magic Cap, an operating system with a friendly interface built from everyday objects, a desk, a hallway, rooms, and Telescript, a language for software “agents” that could travel across networks to act on the user’s behalf. Sony’s Magic Link and Motorola’s Envoy shipped in 1994.

The vision behind both was very close to what the smartphone became: a personal, connected device that managed communication, information and errands.

Why it made sense

The vision was correct. The talent was exceptional; General Magic’s alumni later led some of the most important mobile products of the next two decades. The partners were the companies best placed to make the hardware and networks. And the trends that eventually made smartphones possible, smaller processors, better batteries, wireless data, were real and visible.

How it lost

Every part of the experience depended on technology that was not ready. Handwriting recognition on the first Newton made frequent errors, and the errors became a public joke, lampooned in a Doonesbury comic strip and on The Simpsons. Later versions improved substantially, but first impressions stuck. Wireless data was slow, expensive and patchy. Batteries were limited. Prices were high for devices that did less than their promise.

In 1996 Palm launched the Pilot, a small, cheap device that did a few things, contacts, calendar and notes, reliably, synchronised with a PC and asked users to learn a simplified alphabet called Graffiti instead of recognising their natural handwriting. It sold in large numbers. Apple discontinued the Newton in 1998, as it narrowed its product line during a financial crisis. General Magic’s devices sold poorly; the company went public in 1995, later pivoted to voice services, and wound down by 2002.

Explaining both the same way

Palm’s success is usually explained by simplicity and the Newton’s failure by over-ambition. A symmetrical account notices that Palm made a specific trade. It moved the hardest problem, recognising handwriting, from the machine to the user, by asking people to learn Graffiti. It dropped wireless communication at first, relying on a cradle connected to a PC. It accepted a narrower definition of the product. Each of these trades made the device work reliably with the technology of 1996. The Newton and Magic Cap refused those trades, because their designers wanted the full vision. In 1993–94, the full vision could not be built well enough.

What the loss teaches

This is the clearest case on the site of moral luck in technology. The insight was right and the outcome was failure, because the enabling technologies arrived a decade later. Judged only by results, the projects were mistakes. Judged by what could be known, their core insight was one of the best of their era. Their main error was a calibration error about timing: they planned as if several hard, unproven technologies would mature together on schedule.

The lesson for decision-makers is not to avoid big visions. It is to separate the vision from the vehicle. Ask which parts of the product must work perfectly on the first day, and whether the technology can deliver them. If it cannot, find a smaller product that delivers part of the vision now, as Palm did, and keep the larger vision for later, as Apple eventually did with the iPhone.

Sources and limitations

This account draws on contemporary press coverage, the documentary General Magic (2018), Butter and Pogue’s history of Palm and Computer History Museum materials. Price and date details are approximate where sources differ. Statements about the influence of alumni describe their later roles, not claims about the precise transfer of ideas. The judgement that the Newton was launched before its core input method was reliable is a widely shared assessment, but participants have defended the decision.