Xerox - the company that fumbled the future

The Palo Alto Research Center was founded in 1970, and by the middle of that decade roughly half of the hundred most highly regarded computer scientists in the world worked there. Within five years it had produced the personal computer, the graphical interface with windows and icons, the mouse driving them, what-you-see-is-what-you-get editing, object-oriented programming, the laser printer, and Ethernet.

The sentence that matters most to this catalogue is Steve Jobs's own, from 1995. He said they showed him three things, and that he was so blinded by the first - the graphical user interface - that he did not take in the other two. One of those was Ethernet. The technology on which this entire section rests was demonstrated to the most consequential visitor in the industry's history, and he walked past it.

Robert Metcalfe, who had built it there, left to found a company for the purpose of selling it. That is the third time this section records the pattern: the relational database was invented at IBM and sold by Oracle, the PC architecture was published by IBM and led by Compaq, and Ethernet was invented at Xerox and made purchasable by 3Com. Inventing a thing and being the company that supplies it are separate problems, and organisations that are excellent at the first are not thereby competent at the second.

The received version of this - that Xerox saw nothing and Apple saw everything - is too neat, and the correction belongs in the record. The laser printer was built at the same laboratory in 1971, in under nine months, and Xerox commercialised it into a staple of every office in the world. The lab that supposedly could not ship anything shipped the most profitable office device of the following two decades.

The company was also, at that moment, defending a photocopier business against low-end competition and answering a government antitrust investigation into its dominance of it. A large firm attending to the market it has, while a laboratory it funds invents the market that will replace it, is the ordinary shape of the problem rather than a failure of vision.

Xerox has since given PARC to SRI International, stating that it wished to concentrate on its core business. Fifty-odd years after the Alto, the same judgement, made again, and stated in the same language.

There is also an espionage chapter, and it deserves more than a footnote. In 1962 the CIA approached Ray Zoppoth, a mechanical engineer at Xerox, with a problem: the Soviet Embassy in Washington admitted almost no Americans, but it admitted the photocopier repairman. A camera was built to sit inside the machine and photograph every page that passed through it. After trials in a disused bowling alley the device went into the embassy's own Xerox 914, and from 1963 a trained repairman swapped the film on his service calls. The agency has never confirmed or denied it.

Two things in that are worth more than the anecdote. First, the reason the repairman had standing access was that the 914 was unreliable and prone to catching fire, so it needed constant servicing. The maintenance channel was the way in - which is precisely the modern supply-chain concern about equipment that accepts updates from its manufacturer, arriving sixty years early and in mechanical form.

Second, the copier changed what theft meant. Before it, taking a document meant removing it, and the gap on the shelf was the evidence. Afterwards the original stayed exactly where it was and the copy left instead. Every subsequent argument about exfiltration - and every control designed to detect data leaving rather than going missing - descends from a machine that made duplication cheap enough to be invisible.