Last Friday we wondered what Alan Kay, a former Walt Disney software developer and ex-Apple employee, would do after recently leaving Disney.
Kay is one of the inventors of the Smalltalk programming language. He conceived the laptop computer and was an architect of the modern graphical user interface. Whatever he decides to do next could have major effects on the face of technology.
Kay wouldn’t tell MacCentral exactly what he was planning, but he did give us hints as to what probably wouldn’t occur. In our previous story, we guessed that perhaps he would resume work on “agent based systems” (agents are kernels of intelligence in a computer).
“I have done nothing about agents for years,” Kay told MacCentral. “We have been thinking of making a really comprehensive tutoring agent for certain important ideas (the tradeoff being that it would be very very smart about just a few things, but would not be generally considered to be intelligent).”
What of the Dynabook, his vision of “a portable interactive personal computer, as accessible as a book” and sporting a flat panel display and wireless communications? He once wanted to mass-produce such a product.
In 1993, Kay’s vision of Dynabook finally transformed itself into reality in the form of Apple’s late, lamented (by many) Newton.
“However, what was released was a far cry from Kay’s dream,” Brian Rampersad, computer programmer and systems analyst, wrote. “It transpired into a majorly flawed product and a public relations nightmare for Apple. The Newton was to be the world’s first PDA or ‘Personal Digital Assistant.’ However, in 1993, Apple only sold 80,000 units.”
“The closest thing to a Dynabook right now is the Microsoft Tablet — done by Chuck Thacker and Butler Lampson, two of the principals at PARC in the 70s,” Kay told MacCentral. “However, they, too, made the mistake of leaving off the keyboard. We knew back in 1968, via the first great character recognition system, GRAIL done at RAND in 1966, and better than Graffiti, that even a perfect and instant recognizer would not do the job. The recognizer would be for controls, fixing typos, and short ‘fill in the blank’ type stuff, and you would want a keyboard for paragraph length typing. No one has made a commercial device yet with the particular combination that seems to be needed.”