(2021-10-31) Kay on Sutherland, Engelbart, Licklider leading to PARC then Apple

Alan Kay on the history and where it went awry. The three largest intertwined visions of the computer future we were committed to, all burst into view in 1962: Ivan Sutherland's Sketchpad, Douglas Engelbart’s “Augmenting Human Intellect”, and JCR Licklider’s “Man-machine Symbiosis” and “Intergalactic Network” via ARPA funding. This was all the same community.

The basic sentiment at Parc was that we wanted to do “Engelbart, Sutherland, and Licklider for everyone, including children”. A number of systems considerations, especially UI ones, suggested that the larger schemes thought about for the Internetworking of all should also be employed at the personal level (and that this choice would be better and better as Moore’s Law moved along).

The decision at Parc was to build “time machines” that would start appearing in 1973, but would be in the form of networked personal computers of the 80s ca 1986–88 or so.

The catch was that it would be very difficult to do either “Sutherland” or “Engelbart” with these machines.

One of many important features of Engelbart’s system was that any content could be completely shared real-time between any number of users, and all of them could touch and manipulate. This was built into the bottom-most layer of the system and was a wonderful thing. Just how to do this gracefully on a distributed system was still an as yet to be invented scheme.

These considerations upset Engelbart to the point of causing a schism in his lab at SRI. One group stuck with him and implemented their next system on a mainframe time sharing system (as the previous one had been) and this gave them what they were used to for a few more years. The other group — which could see that the problems had to be solved in a distributed way — mostly wound up at Parc.

The silicon inflection point that allowed us to make our time machines also allowed simple inexpensive “8 bit micro” computers to touch the consumer and low end business markets. This started a wave in the late 70s that had pretty much nothing to do with “how computers should be used in society”.

A deep point is that the Parc and Engelbart approaches required more computing power than people were willing to pay for if they didn’t understand what it was for. The weakness of the 8-bit and early 16-bit micros (and lack of microcoding) took programming back about 15 years into rather low level approaches. New programmers starting learning C etc instead of the much higher level languages that were needed.

If you look at this from the “ARPA Dream” perspective, almost nothing from ARPA made it out except for things for which there was no competition. For example, the Ethernet was manifestly better than other attempts at local area networking. The Internet was manifestly better (and pretty much unique).

The Parc work at the end of the 70s lacked a lot of prime needed features that Engelbart had already done. It is doubtful that Apple would have picked these up if they had been there.

Note that Tim Berners-Lee did not know about Engelbart when he patched together what was to become the WWW. He has apologized for this since.