Earlier quoted context omitted.
The best book about the ARPA/Parc research community (Parc sprouted from ARPA) is "The Dream Machine" by Mitchell Waldrop: it is both the most complete and most accurate. "Dealers of Lightning" is at the next level but far from the bottom. Its flaws are too much "Heroes' Journey" and a very complex and confused jumping around timeline (I had trouble myself orienting in some of the spots). But it also has a lot of goo…
Thanks! I'll add that one onto my reading queue. I actually didn't have too much trouble following Dealers, because it (more or less) mostly followed each project separately, which while creating an interspersed timeline in the book, was mostly coherent within each section. I did read at least part of Fumbling, but I found it a hard book to thread the needle on, and it was in such stark factual disagreement with what…
There are at least two big issues regarding cost that many people miss: (a) the first is the difference between what should be spent on "prototypes for learning and vetting" and what can be done when designing for manufacturing, and (b) the second is the the difficulty most people had with valuing what personal computing might be for them.
In the first public paper I wrote about the Dynabook I pointed out that Moore's Law meant that powerful tablet sized personal computers would likely wind up costing what a color TV set would cost (they would have pretty similar components, and most of the cost in electronics is in packaging).
But we also had another analogy that we though could work via education: that of the personal automobile in the US. People value cars enough to be willing to pay quite a bit more for them than for most consumer devices. This was very interesting because the ARPA dream of an interactive personal computer connected to a world-wide network was a kind of "information and intelligence vehicle".
If people could see this, then they might be willing to pay what they would pay for a car. Certainly most computer people and most scientists and engineers would be able to assign value in this way. We thought most knowledge workers would eventually be able to see this also, and that there would be an intermediate phase before getting to the TV set kind of technologies.
An analysis of what happened to eventually quash this idea is beyond the scope of this note. (But, to make a point in talks, I've tried to get people to think about what "a car's worth" of personal computer could be like (the average car in the US a few years ago was a Toyota Avalon at $28K, so about 10 times what most personal computers go for).
This is a different slant than the problem that DEC and similar companies had (which was to not be able to understand personal computing in any reasonable form).