I think the problem is your sample environment.
I think my sample is good - back then home computers were mostly
purchased by nerdy people who would likely go to an engineering school
(or send their kid there). Home computers were far from being the
ubiquitous piece of home electronics they are now.
The C64 still sold 80,000
units in 1989, building on the already huge extant userbase.
80000 was a not much, even back then. Many of the cheap XT clone
companies were likely doing those numbers or better.
Engineering was never its strong suit
even when it was an even larger market force earlier because it just didn't
have the display or computational chops. But it still was huge as a
general-purpose home computer.
Back then, no engineering school I know of had students do much or any
of the engineering on their home computers, so the power and display
argument is flawed. It was still a traditional (mostly DEC)
minicomputer or mainframe approach - we had a motley crew of machines,
mostly VAX, probably similar to other engineering schools. The home
computers were for writing term papers and reports, and screwing off,
of course. Only at the very end were we allowed to write code on our
PeeCees - and we were the first class to do so.
--
Will