I think a few hundred programmers in the 90s is underestimating the numbers by a factor of at least 1000 and (probably much much) more. Of course such numbers are difficult to estimate, but here some pointers:
It was introduced in school in several countries in the early 80s:
https://en.m.wikipedia.org/wiki/Computer_science_education
And there was vast literature. Petzolds programming windows first edition is from 1988 and second from 1990. While the first might have been a niche publication usually there is only a second edition if there is substantial interest. (number of prints would be nice to know).
You're right. I said "for any interest" but obviously, like all things attention there's a power law distribution. Many Windows programmers, or anything else that had massive interest. But if you were a Unix kernel nerd, there weren't many Unix kernel nerds around. If you were a Perl programmer, there weren't that many Perl programmers.
But even given the prevalence of Windows development, it was still rare to encounter another programmer socially unless you lived in the occasional city where they were common. My point is that nerds finding nerds was a much rarer thing back then.
I'm not sure the OP was making this claim exactly -- I think they just meant for any particular interest (like running Plan 9 on some old hardware) there were probably only a few hundred people around who shared that particular interest.
Could be. But any particular interest is def not true (windows programming), so more like some niche interest. But this is the definition of niche interest, isnt it.
And there was vast literature. Petzolds programming windows first edition is from 1988 and second from 1990. While the first might have been a niche publication usually there is only a second edition if there is substantial interest. (number of prints would be nice to know).