The ideas presented in "Hints and Principles for Computer System Design" are timeless wisdom for systems programmers. They're relatively simple, broadly applicable, and can easily have outsized importance.
Conversation
Replying to
The idea that first order logic is a good language for specification of computer systems is not supported by any evidence.
1
Replying to
I don't think that anybody claimed otherwise. For any definition of "good language". The paper appeals to me as an intuitive thinker. There are probably smart and capable people that would dismiss all this. That doesn't concern me in the slightest - take it or leave it.
1
Replying to
We have to take it or leave it? We can't agree with parts, disagree with others, and think about still others? The original paper was brilliant. The rewrite is something I'm thinking about - but the assumption that 1st order logic is appropriate struck me as wrong.
1
Replying to
I haven't read the original, so cannot comment on that. Your general remarks seem reasonable to me, though. I myself have little patience for monad people (or anybody else whose thinking is 100% top-down). And yet I think their core arguments are definitely worth understanding.
Replying to
Thanks! Your views on C somehow seem relevant. The C standards committee is hell bent on making C more consistent in theory by making it inconsistent with reality. So clearly it's a mistake to dismiss empirical reality. Still, let's not make the opposite mistake somewhere else.
1
Show replies