Re:On the cruelty of really teaching computer science (1988)
by a1369209993 on Wednesday July 29, 2015 @03:53AM
(#9966931)
If you don't have a formal functional specification that adequately characterizes the program's behavior, then Dijkstra's approach is useless.
If you do have a formal functional specification that adequately characterizes the program's behavior, then you have already have source code, and Dijkstra's approach is irrelevant.
[ Parent
| on HN ]
Re:On the cruelty of really teaching computer science (1988)
by AnimalMuppet on Wednesday July 29, 2015 @09:56AM
(#9968624)
Or at least something formally equivalent to source code, even if there is not a "compiler" for it. But then the programmer becomes merely a human compiler, doing a provably-formally-equivalent transform to a formal specification.
Even if Dijkstra's approach is not irrelevant, it is inhuman - it makes the programmer into a machine. (Note that the first "computers" were human machines.)