2 ms·
P.S. If Idris (or a language like it) turns out to be significantly valuable in practice, I don't believe it will be thanks to its ability to formally prove de
by pron 9y ago
P.S.
If Idris (or a language like it) turns out to be significantly valuable in practice, I don't believe it will be thanks to its ability to formally prove deep correctness properties, for the reasons I gave above, but because it may turn out that proving some weaker correctness properties (that don't quite reach the bar of formal verification but that exceed what's possible with weak type systems like Haskell's) empirically turns out to yield a significant improvement in correctness for a reasonable amount of effort. I certainly cannot dismiss such a possibility, but I don't think this is something anyone can foresee. It either empirically turns out to be the case or not.