4 ms·
Clojure Inventor Rich Hickey Doesn't Know Types
- comma_at 8y agoCalling out fallacies while commiting more is, well, funny. > Encoding invariants in a type system is far cheaper than writing tests. You may need to back that claim up with something or you're going to be anecdotal as well. > Rejecting an entire field of study (like types!) is not only anti-intellectual, but also does nothing to move the state of our industry forward. Who says type theory is needed to move our industry forward? Even if you think that why would every language need to use it? This is slippery slope. > I’d say that’s especially harmful now in a time where software errors can bring down a couple of Boeing 787 Max airplanes and kill hundreds of people. Was the crash the result of a bug that a type system would have solved? Otherwise this is quite the straw man too.
- manutter51 8y agoHe's quoting Hickey out of context to try and make it look like Hickey is either ignorant or dishonest about what a type spec like "a -> a" means. In context, Hickey is using hyperbole to emphasize how little information is actually conveyed by a type spec, and explain why he decided to leave types out of Clojure. Thomas' rant is grossly unfair to Hickey, and does a great disservice to the programming community, especially considering the explorations Hickey is currently doing with regards to Clojure spec.