Re:The Ergonomics of Type Checking (2016)
by adrianratnapala on Saturday October 28, 2017 @02:44PM
(#15576210)
The type checker has questions, and it is adamant about them, refusing to let you move forward until they are answered. Often, Iām ok with that. They are good questions, and most of the time they really should be answered. The question is: when?
This is a good point. The poster then proposes being able to nail down your types progressively -- which I think is fine as a long as you don't let the iteration cycle get too long.
Optionally typed languages are one way we approach this ideal. Another way is linters that are stricter than the compiler itself. You must run the compiler to have testable code, but you can run the linter later, perhaps only as part of CI.
It is possible that in future, both of these approaches will look like primitive exteremes of some richer process progressively nailing down the static safety of a program.
This is a good point. The poster then proposes being able to nail down your types progressively -- which I think is fine as a long as you don't let the iteration cycle get too long.
Optionally typed languages are one way we approach this ideal. Another way is linters that are stricter than the compiler itself. You must run the compiler to have testable code, but you can run the linter later, perhaps only as part of CI.
It is possible that in future, both of these approaches will look like primitive exteremes of some richer process progressively nailing down the static safety of a program.