I don't like the article very much, especially the "Disadvantages of dynamic typing" section.
- The first point is about performance, which I doubt is the most fundamental difference between static and dynamic typing. The author even comments on whether performance truly matters, except in very low-level tasks, so why list it first?
- The real difference between the two systems, program correctness and early detection of errors, is hidden under the misleading title of "Debugging". It repeats the classic claim from dynamic typing proponents that "runtime type errors are rarely an issue", which is wrong in my experience (I've seen my share of ClassCastExceptions and NullPointerExceptions on production). Also, it advocates doing Unit Testing as a way to catch type errors, which I also find wrong. I'd rather focus on writing unit tests for other types of errors the compiler cannot help me catch.
Other parts also irk me:
- It lists "interactivity" as a strength of dynamic typing, when lots of languages with static typing have useful REPLs.
- The section on Built-in datatypes is baffling.
- Ditto for refactoring. Static typing for refactoring is a strength, not a weakness. About the only valid point here is that sometimes we want to temporarily break static checking of the whole program when testing a small section, but that probably can be handled by testing in isolation.
- The first point is about performance, which I doubt is the most fundamental difference between static and dynamic typing. The author even comments on whether performance truly matters, except in very low-level tasks, so why list it first?
- The real difference between the two systems, program correctness and early detection of errors, is hidden under the misleading title of "Debugging". It repeats the classic claim from dynamic typing proponents that "runtime type errors are rarely an issue", which is wrong in my experience (I've seen my share of ClassCastExceptions and NullPointerExceptions on production). Also, it advocates doing Unit Testing as a way to catch type errors, which I also find wrong. I'd rather focus on writing unit tests for other types of errors the compiler cannot help me catch.
Other parts also irk me:
- It lists "interactivity" as a strength of dynamic typing, when lots of languages with static typing have useful REPLs.
- The section on Built-in datatypes is baffling.
- Ditto for refactoring. Static typing for refactoring is a strength, not a weakness. About the only valid point here is that sometimes we want to temporarily break static checking of the whole program when testing a small section, but that probably can be handled by testing in isolation.