Re: The most useful form of typing is...

[nothingmuch]

in reply to The most useful form of typing is...

Is everyone's opinion on what the difference between strong and dynamic typing in sync?

I have the impression that static is that it's known in advance, and strong is that it's enforced. Usually strong typing is a prerequisite to static typing, but does not imply when the type check is made.

Does anyone have a better definition?

And what's the diff between weak and dynamic?

[Errto]

I like your definitions, so I'll try to put together a short list of some of the languages I know and where they fit in the scale.
• Strong and static: Haskell, ML, PL/SQL, Java (but only if they implemented generics properly so you wouldn't need typecasting)
• Strong and dynamic: Perl, Lisp, Scheme
• Weak and static: C, C++, Pascal, Java (as is)
• Weak and dynamic: Javascript, VB, bash

The reason (as I understand it) that it is possible to have weakly-typed but also statically-typed languages is because of typecasts: you are essentially instructing the compiler to ignore the type declarations you've already given it. Also note that the first two languages I listed under Strong and Static allow total type polymorphism so that the same functions and data structures can apply to arbitrarily many types, but compile-time resolution still has to be possible.

[hardburn]

Your definitions are close to what is in the litature. Really, "strong" and "static" are orthoginal. Being static means that your language needs to have some means of determining the types at compile time, while strong means those types are difficult to change during execution (runtime). Pascal/C/etc. determine the types via special syntax (int foo, bar; and such), while more modern Hindley-Milner type systems can figure it out based on what you do with the variable the first time you use it.