As I mentioned there's no requirement for typed languages for doing anything. Nothing about OO or any other language really requires the typings per se. But they let the compiler help you find trivial issues that are otherwise harder to track down. There's no reason that you should have to have code fail for a type in a var name, but if you don't have a typed language there's no compiler to help you with that for example.

As I said - it's a choice.

+++ Rick ---

>>Whether you use a statically typed language or a dynamic is a choice. The best choices IMHO tend to be hybrid languages that support both.
>>
>>Personally, I prefer typed languages because they do offer better support for tooling - refactoring, Intellisense, rule based code validation are much easier to implement for tool vendors if you have a type system and language engine in place. There's nothing amateurish about this BTW. These are professional tools, that if you've only used VFP to compare against, you probably can't even imagine.
>>
>>If it's easier for tool vendors they will build those tools, otherwise - not so much. You often see that co-relation when you compare the tooling between statically typed languages and dynamic languages. It's possible to do similar tooling for dynamic languages but it's a heck of a lot harder and often inaccurate. A good case in point are the various JavaScript environments that existed in the pre-Typescript days that would often give iffy results because they had to guess where you had to manually provide hints to the system to do what a typed system can automatically figure out because of the typed metadata.
>
>I guess those tools do help a lot, but OTOH I'll keep a grain of doubt - some of these tools may be needed precisely because of the type restrictions and help resolve them. Since I don't really need any of that (my other languages are js and minimal bits of SQL), I declare myself a non-expert on the matter.
>
>What I had in mind is the situation where, for example, http://beginnersbook.com/2013/03/polymorphism-in-java/ defines overloading of methods just as something you have to do in a typed language - a different version of a method for each applicable set of parameter types. Or all the examples of class factories where you had to create a separate factory for each set of classes, because the class name has to be compiled and can't be ad-hoc taken from external code etc. So I guess many things are done differently because they have to, and I'm not convinced that each of those is better and makes for a more productive work.