Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I'm not confusing anything. It's relatively common to call the use of dynamically typed languages without static type hints "untyped". I'm pretty sure you knew what I meant and are just trying the classic "you used this word in a way I don't agree with therefore you're wrong".


Untyped has an actual meaning, for example, most assembly languages are untyped, same goes for Forth and probably others.

Then there's dynamic and weakly typed, ie. JavaScript. Dynamic and strongly typed, Ruby. Static and weakly typed, C. Static and strongly typed, Rust.


It was very clear from the context that I was not talking about Forth or Assembly.


Those terms have actual meaning, I've never heard Python or Ruby referred to as untyped. Because they both definitely have types and are in fact strongly typed.

Like, no one would call Ocaml untyped if you used type inference and didn't spell them out yourself... Or if you used the auto keyword everywhere in C++... The Ruby and Python runtimes both definitely enforce types.


> I've never heard Python or Ruby referred to as untyped

Well, now you have! Here are some examples:

https://0xda.de/blog/2024/03/untyped-python-sucks/

https://lucumr.pocoo.org/2023/12/1/the-python-that-was/

In fairness I think it is a fairly recent linguistic shift, probably because there's no debate between weakly typed and strongly typed - strongly typed is oblivious the right answer. Instead the debate is between dynamically-typed-without-static-type-annotations and statically-typed-or-dynamically-typed-with-static-type-annotations, but obviously people need a more succinct term for that.

> Like, no one would call Ocaml untyped if you used type inference and didn't spell them out yourself... Or if you used the auto keyword everywhere in C++... The Ruby and Python runtimes both definitely enforce types.

Static type inference still counts as having static types. That's totally different to dynamic typing without static type annotations ("untyped").


Of course it's Python people who come up with this nonsense...


Do you have a better name suggestion?


They said it in their first comment: dynamic typing


Not a good option - then you'll get a whole different set of pedants saying that Python is always dynamically typed, which I think is a lot more reasonable than quibbling over the use of "untyped" to describe Forth.


Adding type annotations doesn't make Python statically typed. Type annotations are borderline meaningless, they're basically just a type of documentation that helps your tools spit out better suggestions and makes your tools yell at junior developers.

Python isn't taking advantage of type annotations to produce better compiled code or anything, it's still a dynamic language that infers types at runtime.


Yeah that's exactly what I was saying. So what word would you use to mean "dynamically typed without static type annotations"?

> your tools yell at junior developers.

Oh you're one of those...


> one of those...

If by one of those you mean a solopreneur who doesn't have any junior devs to yell at then yes. If I wanted to use static types I'd use an actual statically typed language with all of the performance benefits (Go, Java, Rust, Haskell, Ocaml, etc...).

Using a dynamically typed/interpreted language with type annotations is literally the worst of all worlds...


It's relatively common to call the use of dynamically typed languages without static type hints "untyped".

And it's relatively common for people to be simply wrong, or to be careless in their word choice in regard to lots of other matters.

I'm pretty sure you knew what I meant and are just trying the classic "you used this word in a way I don't agree with therefore you're wrong".

It wasn't, i.e. I wasn't considering the possibility that that was what you meant. This is the first time I've encountered this broken usage of the term.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: