Hey that was kind of my point -- but I think you have it in reverse, Haskell has had a lot of this stuff for a long time (as in most of them since it's inception), and it's trickling down to other languages now.
But to make some concrete counter points:
- Apache Spark is not a general purpose programing language (you're totally right about FP and lazy evaluation being important in DAG-land of course)
- "Non-nullable types can be found in Java", yeah except them being the default is the big innovation, along with the recognition of the problem, and facilitation of the worldview that recognizes the issue. Optional didn't show up until a few years ago (Java 8?), first class functions weren't a thing without subclassing till around then too, Function references, Functional interfaces, etc. I'm less familiar with C++ and it's commendable that it's adopting new things and people are moving forward, but it's basically the gold standard of footguns with type-system scopes attached (again I don't write C++ on a daily basis and haven't felt just how much better the new editions are).
- Go and Rust learned from Haskell, Rust heavily so. BTW these days I'm more and more of the opinion that Rust is the one more worth praising of the two
- What go does is kind of error as values, but it's also kind of not -- I mean a near complete lack of use of exceptions at all. The distinction is subtle, but coding to always handle the error case (because it is the result) is different from having a sometimes-present error code that you sometimes check.
- Again, you're right that Monads are everywhere -- Haskell didn't invent the concept, but it is one of the places you can go to see it actually used functionally and learn from what people are doing with it (never mind all the novel papers).
Haskell is one of the few places that all these features come together to form a coherent whole.
But to make some concrete counter points:
- Apache Spark is not a general purpose programing language (you're totally right about FP and lazy evaluation being important in DAG-land of course)
- "Non-nullable types can be found in Java", yeah except them being the default is the big innovation, along with the recognition of the problem, and facilitation of the worldview that recognizes the issue. Optional didn't show up until a few years ago (Java 8?), first class functions weren't a thing without subclassing till around then too, Function references, Functional interfaces, etc. I'm less familiar with C++ and it's commendable that it's adopting new things and people are moving forward, but it's basically the gold standard of footguns with type-system scopes attached (again I don't write C++ on a daily basis and haven't felt just how much better the new editions are).
- Go and Rust learned from Haskell, Rust heavily so. BTW these days I'm more and more of the opinion that Rust is the one more worth praising of the two
- What go does is kind of error as values, but it's also kind of not -- I mean a near complete lack of use of exceptions at all. The distinction is subtle, but coding to always handle the error case (because it is the result) is different from having a sometimes-present error code that you sometimes check.
- Again, you're right that Monads are everywhere -- Haskell didn't invent the concept, but it is one of the places you can go to see it actually used functionally and learn from what people are doing with it (never mind all the novel papers).
Haskell is one of the few places that all these features come together to form a coherent whole.