Would there by any interest in an imperative language with a basic Standard ML type system?
Personally, while I find the imperative paradigm more intuitive than the functional paradigm, I believe functional type systems are simpler and more "right" than imperative ones are (which have been taken over by OOP). There are multi-paradigm languages, but these have large numbers of features, and complicated type systems, and I want something simple. Is there interest in this? Thanks.
By SML type system, I'm thinking:
- Algebraic data types
- Anonymous functions
- Pattern matching
- Simple generics
- Type inference
- Simple modules and implementation hiding
combined with some approach to operator overloading (which SML doesn't consider at all). And to remind you, the paradigm should be imperative, not functional.
I'm thinking a use case might be in numerical linear algebra, but with support for multiple different scalar types including floats, ints, complex floats, dual numbers (for autodiff), "codual numbers" (for better autodiff), bignums, quaternions, symbolic algebra, etc. The functional paradigm is not suitable here, but Julia is perhaps overcomplicated.
OCaml is pretty close already to what you're describing. The OCaml ecosystem fully embraces imperative programming. And if the syntax of OCaml is problematic, then perhaps ReasonML (https://reasonml.github.io/) is what you're looking for: a curly-braces syntax for OCaml.
OCaml and Rust both offer these, OCaml does overloading mostly with module functors and Rust with traits. Of course they’re both multi-paradigm, OCaml is also really good for functional programming.
Personally, while I find the imperative paradigm more intuitive than the functional paradigm, I believe functional type systems are simpler and more "right" than imperative ones are (which have been taken over by OOP). There are multi-paradigm languages, but these have large numbers of features, and complicated type systems, and I want something simple. Is there interest in this? Thanks.
By SML type system, I'm thinking:
combined with some approach to operator overloading (which SML doesn't consider at all). And to remind you, the paradigm should be imperative, not functional.I'm thinking a use case might be in numerical linear algebra, but with support for multiple different scalar types including floats, ints, complex floats, dual numbers (for autodiff), "codual numbers" (for better autodiff), bignums, quaternions, symbolic algebra, etc. The functional paradigm is not suitable here, but Julia is perhaps overcomplicated.