SML pops up now and again on HN, which is always nice to see. I wrote a language server for SML in an attempt to improve the tooling situation around the language: https://azdavis.net/posts/millet/
The main motivator for why I did this is because we use SML as a teaching language at my university and students always seem to struggle with the error messages and tooling from the compiler.
I think pure functions, sum/product types, and pattern matching are generally accepted as an excellent way to model and manipulate pure data. I wonder what the team’s thoughts are about handling less pure things like asynchrony and I/O, as well as more interesting control flow like exceptions/panicking, coroutines, generators, iterators, etc.
The biggest reason why you'd want a formal semantics for your PL, as opposed to just a "human" language specification, is so that you can do proofs about the formal semantics. You can prove your PL, as defined, satisfies certain "safety" properties. And the point of doing these proofs is is to check that the definition of the PL "makes sense".
PLs that don't have a formal semantics (aka, most mainstream PLs) are more likely to be in situations where implementors realize, after the standard has been written, that e.g. some of the PL's features interact in a way that doesn't make sense, such as this example with C++ coroutines and lambdas: https://news.ycombinator.com/item?id=33084431
Two of the most common safety properties of interest are progress and preservation. (I touched on this in the last above linked post near the bottom.) At a high level:
1. Progress states that if you have a program that type-checks, then either that program is "done" or it can continue evaluating.
2. Preservation states that if you have a program that type-checks and that can continue evaluating, then as it continues to evaluate, it continues to type-check.
Note that the conclusion of preservation "feeds back" into progress: the program type-checks. And vice versa: progress may state as its conclusion that the program can continue evaluating, which then lets you apply preservation. This means you can keep applying the progress and preservation theorems in a "loop" until the program is done evaluating.
For each of the 4 posts in my series about formal semantics, I duly translated the rules presented in the blog post into Lean code, and then proved that the rules do satisfy the safety properties. For example, for the first post linked above:
For my part, I think it’s unfortunate that virtually no mainstream language has a formal semantics in the style of SML’s Definition. I wrote a bit about formal programming language semantics in a series of posts: https://azdavis.net/posts/define-pl-01/
Hoping Millet doesn't mean anything offensive in any language… Studio Ghibli had this problem mildly with their film "Laputa" ("la puta" = "the whore" in Spanish).
I note in the caveats that the current approach is to recompute everything whenever even one file is changed. But that’s probably not sustainable, as I admit myself in https://azdavis.net/posts/pl-idea-tooling/ :
> As an example, it would be unsustainable if, every time we changed a single function’s body, the language server had to re-typecheck the entire codebase. This might work for an initial proof-of-concept on a small codebase, but for large ones, the responsiveness of the language server would drop precipitously.