Morphine
This page is about Morphine, a typed functional programming language. Everything's more (fun x -> x) with Morphine.

Morphine is a simple, functional language. It has recursive first-class functions, recursive algebraic data types, let, numbers, symbols, lists, tuples, and a small set of built-in functions.

Morphine was previously known as LamLang2. It is named after the opiate, which in turn is named after Morpheus, the Greek god of dreams (pictured above).

I'll write a more comprehensive introduction one of these days.

• 2008-01-02–The syntax has been significantly extended to allow Haskell-style multi-equation definitions; it also now has a simple implementation of sections. I'm experimenting with allowing infix application of data constructors. I've switched the layout preprocessing engine to a slightly-modified version of the newer one I wrote for Braque. Something's the matter with the tuple / unit rule in the type system, but it shouldn't be too bad to fix (once I actually get round to it). Top-level definitions can be (mutually) recursive, but I still need some sort of "let rec". Following a type inferencer I'd done for Haskell '98, I didn't think to allow scoped "forall" quantifiers (e.g., the type of omega: "(forall a, a -> a) -> forall a, a -> a"); I should fix that, too.
• 2007-11-20–The type system now handles tuples, case, and patterns, and the parser is a bit more sensible. The big thing missing is some sort of Y / fixpoint ... which severely limits how much fun the language currently is.
• 2007-10-25–I've been working a bit on Morphine's type system and parser. It doesn't do much yet. I've also decided that Morphine will be the basis for Oasis, another language I'm working on. More information on that eventually.
• 2007-05-14–I declare LamLang dead and have started Morphine. It is going to be very similar to the first implementation of the language, but is rather a bit more sensible. Finally, lists can be pattern matched. Why? because the list constructor is simply a constructor like any other. "forall" can be written in type annotations. Why? Because the (currently-unïmplemented) type system is going to properly handle universally-quantified types. It will be much easier to extend the set of built-in functions / primitives, and the user will be able to shadow / redefine primitives. Why? Because primitives are simply default bindings in the environment rather than a hard-coded fixed set of procedures. Layout will work. Why? Because I'm taking the code directly from another project of mine, code that I have every reason to believe works. Stay tuned!

• (Morphine is not currently available; eventually, it will be released under some sort of learning-centric license.)

Well, there's really no more information available for the time being, sorry! Please check back soon, though, or e-mail me if you have specific questions.