Slashdot Mirror
Compiling to JavaScript: TypeScript vs. Haxe
lars_doucet writes: Released in 2012, Microsoft's TypeScript is perhaps the best-known "compile to JS" language, but it wasn't the first. One of the earliest was Haxe, whose JS target first appeared in 2006. In his illuminating article, TypeScript vs Haxe, Andy Li gives an excellent rundown of the two languages' various merits, but the bottom line is: "Existing JS developers will favor TypeScript as they are more similar in many ways. They can utilize their existing skills immediately. Non-JS developers with backgrounds like Java/C# or even from the functional programming world will appreciate Haxe more since it fixes a lot of weirdness of JS." The full article includes an excellent rundown of the type systems, syntax, scope handling, compilers, and overall language design philosophy.
How is that web assembly project coming along? It seems like a perfect fit for alternative languages like this instead of having to compile to JS.
WebAssembly is not currently a suitable target for Javascript compilation, because it lacks a garbage collector. Only languages that can base themselves on (the equivalent of) malloc/free can currently target wasm. Eventually garbage collection facilities will be provided, but they're not on the immediate drawing board, and even once garbage collection is possible, it's likely to be a long time before JS->wasm approaches the performance of v8 or similar. (Note that asm.js doesn't do garbage collection either.)
The nice things about wasm are (1) it gives the creators a chance to re-invent the web runtime from the ground up (e.g. native threads will be there almost from the start), (2) languages can evolve independently of the web platform, and the web platform becomes language agnostic, and (3) *all* the major players are fully on-board.
One of the nicest features of wasm, in my opinion, is that the intermediate representation is an AST of expressions to be computed. This is a far more sensible and flexible IR than some stack-based or register-based VM bytecode system. My prediction is that wasm will become the preferred compilation target for desktop/server applications, not just web applications: the same toolchain will be used to generate binaries for running in server containers as well as within a browser.