3 ms·
JVM bytecode semantics were all about Java - it forces the Java object model on you, for example. It was never designed to allow a diverse ecosystem of language
by int_19h 7y ago
JVM bytecode semantics were all about Java - it forces the Java object model on you, for example. It was never designed to allow a diverse ecosystem of languages.
- kasperni 7y agoYet, we still a lot more new languages on the Java platform. Then say .NET which was designed to be polyglot.
- int_19h 7y agoIf you look at those languages closely, they are usually high-level enough that they can be mapped reasonably well to the JVM object model, or else the language itself is intentionally designed around its limitations (e.g. Scala). But something like C++ doesn't really compile to Java bytecode well. .NET CIL is much better in that regard precisely because it was designed to be low-level enough to compile C to it - it has stuff like raw pointers and pointer arithmetic, stack-allocated arrays, unions etc. The reason why not many languages bothered is longstanding lack of cross-platform support - Mono was around for a while, sure, but it was not the official implementation, and nobody knew whether it'd still be there next year. Today that's not really an issue anymore, but by now wasm is a better choice.