I've implemented a rough-in of a VHLL (Very High Level Language) and used it to write a non-trivial program (a mutual multi-tasking kernel that does message passing, about 1,400 LOC). It generates Python, Common Lisp and Javascript (node.js) code, essentially using existing languages as assemblers. This uses a variety of techniques, including a DPL (diagrammatic programming language), t2t (text-to-text transpilation) and a jury-rigged REPL programming language workbench. Several mundane gotchas were wrestled down. Now, I'm wondering what direction to take this next. My thought is that this needs to run in a browser (Javascript, WASM?), but that exceeds my limited knowledge and will require going down new learning curves. Maybe I'm thinking too much inside-the-box? Comments (and help) appreciated. FWIW more detail: https://programmingsimplicity.substack.com/p/rt-now-what?r=1egdky which further references https://programmingsimplicity.substack.com/p/rt-transpiler?r=1egdky

I used to use https://haxe.org/ and quite liked it and thought it might be useful for places where you need to maintain business logic in multiple places, e.g. web app + mobile app. There is still quite a frustrating part where you have to normalize external connections e.g. UI as no cross language thing can really do that. Haxe was a compelling type system and existed before Typescript. I think thats what motivated me, so I would like to understand more on what the language RT actually looks like.

I understand your question to be: "how to do cross-language stuff?". My answer to portability is in this article. [meta: is this answering your question?]