JavaScript/NodeJS code port

[Utopiah]

Hi, the same way the foundational "Structure and Interpretation of Computer Programs" (or SICP for short) has been ported to JS, cf https://sourceacademy.org/sicpjs/index , and arguably how this book core ideas are not language dependent, it could be interesting to port from Python to JS.

This thread is to discuss how, what library to potentially rely on, which one to avoid (because it would "shortcut" the process), etc.

Please share your view!

[pavpanchekha]

Here were my thoughts! https://browserbook.substack.com/p/why-python

Node.js would work better. I recommend QuickJS as the JS engine.

[Utopiah]

Wonderful! To clarify this wasn't a criticism on the initial decision and from a pedagogical standpoint I definitely see a lot of value in using Python.

I'll check on QuickJS https://bellard.org/quickjs/ but looks quite adequate indeed.

You also mention in the blogpost a student focusing on JS. Could you please point them here so that could share a bit of what worked, what didn't, the challenges encountered, and more?

Thanks for the amazing book and the quick reply.

[pavpanchekha]

That student has long graduated, but every chapter in the book has a JS version (like this: https://browser.engineering/widgets/lab10-browser.html) at the end of the Summary section. It's automatically generated from the Python code and runs against a small harness (https://github.com/browserengineering/book/blob/main/www/widgets/rt.js) that emulates the Python classes using JS APIs. If you wanted to run it all in Node you could probably modify the harness and it would work.

[Utopiah]

Feel free to share my contact information with them. They might have thrown away the code but maybe it's in a directory somewhere and they wouldn't mind open sourcing their effort, with code itself or just some pointers for others going down that path!

[Utopiah]

Beside the pedagogical aspect, one use case that might also be interesting for a JavaScript (with potentially NodeJS "backend" if truly needed, e.g. proxying to avoid CORS issues) could be an in VR/AR (or XR) for short browser ... in the browser.

This might sound very convoluted but currently XR headsets (e.g. Apple Vision Pro, Quest 3, Lynx XR1, etc) do have browsers. Those browser do support WebXR, also known as Immersive Web from W3C specifications, but none of those browsers support browsing while being immersed. What that means is that if you open a Web page that support WebXR (e.g. https://companion.benetou.fr with the bottom right VR/AR button if you visit it in headset) you can only see it's content but there is no iframe support.

According to specifications https://www.w3.org/TR/webxr/ and numerous discussions this will never be implemented due to security reasons. There are some foray in the domain, e.g. immersive-web/webxr#1365 but this is not compatible across devices.

Consequently having a browser that can render to canvas, than a 3D engine (e.g. threejs, BabylonJS, etc) would help a lot for WebXR content.

A typical use case is to have a WebXR experience then want to display, without leaving immersive mode, its documentation.