4 ms·

Seriously... do we really need to invent yet another implementation of Python that runs on the web in Javascript? I mean it with heart. How about we just have o

by yeukhon 11y ago

Seriously... do we really need to invent yet another implementation of Python that runs on the web in Javascript? I mean it with heart. How about we just have one language and one implementation forget about it? Wouldn't that be super ideal?

emily-c 11y ago
Even if it's not extremely useful it's an interesting project. People tend to like doing interesting things for fun.
- jawbone3 11y ago
  Per comments above, pypy.js is not presented as an trivial but interesting thing, but as something useful.
  - Mizza 11y ago
    It's easy to see how this could be useful for somebody though. Imagine you're a scientist, and you have a large, pure python codebase - let's say for intelligently predicting the number of foos in a bar - and you want to make that available via the browser, in order to visualize the progress of your foo-hunter. This makes it a trivial task - and means you can use cool web-based visualization tools (D3, etc), and combine with other useful APIs.
    - dorfsmay 11y ago
      I think most scientific apps written in Python rely on native non-pure-python (written in C, Fortran etc...) modules which I'd expect to not work in this context.
      - voltagex_ 11y ago
        If you wanted to go crazy you could use Emscripten but I could easily see pushing the C+Python parts to a server-side web service and then using Pypy.js for the rest.
klibertp 11y ago
> Wouldn't that be super ideal? No! Languages (and language implementations) have different strengths and weaknesses. No one language may hope to serve all needs equally well. While for what you do your single language may be sufficient, there are people and use cases where it's better to use something different. I'd rather embrace the diversity than try to artificially limit available choices.
LBarret 11y ago
Except that for historical reasons, js is a very rushed/badly designed language. More precisely it mixes moment of brillance and very stupid design choices. Ruby, Python, Lua are in term of language design way superior. Coffeescript and co emerges because it was more reliable to have a complex stack than to use the JS language directly. That's the only high-level language that produced this situation. Python/Ruby & co are not perfect but good/reliable enough. You don't program in a another language that compile to python to avoid its verbosity and its traps. If python had been choosen for the browser, we would have used for the last 15 years : module, sane scoping rule, classes, iterators, etc. Yep, that's something that it ecma6 without the ugly baggage. 15 years, damn it. So no, please give at least Python in the browser and let people experiments with other language design choices and let them find something superior to Python. At one moment, it will be obsolete like any other tool. I don't want to be forced to use it because it is the only choice.

by yeukhon 11y ago

emily-c 11y ago

Even if it's not extremely useful it's an interesting project. People tend to like doing interesting things for fun.

jawbone3 11y ago
Per comments above, pypy.js is not presented as an trivial but interesting thing, but as something useful.
- Mizza 11y ago
  It's easy to see how this could be useful for somebody though. Imagine you're a scientist, and you have a large, pure python codebase - let's say for intelligently predicting the number of foos in a bar - and you want to make that available via the browser, in order to visualize the progress of your foo-hunter. This makes it a trivial task - and means you can use cool web-based visualization tools (D3, etc), and combine with other useful APIs.
  - dorfsmay 11y ago
    I think most scientific apps written in Python rely on native non-pure-python (written in C, Fortran etc...) modules which I'd expect to not work in this context.
    - voltagex_ 11y ago
      If you wanted to go crazy you could use Emscripten but I could easily see pushing the C+Python parts to a server-side web service and then using Pypy.js for the rest.

klibertp 11y ago

> Wouldn't that be super ideal? No! Languages (and language implementations) have different strengths and weaknesses. No one language may hope to serve all needs equally well. While for what you do your single language may be sufficient, there are people and use cases where it's better to use something different. I'd rather embrace the diversity than try to artificially limit available choices.

LBarret 11y ago

Except that for historical reasons, js is a very rushed/badly designed language. More precisely it mixes moment of brillance and very stupid design choices. Ruby, Python, Lua are in term of language design way superior. Coffeescript and co emerges because it was more reliable to have a complex stack than to use the JS language directly. That's the only high-level language that produced this situation. Python/Ruby & co are not perfect but good/reliable enough. You don't program in a another language that compile to python to avoid its verbosity and its traps. If python had been choosen for the browser, we would have used for the last 15 years : module, sane scoping rule, classes, iterators, etc. Yep, that's something that it ecma6 without the ugly baggage. 15 years, damn it. So no, please give at least Python in the browser and let people experiments with other language design choices and let them find something superior to Python. At one moment, it will be obsolete like any other tool. I don't want to be forced to use it because it is the only choice.