On Feb 10, 2014, at 5:19 PM, Guido van Rossum <gu...@python.org> wrote:

> Then we're doomed, because this is entirely political (the companies making 
> browsers must want to do it).

Not at all.  There are a number of promising proofs of concept, demonstrating 
different aspects of this problem: Pyjamas, the PyJS fork of Pyjamas, Skulpt, 
Empythoned (which actually compiles CPython to run in the browser with 
surprisingly reasonable performance!), Brython, and PyPy's now defunct JS 
backend to name a few.  With source maps, it's even possible to use native 
JavaScript debugging and profiling tools on Browser-hosted Python.

The political problem is entirely based around getting the various interested 
parties together and trying to hammer out some kind of common core that they 
can all get running and contribute to, rather than duplicating tons and tons of 
effort and only ever getting to a 60% solution.

Browser vendors seem perfectly happy to treat JavaScript as a compilation 
target; efforts like ASM.JS are in fact trying to formalize this process and 
provide support for it.  Most of these Py-to-JS converters work fine on mobile 
devices, too.

Guido, if you haven't checked it out, <http://www.brython.info> is a very 
interesting demo: it literally makes <script type="text/python3"> work in a web 
page, with nothing beyond a single JavaScript file include (no local 
compilation, no offline processing).  The only problem is that the language 
really isn't quite Python - see, for example, 
<https://bitbucket.org/olemis/brython/issue/147/generators-send-method-missed>. 
 But they've made a surprising amount of progress since I last checked, and 
they seem to be closing issues all the time :-).

-glyph

_______________________________________________
IDLE-dev mailing list
IDLE-dev@python.org
https://mail.python.org/mailman/listinfo/idle-dev

Reply via email to