[Idle-dev] The Future of Python

[Glyph Lefkowitz]

On Mar 2, 2014, at 8:53 PM, Tom M <letterrip at gmail.com> wrote:

> Regarding her comments on games.
> 
> Python is a non starter for platform (Nintendo, PS3, Xbox) games - I
> discussed this with Erwin Coumans (created and develops the Bullet
> physics engine and also consults for most major physics engines) many
> years ago and he gave a number of reasons why major game developers
> are unlikely to ever use Python (this was quite a few years back but
> the comments as I recall were - lack of good embedding which she
> covers; too heavy a set of mandatory libraries; slow execution speed;
> high overhead)  which was why he went with Lua for 'gamekit' (his
> hobby game engine based on Ogre and Bullet).

This is a very real perception in the game development community, but is also mostly an issue of configuration and build toolchain.  Console games can be and have been built with Python with no particular detriment to performance.  (Case in point: you don't know which ones they are.)

> Python is a non starter for mobile games (not available on most of the
> platforms in any reasonable manner)

Also not true.  Play around with some of the simple 2D demos included with <http://omz-software.com/pythonista/> and you'll see what I mean.  Heck, in Pythonista itself, there's an "export to Xcode" button, so you can make a game jam game right there on your iPad, and actually ship it on the app store, as long as you have a mac to bounce the final build process off of.

> Since most games will originate for platforms or mobile, there isn't
> going to be much in the way of python used in games or likely to be
> used for games.

So, this is basically all nonsense but I completely understand where it's coming from, and I understand how it's nearly impossible to arrive at my perspective without decades of folk knowledge of the Python community's history.

In fact, the main thing that makes Python still interesting to me after all this time is its complete versatility.  You can put it just about anywhere.  But there needs to be better community support for these sorts of things.  (And the place to start with game developers is not to create a professional toolchain for PS4 developmen, but rather, as you point out...)

> She does make a good point that for hobbyists, one of the issues is
> making a bundled executable on windows platform, which is an issue for
> 'my first python game' type developers.

This is absolutely the problem.  The toolchain for making some code that you can share has evolved so much and requires so much arcane understanding of history (distutils, modulefinder, importlib, setuptools no distribute no setuptools again, easy_install no pip no actually easy_install because windows no actually pip with wheels) that it's basically impossible.  _technically_ the problem is all relatively straightforward and tools exist to do all the things you need, you just need to know which 9000 things to avoid and ignore while you're doing it.

Would anyone be interested in working with me on a project to create a build platform to make shippable artifacts out of Python scripts easily?

-glyph
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/idle-dev/attachments/20140304/b6d68763/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4124 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/idle-dev/attachments/20140304/b6d68763/attachment.bin>