[Edu-sig] Python for CS101

[Kirby Urner]

> Not arguing against closed source - per se.  But if that is important to
> industry, let industry bear the training costs.
> 
> 
> Art

I don't yet have a clear grasp of what parts of the .NET architecture would
be closed to a Python developer.  Good question though.  

In the case of Mono, the .NET clone for Linux, you can get the source code,
including for the C# compiler.  And even on MSFT, there's nothing to keep
Python developers from sharing their .py source files -- they'd compile to
the CLR when first run, then wouldn't need to be recompiled again (like now,
with .pyc files, which any developer may share in lieu of .py files).

The situation with Jython is comparable.  You can share the source, or just
ship the .class files.  Either way, the source for the platform-specific JVM
itself (e.g. Sun's) may not be open to you.

Most Windows users of Python are aware the source is available, but few
compile it, as Windows doesn't even include a free compiler.  And look at
the Win32 extensions module, available with all Windows versions -- that
sends hooks deep into the Windows OS, lets you deal with COM objects and
provides bindings to the Microsoft Foundation Classes (MFC).  That stuff is
all proprietary/closed, even if Mark Hammond's win32 source isn't.

I don't know to what extent Jim's project at Microsoft is going to yield a
transparent stack, i.e. will all the C# code for implementing Python atop
the .NET framework be open?  If not, I'd expect the Python community to come
up with another version that was.

The metaphor of a stack (stuff on top of stuff) is a good one.  On the
Windows side, you always come down to a closed kernel at the bottom, so your
Python can't be considered "pure open source all the way down to the chip"
no matter what you do.  Besides, the chip itself is proprietary (Intel or
AMD inside).

Kirby