Python questions -> compiler and datatypes etc

[Paul Rubin]

Richard Jones <richard at bizarsoftware.com.au> writes:
> > I think the real answer to why there's no Python compiler is no one
> > has really wanted to write one.
> 
> That's not true - as I have mentioned in my last post, there have
> been efforts and as far as I recall there is an ongoing effort right
> now. I just can't find the reference at the moment.

Well, if they're determined enough, they'll see it through to completion.

> >  Python's runtime semantics are pretty
> > similar to Lisp systems, and native-code Lisp compilers have existed
> > since the 1960's or maybe even earlier.  It might even be feasible
> > to adapt one to compile Python.
> 
> I believe I've seen a similar proposal along those lines in the past too.
> 
> I don't believe it's a case of people not wanting to do it - it's
> would just required huge resources. The kind of resources only a
> commercially funded operation could provide at the moment. And I
> guess that at the moment, no-one can make a business case for it.

Oh come on, it's self-defeating to think that nothing can get done
unless a corporation does it.  It didn't take "huge resources" to
write the Lisp compilers in the first place.  The lack of a Python
compiler just isn't seen as an urgent problem right now.  Most users
are content to leave Python implementation to the central Python
developers, and the central developers are mostly concentrating on
interpreted implementations.  And Python has a useable-enough
native-code interface that bottlenecks in Python programs can usually
be relieved by converting the critical parts to a C module.  That
relieves a lot of the tension.  A Python compiler would be nice to
have, but I can think of lots of things that are more important.