[Fwd: Re: [Uuu-devel] languages] <-- Why Python

Mike Meyer mwm at mired.org
Mon Feb 21 01:05:09 EST 2005


paul at boddie.org.uk (Paul Boddie) writes:

> Mike Meyer <mwm at mired.org> wrote in message news:<86r7jdh5vw.fsf at guru.mired.org>...
>> Arich Chanachai <macrocosm at fastmail.fm> writes:
>> 
>> > When the CLR is integrated, it will allow a wide array of problem
>> > solving choices for uuu users.
>> 
>> You've missed the point. Allowing a wide array of problem solving
>> choices is a goal, not a means. Instead of concentrating on adding
>> langauges, you should be provding an infrastructure that makes adding
>> langauges simple.  The Plan 9 example does this best, as any language
>> that can do file I/O is supported.
>
> This discussion collides quite nicely with two of the more quoted
> articles on software development:
>
> http://primates.ximian.com/~miguel/bongo-bong.html

I've actually done some work on using CORBA as a COM for Unix (or, as
I think of it, an ARexx for Unix). After being exposed to Plan 9, I've
decided that's a better solution. CORBA has the advantage that you can
work on it without getting buy-in from kernel vendors.

> In contrast, the standard UNIX model of interprocess communication and
> introspection, extended substantially by Plan 9 and more
> conservatively by modern UNIX variants, is said to leave something to
> be desired, and those wishing more informative details about the
> interfaces exposed by programs and services usually end up putting
> something more complicated on top - see the emergence of Bonobo, DCOP
> and D-BUS.

The crucial question isn't whether the underlying mechanism is
complete, but whether it's sufficient. The underlying Unix facilities
provide byte streams - because that's what Unix does. What that leaves
to be desired is a standard high-level way for applications to
interact with each other, passing back and forth things that are more
complicated than byte streams. Having standars for these high-level
objects allows for code reuse and data sharing. This is a good thing,
and there's nothing inherently wrong with doing things this way.

    <mike
-- 
Mike Meyer <mwm at mired.org>			http://www.mired.org/home/mwm/
Independent WWW/Perforce/FreeBSD/Unix consultant, email for more information.



More information about the Python-list mailing list