Copied from comp.programming

Bennett Todd (bet@std.sbi.com)
Fri, 14 Jan 1994 01:41:12 -0500 (EST)

I just shoved off this little note into comp.programming; Python folks might
want to peek there, in case it draws any interest.

-Bennett
bet@sbi.com

>From: bet@std.sbi.com (Bennett Todd)
>Newsgroups: comp.programming,comp.unix.advocacy,comp.lang.c++,comp.lang.c
>Subject: The next programming language (was K&P, and before that Windows NT)
>Followup-To: comp.programming
>Date: 14 Jan 1994 06:38:03 GMT
>
>[NB: I'm redirecting followups to comp.programming, since I'm talking about
> language choice, not C]
>
>In article <1994Jan12.222149.15241@vulcan.iss.bnr.com>,
> <rpineda@alpha.ftw.bnr.com> wrote:
>>>As a programmer (employee or prospective employee), I don't really
>>>care *why* an employer is using a certain language, just that they
>>>*are* using it and *will continue* to use it for a while.
>
>>As a battle weary, in the trenches, UNIX hack, may I jump in and say
>>that I DO care.
>>
>>Software is broke folks...its too big for its britches...efficiency, in
>>all respects, is nil. If you've done more than 3 major projects...you
>>know the ugly little secret about the software industry, it ain't
>>pretty. We can't continue doing software the way its been done in the
>>past; if we want to live up to the promise of complex software systems.
>>SOMETHINGS got to give.
>
>
>I've been feeling about the same way. I pretty well mastered C a few years
>ago. Then I learned Perl. I haven't written much C since then. Still, I'm
>finding some jobs are hard to do really cleanly with just C, Perl, Bourne
>Shell, and all the tools shell can invoke.
>
>So I started looking again. TCL is a strong effort to add a missing bit, but
>I'll confess I'm not overmuch fond of the core language. Scheme has more
>architectural strength, but I haven't managed to get over a distaste for all
>those parentheses. Then I found Python.
>
>Like some implementations of Scheme, and like TCL, Python is designed to be
>extensible and imbeddable. Python is strongly object-oriented, and its OO
>structure fits _beautifully_ with built-in support for dynamic linking to
>shared libraries. Like Scheme, it's another effort to do The Right Thing.
>
>It has a small, simple language definition. Programs written in Python are
>spare and clean; Python code is extremely uncluttered, with few noisy
>punctuation marks. Its language focus is namespace management, and its
>namespace management seems to be a good fit to software engineering needs.
>
>
>I'm not yet sure that Python will be the next revolutionary language to
>extend our ability to write large, sophisticated programs and still have
>clean, simple, maintainable code. It might be some other language. But I'm
>inclined to bet that it will be an object-oriented, extensible, imbeddable
>language, with many dynamically loadable libraries providing convenient
>access to all _kinds_ of higher-level functionality.
>
>Also, I suspect both the next revolutionary language, and most of its
>libraries, will be written in C.
>
>And as long as I'm venturing predictions, I also expect that the next
>language will have an exception system well-integrated in the core
>implementation, so that straightforward code, written the most terse,
>obvious way, will properly check for errors and exit with an informative
>message if any occur, unless the programmer elects to handle the error
>condition explicitly. Python nails this one.
>
>-Bennett
>bet@sbi.com