Re: [Gnu-arch-users] OT: Lisp

[Andrew Suffield]

On Wed, Nov 19, 2003 at 06:18:19PM -0800, Tom Lord wrote:
>   Programming language "fads" show up from time to time:
> "object oriented programming"

Predates lisp, although it wasn't recognised as such until later. It's
merely a formalisation of the "natural" way to think about large-scale
programming, to make it easier to teach. In the past 10 years, I've
seen maybe three significantly large programs that weren't object
oriented (and they were all crap).

> "constraint-driven data flow"

That's been around for decades too. Furthermore, it's not a "fad",
it's part of a suite of techniques that are the only way we currently
have to write verified code (that is, code which has been
mathematically proven to be correct and cannot contain any bugs; yes,
it's possible). It's the opposite of a "fad" because it isn't, and
likely never will be, popular. It's something you do only when you
absolutely have to (and increases development time and cost by
orders-of-magnitude).

> "aspect oriented programming".

This one's new, but again, not popular. Aspect weaving is still in its
infancy. It's about as much a "fad" as quantum computing, and about as
far along development (ie, "maybe it'll be ready within 10 years").

>   With the fads often come
>   new language designs that try to use the fad as the central,
>   organizing concept --- such languages attempt to simplify the
>   problem of programming language design by finding a "magic bullet"
>   that eliminates the need for any other approach.

I can't think of any languages like this. Examples?

Certainly all the languages common in the real world (C, C++, Perl,
Java, COBOL, PL/X, ...) don't have any singular focus on one
particular technique.

>   I think that such "fad languages" are perpetually doomed to
>   long-term failure or severe mutation.  There is no magic bullet in
>   language design.

That sounds plausible enough. It's probably one of the reasons why the
popular languages don't fall into this category.

>   Rather, there is a rich tapestry woven from many
>   divergent lines of thought about language design -- the tapestry
>   itself an attempt to bring those lines of thought together to form a
>   pleasing and coherent pattern.  That rich tapestry is lisp.

...and that's just post hoc ergo propter hoc, combined with the
fallacy of exclusion (you could just as easily have said the same
thing about most of the popular languages, except possibly COBOL, and
it would be about as valid).

-- 
  .''`.  ** Debian GNU/Linux ** | Andrew Suffield
 : :' :  http://www.debian.org/ |
 `. `'                          |
   `-             -><-          |

signature.asc
Description: Digital signature