Re: conditional compilation

[Robert A Duff <bobduff@world.std.com>]

Brian Rogoff <bpr@shell5.ba.best.com> writes:

> On Wed, 2 Aug 2000, Robert A Duff wrote:
> > In any case, from a language design point of view, if you think the
> > compiler should see things in a different order or different
> > organization than human beings, then clearly the programming language is
> > poorly designed.  
> 
> That isn't so clear to me. I don't think Common Lisp, Scheme and Dylan are 
> poorly designed because they include powerful macro systems.

I completely agree with you, so if you think we disagree, I think you
misunderstand what I mean.  I have nothing against Lisp macros -- they
are quite powerful and useful.  Note that one feeds them directly to the
compiler, and also reads them as is.

What I object to, is using Literate Programming to stir the code around
in an obfuscatory manner.  In cases where Literate Programming improves
the structure of the code, that indicates a flaw in the underlying
programming language.  In cases where it damages the structure of the
code (see my previous post about loop exits), well, that's not a good
thing.

Note that Lisp macros operate at a completely different level than
text-based macros.  Lisp macros don't stir the code around in ways that
totally violates the language rules.  Text-based macros are evil.

>... I believe 
> their designers feel that syntactic abstraction is important enough to 
> include facilities in the language for extensible syntax. That may
> conflict with some of the design goals of Ada, but I don't thnk it
> clearly indicates poor design. 
> 
> > Macros can be used to get around poor language
> > designs.
> 
> They can also be used to change the syntax of the language so that it is
> more suitable for a particular problem. There are lots of things I don't
> like about Lisp, but the inclusion of a powerful macro system is not among
> them.
> 
> -- Brian

- Bob