[plt-scheme] Code for Krishnamurthi's "Swine Before Perl"

From: Shriram Krishnamurthi (sk at cs.brown.edu)
Date: Thu Jun 10 17:21:43 EDT 2004

Bruce Butterfield wrote:

>						 Obviously it is simpler 
> to implement a language like Scheme in Scheme since then I can leverage 
> all the transformational power of the embedding language. 

Well, there's syntax and there's semantics.  Greg Cooper's FrTime
language is a good example of something that sytactically looks like
Scheme, and superficially also seems to behave similarly, but is
semantically quite different.  Implementing it in Scheme wasn't such a
great win, even though the language is parenthetical.  (I mean, it
wasn't a *loss* -- obviously it was better than doing it in C++, say.
But it's not clear ML would have been worse.  This is also orthogonal
to the fact that *DrScheme* was a huge win.)  So you'd have to be
quite close to Scheme's semantics and/or libraries for the
transformational power to really kick in.

> If your point is that one will never be able to build a rich
> language around a syntax that is not inherently s-expression based
> -- I defer to your greater experience.

I didn't say that, and I don't believe that.  As I get older, I've
begun to think kinder thoughts about infix.  That said, there's a nice
quote (that I bet Joe Marshall can scare up in a moment's notice) in
the Steele and Gabriel survey of Lisp that describes how and why
languages come to look like Lisp.


Posted on the users mailing list.