Path: utzoo!attcan!uunet!wuarchive!psuvax1!rutgers!mcnc!uvaarpa!murdoch!uvacs!dmw9q
From: dmw9q@uvacs.cs.Virginia.EDU (David M. Warme)
Newsgroups: comp.lang.scheme
Subject: Re: what makes scheme?
Message-ID: <1990Sep4.210708.23763@murdoch.acc.Virginia.EDU>
Date: 4 Sep 90 21:07:08 GMT
References: <9008031618.AA02461@mailhost.samsung.com> <26569@nigel.udel.edu>
Sender: news@murdoch.acc.Virginia.EDU
Reply-To: dmw9q@uvacs.cs.Virginia.EDU (David M. Warme)
Organization: University of Virginia Computer Science Department
Lines: 48

In article <26569@nigel.udel.edu> carrol@udel.edu (Mark Carrol <MC>) writes:
>GJC writes:
>]Maybe it is possible to have an intelligent discussion about this.
>]Q: What is Scheme?
>]
>]What makes something scheme?
>
>Compliance with the specification of the language Scheme - R3RS, or soon R4RS.
>

     There seem to be two camps here.  One camp says that
implementation X is not language Y unless it conforms to ANSI/ISO
standard 3.162277*pi/2.  These are the hackers who want to buy a
Scheme implementation to hack (or work) with, and wouldn't be caught
dead with something that wasn't the "real" thing.  They spend money,
implement systems and maybe even port them, so their concerns are
real.

     The other camp says that when scheme was born as a research hack
in the 1970's, it came out of a few adjustments to LISP:

	- lexically scoped variables
	- one name space
	- functions are first-class
	- recursion is the only loop (tail-recursive implementation)

As it turns out, these few "small" tweaks produced a very different
critter, as we all know well.  I suspect that continuations came into
the language *after* they were first used to IMPLEMENT point 4 above
using CPS semantics.  Care to shed any light on this Guy?

    LISP came from humble beginnings and is now standardized in the
form of Common Lisp.  Likewise, Scheme began humbly and developed into
RnRS.  It is difficult to deny that LISP 1.5 is in fact LISP.  So too
is it difficult to deny that a dog with its tail removed is still a
dog.

     Scheme was and is a very successful research vehicle.  Language
researchers tend to balk at standards -- from their point of view,
a language stops being interesting the moment it is cast into stone.

     I believe that GJC knows all of this, and is himself doing
further research -- how can we combine the Scheme and C worlds?  Yes,
it may be missing its RnRS tail.  Compare it against Scheme's humble
beginnings, however, and it still looks like a dog.  :^)

				- Dave Warme
				  Dept of CS, Univ. of Virginia