Newsgroups: comp.lang.dylan
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!news.mathworks.com!uunet!in1.uu.net!sytex!smcl
From: smcl@sytex.com (Scott McLoughlin)
Subject: not "Scott's proposing" or thanks K. Anderson
Message-ID: <sVac1c3w165w@sytex.com>
Sender: bbs@sytex.com
Organization: Sytex Access Ltd.
Date: Sat, 11 Feb 1995 03:43:51 GMT
Lines: 37

Howdy,
        Just to set the record straight. I was in no 
way proposing (1) that Dylan function call be implemented
like C++ vtable dispatch (2) that multimethod dispatch is
inferior to distinguished receiver or (3) that C++ is a
"good language".
       Second. Thank you Ken Anderson for posting
nifty example code. I've enjoyed your "Courage in
Profiles" and look forward to your upcoming column
in Lisp Pointers.
        Third. What I _WAS_ saying is that I have more
hope for (would prefer?) a base language (Dylan, whatever)
that compiled to speedy code so that 3rd parties could
write portable GUI libraries of various scales, target
markets, etc. More hope for such a base language than
for a common GUI interface standard that implementor
would run around and "tweak" behind the scenes as it 
were to make it run fast enough.  One way of saying
this is that I would like to see lots and lots of
"real world" Dylan code, Scheme code and Common Lisp
code without tons and tons of #+ and #- (or equivalents)
to make it go fast enough. Companies should be able
to write fairly portable X-libraries (GUI, database,
whatever) that go fast without screwing with the
base language implementation details.  #+ and #- would
then be reserved for OS dependencies, endian order
issues, and other "conditional compilation" issues
        I full recognize that is is more easily
said than done and I have great respect for the
folks who are out there pushing the envelope with
type inference systems, fancy runtime libraries, etc.
etc. etc.

=============================================
Scott McLoughlin
Conscious Computing
=============================================
