Stuff

[Tom Novelli]

----- Forwarded message from Tom Novelli <tcn> -----

Date: Sun, 25 Apr 1999 15:33:45 -0400
From: Tom Novelli <tcn@tunes.org>
To: Maneesh Yadav <97yadavm@scar.utoronto.ca>
Subject: Re: Stuff
X-Mailer: Mutt 0.95.3i
In-Reply-To: <Pine.GSO.4.05.9904241714060.18986-100000@fissure.scar.utoronto.ca>; from Maneesh Yadav on Sat, Apr 24, 1999 at 05:47:04PM -0400

Well, I'm inclined to disagree with you.  But I guess that's because we have
different goals, and something called "Tunes" is just one of them.  My main
goals are efficiency and simplicity... the theoretical stuff we're working
on in Tunes is only part of that; the other part is a clean, efficient
implementation that fits the philosophy of Tunes, the way a Lisp Machine
fits Lisp.

Let me stress that Retro is a prototype.  It's constantly changing;
sometimes big chunks of it are scrapped and redesigned.  We're writing code
and we're learning a hell of a lot... it'll all pay off when we're writing
the "real" Tunes and we have all that to draw from.  Right now, it gives us
the freedom to modify the languages to our liking.  Scheme, Common LISP,
Forth... none of those is perfect.  We can do better.  Also, Retro already
includes some foundations for metaprogramming.  It's moving on to higher
levels.  Anyway, at least we've got the ball rolling... Tunes is no longer
"just talk."

As for drivers, we don't need to support everything, just the basics...
floppy, IDE, SCSI, keyboard, standard VGA.  It gives us something to play
with.  The way I see it, Tunes is for the next generation of computers,
where software will have to be written from scratch anyway (hopefully
they'll be designed better, so it won't be so hard).  If Tunes becomes
popular on existing PC's, specialized drivers for all the various video and
network cards will follow.

-- 
Tom Novelli <tcn@tunes.org>

----- End forwarded message -----