Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site harvard.ARPA
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!godot!harvard!winkler
From: winkler@harvard.ARPA (Dan Winkler)
Newsgroups: net.lang.st80
Subject: Stupid Questions about Berkeley SmallTalk
Message-ID: <419@harvard.ARPA>
Date: Mon, 4-Mar-85 00:02:43 EST
Article-I.D.: harvard.419
Posted: Mon Mar  4 00:02:43 1985
Date-Received: Tue, 5-Mar-85 01:49:51 EST
Distribution: net
Organization: Aiken Computation Laboratory, Harvard
Lines: 36



I'm about to embark on a large programming project on a Sun and I'm trying
to decide between using Berkeley SmallTalk and using C.  I've never used
SmallTalk, but from reading the books it looks easy and fun to pick up.

I'm attracted to SmallTalk because the environment looks much nicer than
C and dbx.  Also, the graphics interface looks easier to use than Sunwindows.
However, noone at Harvard uses SmallTalk and I have a few questions: 

	1) How much slower is Berkely SmallTalk than C on a Sun 2 workstation? 
	2) Is it stable and fairly bug free? 
	3) When speed becomes a real issue, will I be able to load different
		microcode on our Xerox Lisp machines and run my SmallTalk
		programs there?
	4) How easy is it to work with digitized images that have been
		ftp'd from some other machine?

The project I'm doing is an implementation of Valiant's learning protocols
applied to handwriting recognition.  On the graphics side, we will want to 
do things like run thinning algorithms on bit images.  On the learning side,
we will be working with boolean functions (mainly monomials).  Lisp is great
for working with monomials since you can evaluate p1p2p3...pt with:

		(and p1 p2 p3 ... pt)

How would you represent monomials and more general boolean expressions in
SmallTalk?  (I told you these were stupid questions!)

Thanks in advance.

By the way, Apple just released SmallTalk for the Lisa.  They had never
released it before because it was too slow, but with recent speed improvements
it is now (allegedly) reasonably quick.

Dan. (winkler@harvard)