Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!asuvax!noao!arizona!gudeman
From: gudeman@cs.arizona.edu (David Gudeman)
Newsgroups: comp.lang.misc
Subject: Re: Algol, and language design
Message-ID: <23847@megaron.cs.arizona.edu>
Date: 3 Aug 90 19:00:10 GMT
Organization: U of Arizona CS Dept, Tucson
Lines: 27

In article  <1990Aug2.224828.2867@rice.edu> preston@titan.rice.edu (Preston Briggs) writes:
>
>Well, for research purposes, maybe.  Link-time optimization
>means waiting 'til you have the entire program in one place
>so you can optimize over the whole thing (think expensive).
>Remember that every change means re-optimizing the entire
>program (as you noted), including all the libraries you normally link in.

My model of optimization is that it is something that is done after
you have done all your testing and are producing a production version
of the program.  In this model, it doesn't matter if it takes a
hundred times longer to compile.  (Just because I know someone else is
going to say it if I don't: Yes, you should run your test cases again
on the optimized code.)

>In my sample languages (C, Oberon, Scheme), I tried to be fair and take 
>into account normal usage.  C is perhaps a counter-example, and it's 
>only included because it's relevant to most readers.

I don't know much about Oberon, but I know it is easy to write a lot
of Scheme code without resorting to procedure-valued variables. Of
course, they do tend to crop up a lot more than in C...
-- 
					David Gudeman
Department of Computer Science
The University of Arizona        gudeman@cs.arizona.edu
Tucson, AZ 85721                 noao!arizona!gudeman