Path: utzoo!attcan!uunet!lll-winken!elroy.jpl.nasa.gov!usc!wuarchive!mit-eddie!rutgers!cmcl2!kramden.acf.nyu.edu!brnstnd
From: brnstnd@kramden.acf.nyu.edu (Dan Bernstein)
Newsgroups: comp.lang.misc
Subject: Re: Aggressive optimization
Message-ID: <2205:Nov607:20:3090@kramden.acf.nyu.edu>
Date: 6 Nov 90 07:20:30 GMT
References: <2301@wn1.sci.kun.nl> <8960020@hpfcso.HP.COM> <4179@goanna.cs.rmit.oz.au>
Organization: IR
Lines: 14

In article <4179@goanna.cs.rmit.oz.au> ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe) writes:
> Surely the cardinal rule for an assembler is not "one-to-one translation"
> but "full access to all machine operations".

Exactly! As a matter of principle I'm even willing to argue that the
programmer should be able to exert control over instruction scheduling,
even though I am reasonably confident that the compiler does this well.

> Why is it evil to offer
> optimisation as well?

It's not. I hope we can get Herman to agree on this.

---Dan