Path: utzoo!attcan!uunet!lll-winken!lll-tis!oodis01!uplherc!sp7040!obie!wsccs!terry
From: terry@wsccs.UUCP (Every system needs one)
Newsgroups: comp.lang.c
Subject: Re: Optimization (was Re: volatile)
Message-ID: <537@wsccs.UUCP>
Date: 17 May 88 03:57:22 GMT
References: <13074@brl-adm.ARPA> <3938@killer.UUCP> <511@wsccs.UUCP> <4628@ihlpf.ATT.COM>
Lines: 18

In article <4628@ihlpf.ATT.COM>, nevin1@ihlpf.ATT.COM (00704a-Liber) writes:
> In article <511@wsccs.UUCP> terry@wsccs.UUCP (Every system needs one) writes:
> 
> >I didn't say they shouldn't effect the generated program, I said that they
> >shouldn't effect the OPERATION of the generated program.
> 
> What's the difference between your two statements?  Optimizers, almost by
> definition, do effect the operation of the generated program.

Not from a black-box point of view.  The only thing that's effected from the
user's point of view is execution speed and/or resource usage, not what the
user thinks of as "operation".  The same information in produces the same
information out.

My qualm is that badly written optimizers operating on "good code" (code not
dependant on side effects) sometimes change the shape of the black box, and
they should not.

				terry@wsccs