Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site brl-tgr.ARPA
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!think!harvard!seismo!brl-tgr!tgr!jon@cit-vax.ARPA
From: jon@cit-vax.ARPA (Jonathan P. Leech)
Newsgroups: net.lang.c
Subject: Re: more questions about efficient C code
Message-ID: <11443@brl-tgr.ARPA>
Date: Tue, 9-Jul-85 22:03:53 EDT
Article-I.D.: brl-tgr.11443
Posted: Tue Jul  9 22:03:53 1985
Date-Received: Fri, 12-Jul-85 00:44:56 EDT
Sender: news@brl-tgr.ARPA
Lines: 24

> From: DHowell.ES@XEROX.ARPA
> ...
>				    Idioms which are believed to be more
> efficient (sometimes mistakenly) should be given up in favor of
> constructs which are common to most programming languages.  I know that
> things can't be done the same in all programming languages, but what can
> be, should.  It is simply a matter of making it understandable to all
> who are involved with a project, programmers and non-programmers alike.
>
> Dan

    Would you then recommend that we eliminate pointers and structures
from our code? BASIC, FORTRAN and APL don't have them.	I  will  place
myself	out  on  a  limb  by   stating	 that,	 in   my   experience,
non-programmers don't look at code and care that  is  WORKS,  not  how
elegant it is. I would hate to program	in  a  common  subset  of  all
languages. I far prefer to let the comments describe the algorithm and
the implementation be hand-optimized and obscure, if that makes it run
significantly faster.  And sometimes it does. I've been doing a lot of
ray tracing recently, and at >8 hours frame, I CARE about efficiency!

    Jon Leech
    jon@cit-vax-arpa
    __@/