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 __@/