Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!mips!apple!agate!riacs!stanford.edu!neon.Stanford.EDU!flamingo.Stanford.EDU!espie From: espie@flamingo.Stanford.EDU (Marc Espie) Newsgroups: comp.sys.amiga.programmer Subject: Re: Lemmings - a tutorial Part IV Message-ID: <1991Apr3.104218.5224@neon.Stanford.EDU> Date: 3 Apr 91 10:42:18 GMT References: <1991Mar31.195148.6667@starnet.uucp> <24001@well.sf.ca.us> Sender: news@neon.Stanford.EDU (USENET News System) Organization: LIENS, ENS, 45 rue d'Ulm, Paris (France) Lines: 53 In article <24001@well.sf.ca.us> farren@well.sf.ca.us (Mike Farren) writes: >sschaem@starnet.uucp (Stephan Schaem) writes: [most of the discussion deleted] >>And if you have a good algorithm a C implementation will alway be slower >>and bigger than an optimized ASM one.(even if its not by mutch). > >Algorithm A: 8 hours implementation time, including debugging, in C - >32 hours, including implementation time, in asm. How many times will >you have to run it to recoup that 24 hours? Assembly where you need it, >C where you don't. You don't ALWAYS use C (at least, I don't), but you >don't always use assembler, either. > >-- >Mike Farren farren@well.sf.ca.us One very strong point of the amiga is that you can use whatever tool is best suited to what you want to do. Just for starters, you have assemblers, basics, FORTRAN, C, Pascal, ModulaII, C++, forth, scheme, lisp, prolog, rexx, postscript, tex, awk, bison, flex, mathlab, maple, oberon, draco, scriptit, etc (sorry for those I've forgotten). Ok, some of these are not public domain. But... most of them are and you've got the source in many cases. Other stuff is being ported right now (wishful thinking: mathematica). So you can CHOOSE the tool which is best suited to what you want to do. Assembler has only to do with speed issues. Except for critical portions, it is a bad idea to code in assembler. There were some very serious studies done about the way an average programmer works (netters: any reference? I don't have any on the tip of my hat). Some results are intuitive. Most are shocking to the ego of a programmer. One of the most interesting result was that, WHATEVER the language, MOST programmers can write only six (6) lines of working code in a day, if the project has a ``real-life'' size---this accounts for debugging, upgrades and such. Of course, a game rarely needs upgrades. Since Simcity was written mostly in C, there WERE indeed upgrades. Usually, once a game is finished, it's out-of-the-door- and-get-as-much-money-as we-can. Mr Farren apparently doesn't work that way and, *as he said*, some of his games still sell *4 years* after he wrote them. Assembler is not always the best tool, C neither for that matter. For most quick projects, ARexx is a good choice. Personnally, one of the language I use all the time is... postscript. I write lots of graphical applications (I'm a researcher in computational geometry). I need - device independance. - high quality output. - easy to read and debug code. As a bonus, I even get many high level features... guess if I wanted to do the same thing in assembler, it would take forever to port the stuff from my amiga to the dec-station/Laserwriters II of the lab... Food for thought. -- Marc (espie@flamingo.stanford.edu)