pete@octopus.UUCP (Pete Holzmann) (07/15/88)
In article <1095@maynard.BSW.COM> campbell@maynard.UUCP (Larry Campbell) writes: >Gimme a break. You can afford to spend 1000 bucks on a computer and can't >afford 60 bucks for a compiler? Of course. But not all useful stuff can be distributed as Turbo C source code alone. Perhaps because of the widespread proliferation of PC's (and PC developers) [I think the order of magnitude is something like 1000 PC's for each Unix box; certainly much more than 100:1]... there are a zillion different tools used for developing useful stuff on PC's, all for various good reasons. It would cost a small fortune, and be silly to boot, to require all net-readers in such an environment to maintain a copy of all of the useful development tools needed to recreate all of the useful utilities out there. Just a few examples: the recent keyboard-remapping TSR programs (*tiny* things) were written in Forth. I don't have Forth, never want to have it. The posted binary utility programs are *real* handy though. Many PC-based programs are linked using PLink [an overlay linking loader that works]. Few programmers have PLink. dBase files are the database standard on PC's, not dbm. Thus, a useful data oriented program might be written in the dBase language. If compiled by the author with a dBase compiler, the resulting binary could be used by everyone, whether or not they have a copy of dBase or a clone. I could go on and on. My point: Being a PC programmer is different from being a Unix programmer in many ways. Please don't cast PC programmers in a mold based on your non-PC experience and preconceived notions. I happen to be both a Unix and PC programmer (doing both heavily), thus I am (hopefully) able to see both sides of the fence a little better than some. >[further comments about source being better than binaries] There is no question [in my mind] that source code is better than binary code for revealing algorithms used [for educational and malicous-avoidance purposes]. Source code can also be ported and improved by the recipient. Binary code also has its advantages, as pointed out in other articles [a few: reliable operation due to independance from recipient's tools; author's control over quality of porting, 'improvements' and other maintenance issus; ability for author to use specialized tools in creating the program; ability to eliminate malicious 'improved' versions; educational value of user interface and implementation in proprietary code-- you may not mind letting me see exactly *how* your $5000 CAD program works, even though you would never give me source code--]. The key is that *both* methods of distribution are valuable. Ideally, I'd like to see most PC postings in binary AND source form. If some authors want to only distribute their code in binary form though, I can go along without trouble. Pete -- OOO __| ___ Peter Holzmann, Octopus Enterprises OOOOOOO___/ _______ USPS: 19611 La Mar Court, Cupertino, CA 95014 OOOOO \___/ UUCP: {hpda,pyramid}!octopus!pete ___| \_____ Phone: 408/996-7746