Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!cornell!uw-beaver!rice!titan!preston From: preston@titan.rice.edu (Preston Briggs) Newsgroups: comp.lang.misc Subject: Re: BISON, GCC, and the GNU public license. Message-ID: <669@brazos.Rice.edu> Date: 16 Aug 89 17:22:32 GMT References:<26@ark1.nswc.navy.mil> <26609@agate.BERKELEY.EDU> <5271@ficc.uu.net> <14699@umn-cs.CS.UMN.EDU> <95@euteal.ele.tue.nl> <714@skye.ed.ac.uk> <105@euteal.ele.tue.nl> Sender: root@rice.edu Reply-To: preston@titan.rice.edu (Preston Briggs) Organization: Rice University, Houston Lines: 89 In article <105@euteal.ele.tue.nl> mart@ele.tue.nl (Mart van Stiphout) writes: >We have several commercial packages at our side and they all work >well and have excellent support. >Gnu starts out by relying on the vendor supplied !commercial! software >they despise. If they were honest guys, they would start off with writing >their own operating system. Anyone can write an editor or a diff remake >or maybe even a C compiler. I guess their C compiler is so popular, >mainly because it is better than the one supplied by Sun which is bad. >By the way the Gnu C compiler doesn't run on any machine we have >(hp, alliant, apollo). >When you talk about the quality of the >gnu software: several years ago we have been porting Emacs to our >hp9000s500. It contained lots of bugs and faulty programming. Just >think of how long ago this program was written. Its about time >it worked. >As far as I know, most of the gnu software is written by students and >by people with a job during lunch breaks (don't take this too literally). >Anayway I don't see why the gnu stuff must be better than commercial >software. People writing programs make bugs. Ever scanned all the bug >reports and fixes of gnu tools?? >Mart van Stiphout I'm glad your happy with commercial products. I use lots of commercially supplied software too, and I'm pretty happy too. However, I am sometimes (and I assume lots of people are) unhappy with certain aspects of different commercial tools. Compilers are easy to pick on, but other examples come to mind. If you have the source, you may be able to fix it. If you rely on commercial support, all you can do is report the problem and they may or may not fix it, perhaps within the next year. An easy non-compiler example: Yesterday a friend was debugging an incremental attribute grammer evaluator. He was using grep to wade through thousands of lines of debugging output. But grep (on our machine) truncates lines at 1000 or so characters, and this was a problem. GNU grep however, doesn't seem to have the same arbitrary limit. Even if it did, we could have edited the source and recompiled. You suggest that the FSF people are dishonest!? I've always thought they were scrupulously honest. And they are working on there own operating system; but first things first. I don't believe just anyone can write a {\em quality\/} editor, diff, make, compiler, ... ; there are so few examples extant. I expect GCC is popular because it is faster, and generates better code, and is cheaper, and comes with more source, than many commercial compilers. You wonder "why use GCC"; I would wonder "why (pay to) use another" (especially in a research setting). Lots of the GNU stuff was written by rms, who is not a student, nor is he an employee. Some GCC ports were performed by people who wanted the compiler to run on their system. This isn't illegal or even immoral. If we want to use C++ for a project, we can pay AT&T a lot of money (assuming they have a port to our system) or we can port GCC/G++ ourselves. My boss thinks that investing some time in a porting effort is not such a bad deal. From one of your earlier postings, I understand that you would rather not spend your time coding. I'd rather not have to re-invent the wheel either; but isn't that what FSF is all about? If the source is available, and of high quality, then we won't have to re-invent the wheel all our lives. Finally, why is FSF software necessarily of high quality? To some extent, it's better because rms (and others) have deliberately made it better. That is, they examine a tool (like grep) and say "What's wrong with this tool?" Well, it has thus and so limitations and it's too slow. Then they work hard overcome these obstacles. They don't clone tools; they make better tools (better mousetraps). And they make mistakes; so does everyone else. But GCC's bugs are published; commercial companies simply keep their buglists private. Preston Briggs preston@titan.rice,edu