Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!iuvax!purdue!mentor.cc.purdue.edu!l.cc.purdue.edu!cik From: cik@l.cc.purdue.edu (Herman Rubin) Newsgroups: comp.misc Subject: Re: The GNU Public License Message-ID: <1469@l.cc.purdue.edu> Date: 3 Aug 89 15:01:13 GMT References: <26@ark1.nswc.navy.mil> <99@euteal.ele.tue.nl> Organization: Purdue University Statistics Department Lines: 38 In article <99@euteal.ele.tue.nl>, mart@ele.tue.nl (Mart van Stiphout) writes: > In article <1463@l.cc.purdue.edu> cik@l.cc.purdue.edu (Herman Rubin) writes: < >I am in no way associated with FSF, but I support the idea of software < >which the user can modify easily. I think it should be simpler and more < >powerful, and somewhat less user-inimical. I would object to a car being < >programmed in such a way that it could select a path to a destination and < >not let me override that path. > > Your analogy is rather useless. What you actually mean is: > When I buy a car and its performance does not suit me, I can > take my toolbox and start improving it. I disagree. The control mechanism here must operate by software, as the maps change. And this is the problem in many situations. > Unlike many other posters, I can't subscribe the opinion that having > the sources is necessary because the software is usually bad. > I've worked on several computers (HP, DEC, Apollo, Alliant, Sun) > for a number of years and for most of my activities (program development, > reasearch, text processing) the utilities provided are more than > sufficient. I really don't need the diff sources or the compiler > sources. > I assume people who are into operating systems and stuff like that > want to make changes in the operating system sources but > how many people are into such things? I am in no way into operating systems. The problem is mathematical software. Even worse is the problem with statistical software. Take, for example, the problem of comuting something as simple as x - ln(1 + x), and a procedure with good relative error is wanted. Speed is also important. Using the provided software will not do, although the sources make it much easier than starting from scratch. And the HLLs will not be efficient, either. I have worse problems than this. -- Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907 Phone: (317)494-6054 hrubin@l.cc.purdue.edu (Internet, bitnet, UUCP)