Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!accuvax.nwu.edu!mccoy From: mccoy@accuvax.nwu.edu (Jim McCoy ) Newsgroups: comp.lang.c++ Subject: Re: libg++ and copylefts Message-ID: <904@accuvax.nwu.edu> Date: 18 Jul 89 00:56:57 GMT References: <2053@dataio.Data-IO.COM> <6590203@hplsla.HP.COM> Sender: news@accuvax.nwu.edu Reply-To: mccoy@accuvax.nwu.edu (Jim McCoy ) Organization: GNU's not Usenet Lines: 104 In article <6590203@hplsla.HP.COM> jima@hplsla.HP.COM (Jim Adcock) writes: >>Might I suggest to potential developers for GNU a solution to the libg++ >>problem: >> >> Redevelop libg++ from scratch. Explicitly make it *PUBLIC DOMAIN*. >> This would solve those nasty copyright problems. > >Or just remove the copyleft from the existing libraries. I keep getting >tempted to just rewrite this stuff on my own, but to do so is a bore, and >clearly to have to waste my time doing so is destructive to the interests >of humanity at large. How much better if Gnu would simply lift these >restrictions, and I could spend my free time developing new software >to share? > Yes, it probably would be easier, but since you seem to think that PD libraries are vitally need, then please make them and release them. "If you aren't a part of the solution, you are a part of the problem." >>Copylefting library code simply insures that nobody will use them. > >I wish this were the case. I'm afraid some people get sucked into believing >Gnu software is really "free"; becoming mislead by the "FSF" nomiker >that they can use Gnu software without restriction. In fact many people >on the net have mistated this many times. Example: the number of people >who don't understand the use of bison is restricted. > Copylefting library code means that no one will steal them. If you want to steal code, then I think GNU should do everything it can to stop you. Bison, as is stated in the manual, includes a part of the GNU code into the generated file (e.g. you are using GNU code). If you don't RTFM, I have no sympathy for you. >>It's even worse than when some misguided compiler vendors tried >>to charge royalties for products compiled by their compilers. This practice >>was soundly rejected by the marketplace (though every once in a while, >>somebody tries it again!). > >Compiled? Or compiled and linked? It would be nice to know that if one >bought a compiler one was automatically allowed to distribute executables >produced by that compiler, but I don't know how widely allowed this practice >is among compiler vendors. I do know we have to go through a painful process >of figuring out how to make sure we're not stepping in someone's licensing >restrictions everytime we design a product where I work. > >As a compiler user, this is a pain in the butt. Using an editor to create >source code, or a laser printer to print it, doesn't put copyright >restrictions on your work. Why should using a compiler to compile >and link it put copyright restrictions on your work? > Using an editor to create source code does not put the source code for the editor (which is probably copyrighted) into your product. Using a laserprinter to print your source code doesn not include font descriptions into your output (although there is currently contention and accusations concerning reverse-engineered fonts) or printer driver code into your printed source code. Using a compiler to compile your source code does not include compiler source code into your source code. Using bison to generate your compiler includes most of the bison parser into your source code. GNU claims that linking in their libraries is using their code, although this is arguable. If you intend to steal GNU code, then use the bison parser and sell your product. If you want to make c++ libraries that everyone can use for commercially hoarded products, MAKE THEM! Don't look to GNU to cater to your wish list. >What it seems we need is some clever court or congress to explicetly state >that one can use a compiler for a language -- including necessary and standard >libraries -- to express oneself without reprisal. God help us if people who >wrote editors were allowed to restrict what we could do with our work. >Why then do we tolerate compiler vendors placing these restrictions on >us? This is total bullshit! What is "necessary and standard"? Who determines what is "necessary and standard"? What if this "clever court or congress" decides later that the new library that you or your company has just released is so usefull and revolutionary that is is "necessary" for rapid progress? The court protects copyrighted source code. If you want to steal code, don't look to the net for help. Currently people looking at C++ have two usefull options: 1) Pay AT&T lots of money and get cfront 2.0 2) Get the GNU g++ compiler for free and make their own libraries if they are doing comercial development. Perhpas we should worry less about how we can "boost g++ libraries" and how any company that offers a free compiler, but doesn't let you steal their code is some kind of devious crypto-fascist organization. Maybe if people keep whining enough about how the GNU copyleft applies to products linked to their library, someone will get fed up with this political BS and make a library. I am comfortable with the GNU restrictions and will continue to use their products, if you are not, then that check should be made out for AT&T as soon as possible. jim ------------------------------< Jim McCoy >------------------------------------ mccoy@acns.nwu.edu | "...far too many notes for my taste" #include | -Phantom of the Opera "To thine own self be true"