Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!tut.cis.ohio-state.edu!purdue!mentor.cc.purdue.edu!pur-ee!hankd From: hankd@pur-ee.UUCP (Hank Dietz) Newsgroups: comp.lang.misc Subject: Copylefting Summary: After much thought, I think pure PD is best. Here's why. Message-ID: <12344@pur-ee.UUCP> Date: 25 Jul 89 01:31:37 GMT References: <26@ark1.nswc.navy.mil> <23837@santra.UUCP> <5207@ficc.uu.net> Reply-To: hankd@pur-ee.UUCP (Hank Dietz) Organization: Purdue University Engineering Computer Network Lines: 43 I have thought about this "copylefting" business quite a bit lately. T. Parr and I are on the verge of releasing Antlr (an extended-BNF LL(1) compiler-compiler which is part of his MS thesis), so it has been more than a theoretical concern.... When I first saw the info on copylefting I thought to myself, "Well, that should stop the scum from selling what we did for free." However, I have since seen that "the scum" don't really care if it is copylefted -- which only makes sense, because none of them ever seemed to care about copyrights before. They know that prosecution is mostly an empty threat. I have also seen that copylefts send a shudder down the spine of nearly every major corporation. Since they fear that they'll accidentally lose rights to some vital piece of in-house developed software by some nieve employee using a copylefted tool, many companies simply refuse to touch copylefted software... they will not even allow it on their machines. The propagation of copyleft rights via use of copylefted tools is an absurd power trip: imagine a fellow inventing a new type of nail and then claiming ownership/control of all houses built using at least one such nail because the assemblage of the house is fundamentally the use of the nails, not a function of the lumber used. This certainly violates the spirit of pure academic research as I see it; all we're supposed to want is proper credit for our good work, not a claim on somebody else's efforts. Hence, I've slowly become convinced that true "public domain" (PD) is the only way to go -- perhaps augmenting this by establishing a user group at the time of release. Companies are a very large and important user community, and I don't want to exclude them from using our software. Further, if we publicize the PD software enough, few people will pay big money for a commercial rip-off of it (maybe somebody else will get rich off of our work, but in some sense that's what academia is all about ;-). Anyway, here's what I think we will be doing: a pure PD release with a credit/modification history which must be maintained, but which carries no further rights. In other words, people would be free to sell modified versions of our software... but they must acknowledge the lineage of their product in some way. I'm not quite sure of the details... but it shouldn't be anything that would frighten away any potential users. And that's what this is all about. -hankd@ecn.purdue.edu