Return to BSD News archive
Path: sserve!manuel!munnari.oz.au!news.hawaii.edu!ames!olivea!uunet!mcsun!uknet!edcastle!aiai!jeff From: jeff@aiai.ed.ac.uk (Jeff Dalton) Newsgroups: comp.unix.bsd,alt.suit.att-bsdi Subject: Re: UNIGRAM's article on the USL-BSDI suit Message-ID: <7079@skye.ed.ac.uk> Date: 6 Aug 92 11:52:12 GMT References: <7065@skye.ed.ac.uk> <o772klk@twilight.wpd.sgi.com> <1992Aug06.010408.2470@kithrup.COM> Organization: AIAI, University of Edinburgh, Scotland Lines: 48 In article <1992Aug06.010408.2470@kithrup.COM> sef@kithrup.COM (Sean Eric Fagan) writes: >In any event, the USL suit is either claiming outright infringement (i.e., >there is USL code in Net/2 and/or BSDI's code), or a form of infringement I >can understand, if not agree with. Namely, that the Net/2 code was written >to replace USL code, piece by piece, fragment by fragment, by people who >knew the USL code, and, therefore, the code is based upon USL code. Or a >combination of both, of course. There's certainly a sense of "based on" that doesn't even require piece-by-piece replacement. Eg, suppose I learn about C compiler by reading a particular compiler and then write a new compiler from scratch using the same strategies and tactics (though w/o lining things up procedure by procedure). The person who wrote the 1st compiler might well think "Dalton wouldn't have his compiler if he hadn't read mine. And he's taken all my neat ideas." Whether he would have a legal remedy in this case I don't know. On the other hand, do we really want everyone to have to start over from nothing when learning CS (if it can still be called learning)? Surely there are some sources (books, articles, public domain code, etc) that we can use? What if the "person who wrote the 1st compiler" wrote a book too, and I read his book, his compiler, lots of other books and articles, other compilers? What if I started with his code, added lots of neat ideas, and eventually replace all of his code with my own? And what if all the neat ideas in his code had also been published in books and articles? It seems to be in this case, which is much more like the UCB/BSDI one, that what I've done should be allowed unless my replacement code follows his very closely, thus infringing his copyright. And then I ought to be able to replace it with code that doesn't infringe. In short, I think a copyright claim might make sense, if there's evidence to support it. And no one seems to be suggesting that patents are involved. That leaves trade secrets. Can it really be a valuable trade secret that someone implements something by writing two procedures F and G rather than dividing things up differently as F', H, and I? So it seems to me that if there are no copyright violations and no use of interesting, unpublished algorithms and techniques, then AT&T/USL do not have a claim I would find reasonable. -- jd