It seems that there is a candidate (for at least partial) work-around to copyright on the net after all. This is what is called pass-by-reference in computer programming. Instead of passing a given expression around, you pass only a pointer to the expression. This technique is used, for example, by the "dpans2texi" package: see http://www.ifa.au.dk/~harder/dpans.html for details. Basically this package helps you turn a (presumed) copyrighted set of documents into a new, browsable Texinfo document. I say "presumed" because the actual copyright status of this work happens to be unknown at present; the technique would work just as well on a known-to-be copyrighted document – I think! Anyone think anything different?
This method can be extended to extremely complicated cut-up-technique programs that take various copyrighted documents from all over the net, do various things, and spit them back out in some new form. In a peer-to-peer setting, these documents can again be redistributed, by reference -- and it would seem that no one has violated any copyright laws, because everything was passed by reference, not expression. The only thing that has been passed around is a set of instructions for making a derivative work. These instructions do not appear to encourage copyright infringement, because everyone makes their own copies under fair use. Furthermore, the tool that processes these instructions (say, some future version of Arxana) does not encourage infringement either – for the same reason!
This will need more analysis, I'm sure, but if I'm right, it seems that copyright has been rendered essentially irrelevant by the combination of p2p infrastructure and sophisticated hypertext tools.
Comments?
--jcorneli Sat Mar 11, 2006