There is another intersting issue in the SCO case besides the user liability for (alleged) infringment issue. In reality, I seriously doubt that the facts will support SCO's infringment claims, but suppose for now that they do support SCO. SCO is claiming that Linux contains some their copyrighted work and is thus a derivitive. However, it clearly also contains copyright work by hundreds of other parties, collectively known as kernel.org (eg Linux Torvalds, Alan Cox, Dave Miller, etc...). That copyright is licenced under the "GPL" licence: http://www.gnu.org/copyleft/gpl.html
Let's say that kernel.org created work A and licenced under the GPL. Later, IBM supposedly took parts of SCO's work B to create a new work C = A + B. The overwhelming majority of C is A (99%).
Question 1: Isn't it true that in order to distribute C, a party needs licences from both kernel.org and SCO. This applies to both kernel.org and SCO as well, in that they would each need a license from the other to distribute C.
SCO in fact took C and distributed it. As of yesterday, you can still download C from their ftp site. The GPL licence for A binds _distributors_ in two cases: covers if they modify it or if they change it from source code to binary form. SCO did both. According to GPL 2b and 3, in either of these two situations, the distributor of any work based in part on GPL'd work is required to independently license the whole of what they distribute under the GPL as well, and to do so to "all third parties" (which includes IBM).
Question 2: If question 1 is "true", then how can SCO distribute C legally and still claim C is improperly licneced? Isn't SCO essentially proving their own infringement of kernel.org's rights to A if they prove IBM infringed B by making C=A+B? How can they distribute a derivitive of A?
This archive was generated by hypermail 2.2.0 : Mon Mar 26 2007 - 00:35:49 GMT