This is one of the most ridiculous definitions of "derivative work" ever
conceived. It makes absolutely no sense whatsoever.
Two classic tests for derivative work are "access" and "substantial
similarity". Access is a given here. But why should a filesystem first
developed for Linux necessarily have greater similarity to other operating
system source code than one that was ported from another operating system?
A difference so great that one is a derivative work and the other one
isn't? The end products are nearly equivalent. Both rely on access to the
Linux source code. The order of development or porting is completely
irrelevant - either both (in their final form) are derivative works of
Linux or both aren't.
And the theory that either one is a derivative work is pretty thin in any
case. Is a Windows application a derivative work of Windows because that
is the first platform it was developed for? Because it makes extensive use
of hundreds of Win32 APIs? What about a business management application
that relies on the special features of a certain database? Should it be
developed for some brain dead database first, just so it can avoid the
accusation of being a derived work of the good one?
Is the contagion of being ported to a kernel or operating system so great,
that the taint can never be removed in a version for a different kernel or
operating system. That is what SCO's theory was. Derivative status as a
Copyright © 2017, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds