Re: Bug#570621: Parsing output = derivative work?
The distinction between a derivative work and a separate work is not
based on technology but on functionality.
Parsing the output of a program doesn’t make a derivative work. However,
if this parsing is vital for the operation of the application and makes
it useless without that program, what is the difference with dynamic
linking to a library? To a programmer, there might be one, but to a
court, there wouldn’t be any.
By this reasoning, if I write a program which converts another word processor's
output to Microsoft Word format, then that program is a derivative of
Microsoft Word, at least until Open Office gets a filter good enough to read
Moreover, by this reasoning, if I write a program that runs only on Windows,
or which interfaces with some proprietary Windows protocol, Microsoft can
legitimately claim that I am violating their copyright by creating an
unauthorized derivative of their work.
This definition of "derivative work" is something which the FSF claims, but
which many people outside the FSF are skeptical of precisely because of
absurd consequences like these.