Approval of IWL - Consolidated Response

[Bruce Perens]

Gernot Heiser wrote:
> I really have a problem understanding where you draw the boundary
> between this and the GPL situation, and how you can do it
> unambiguously. "All code" is bad, but some code is ok? How much is
> "some"?
Discussion of this topic requires more precision than above if you're 
going to get anywhere :-)
It's not a matter of what code is good or bad. It's a matter of what 
constitutes a derivative work.

Russ is speaking from an assumption that an API is always the boundary 
of a derivative work. There is argument about this, and attorneys that 
I've asked say "yes", "no", or "nobody knows". The "nobody knows" is 
because copyright law doesn't define what is a derivative work in 
software. We have some cases that have touched on the issue, but mostly 
in courts too low for them to be definitive in more than their own 
jurisdiction.

The Open Source Definition doesn't state an opinion on the API issue 
either. But it does state an opinion on entirely separate programs being 
pulled in by the license, which is what you are trying to do.

It strikes me that you are being confused by issues that aren't really 
germane to the question about your own license.

The definition of what constitutes a "separate program" has been blurred 
by all of the various RPC schemes like Corba, DBus, etc. that can make 
two programs function as if they are effectively one. But that argument 
could as easily be applied to any server-client protocol all the way 
back to the days of FTP and Telnet. If it's in a separate executable 
file and runs in a separate process, the OSD treats it as a separate 
program. So does FSF in licenses like the GPL.

    Thanks

    Bruce