Approval of IWL - Consolidated Response

[Gernot Heiser]

>>>>> On Wed, 11 Jun 2008 22:08:13 -0700, Bruce Perens <bruce at perens.com> said:
BP> Russ Nelson wrote:
>> Gernot Heiser writes:
>> > >>>>> On Wed, 11 Jun 2008 17:13:10 -0700, Bruce Perens <bruce at perens.com> said:
>> > BP> FYI, I wrote OSD #9 and agree that its language is intended to prohibit
>> > BP> the terms in your license section (c)(ii).
>> > 
>> > Ok, so please excuse my ignorance. If this is a problem for the IWL,
>> > why is it not a problem for the GPL?
>> 
>> Because the GPL relies on there being a single work.  The GPL doesn't
>> try to reach across an API into other pieces of software.

That's not what the FSF says, they claim that linking against GPLed
code makes your code GPLed, a clear case of reaching across APIs. And
if I'm not mistaken, this view has been upheld by courts in Germany
(but I'm not a lawyer). So, is the FSF wrong? Are the courts wrong?

BP> Essentially this is the difference between "use" and creation of  a 
BP> derivative work.
BP> The OSD was created as a policy of the Debian distribution, when I was 
BP> the Debian project leader. One of my goals was that plain users, those 
BP> who weren't modifying the software in the Debian system, would be able 
BP> to use all of the software in the Debian system without concerning 
BP> themselves with the licenses or requiring an attorney. Debian would not 
BP> have been practical to use otherwise. But that applies to Open Source as 
BP> a whole, too. Any significant restriction on use makes it impractical.

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"?

Gernot