RE: Roaring Penguin's view

> Am Mit, 2002-06-12 um 16.11 schrieb Patrik Stridvall:
>  
> > Ah, now we get to one of the core problems with the GPL.
> > The difference between "derive a product from" and
> > "develop a product for" is not defined by the GPL
> > quite the opposite. However note that regardless of what the
> > GPL say copyright law is unlikely to let the GPL push the
> > issue too far. Even the FSF recognize this.
> 
> The FSF if pretty clear about its viewpoint: Linking (static 
> or shared)
> forms a derived product, communication through IPC mechanisms doesn't.

IIRC they state that communicating between IPC might not create 
a derived product. However that this depend on details that ...

> The FSF itself states that that's eventually up to a judge to decide.

... are up to a judge to decide.
 
> In this respect, the GPL is indeed more restrictive than some 
> commercial
> licenses (Microsoft grants you to link with their core DLLs - is that
> true for _all_ DLLs?)

The main point is not was Microsoft does or does not.

It is whether "develop a product for" lies out of the
scope of the copyright of the orginal product.

I think most people intuitively think it does,
with the possible exception of GPL hardcores.
 
> Thus, if glibc became LGPL, you could still develop a GPL'd 
> wrapper app 
> that would communicate with your non-free program through 
> e.g. a socket,
> transferring the library calls to glibc, just like wineserver is
> currently doing it for wine (though not for licensing reasons).
> Actually, I am pretty certain that such a wrapper would be among the
> first things created if glibc became GPL, probably by 
> community members
> who'd disagree with that move.

I you do that I'm 100% sure that you will be sued by the FSF.
Whether they would win is another matter.