License debate conclusion

[Dan Kegel]

James A Sutherland wrote:
> > > before this is concluded, I will add one thing just for the sake
> > > of archival (in case this comes up again).  If ALL of wine were strictly
> > > LGPLed, basically, *all* propriatary extensions could no longer be
> > > included.  Why?  Because the output of winebuild will be LGPLed, thus
> > > no propriatary programs could use its output in creating a shared
> > > library.  winebuild is like bison in this case, and the output of
> > > bison has a special exemption so that you can use it in something
> > > other than *GPLed programs.
> >
> > Agreed, programs that generate code should specify what license the
> > generated code is under.
> 
> "Agreed"? My view is that compilers (or any "program that generates code")
> should have *NO* say whatsoever in licensing of the output. Or do you want to
> see VC++ prohibiting the development of software competing with MS offerings?

Because generated code could be considered a derived work of the compiler,
programs that generate code need to expressly disclaim rights to the
generated code to avoid restricting the use of that code.  If the
compiler vendor fails to expressly spell out the licensing of the
generated code, users of the compiler are left in an unclear legal position.

- Dan