[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

unsubscribe




Byju Michael
Dell | Product Group 
Bangalore Development Centre 
Bangalore, India
Direct Dial: +91-80-280-77048
		 +91-80-511-77048



-----Original Message-----
From: Andrew Suffield [mailto:asuffield@suffields.me.uk]On Behalf Of
Andrew Suffield
Sent: Tuesday, January 18, 2005 2:13 PM
To: debian-legal@lists.debian.org
Subject: Re: SableVM/Kaffe pissing contest (Was: GPL and Copyright Law)


On Sun, Jan 16, 2005 at 11:18:30PM -0500, Michael Poole wrote:
> Andrew Suffield writes:
> 
> > On Sun, Jan 16, 2005 at 08:14:32PM -0500, Michael Poole wrote:
> > > The rest of your post is either intentionally or incompetently
> > > misleading, since Java's idea of binary compatibility means that a
> > > compiled Eclipse package does not contain any copyrightable
portion of
> > > the class libraries that provide declarations to the compiler.
That
> > > is what determines whether the binary package is a derivative work
of
> > > the class library package.
> > 
> > That's not entirely true. The binary package is a derivative work of
> > the class library package if:
> > 
> > (a) it contains a literal creative part of the class library,
> >     or a derivative of such a part
> > 
> > or
> > 
> > (b) it contains a literal creative part of the java source or a
> >     derivative of such (pretty much a given or the compiler wouldn't
> >     be much use), and the java source is a derivative of the class
> >     library
> 
> Does (b) refer to the use of features that are specific to one
> implementation (or at least one license)?  Or do you mean something
> broader?

It means anything that a court would consider derivation. The stuff
about using features is just a convinient rule of thumb.

> > About the only thing I've seen that will do (a) is static linking in
> > an ELF object, or anything comparable. (b) is the one that we
normally
> > deal with in Debian.
> > 
> > [Always remember: derivation is a transitive relation. If a is
derived
> > from b, and b is derived from c, then a is derived from c]
> 
> This is not true.  The parts that make A a derivative of B may be
> disjoint from the parts that make B a derivative of C.  (When those
> works are virally licensed, the license is transitive.)

It's still true, you've just introduced an aliasing error. Set the
resolution to 'lines of code', not 'packages'.

-- 
  .''`.  ** Debian GNU/Linux ** | Andrew Suffield
 : :' :  http://www.debian.org/ |
 `. `'                          |
   `-             -><-          |



Reply to: