[sf-lug] more re Novell/MSFT and GPL

[Rick Moen]

Quoting jim stockford (jim at well.com):

> per government code, there is a lot of code written by government
> employees, not contractors.

Yep, you're right that the former is automatically public domain.
(Of course, other people can then further modify that PD work, and their
code contributions _would_ have copyright title.)

> per patent vs copyright, i don't know how things have been resolved
> recently: there was seeming confusion over whether software is a
> written expression of something or an algorithm controlling a machine.
> What's the latest?

Your question assumes a mutually exclusive choice, where there in fact
isn't any.

Let's consider, oh, writing a Web browser.  

Case 1:  Your name is Tim Berners-Lee, you're at CERN in Genevea, it's
the year 1989, and you're using a NeXT workstation.  You write a
hypertext editing and viewing program for NeXTStep (a 4.2BSD fork with a
nice graphical desktop based on Display PostScript).  You call your
hypertext program "WorldWideWeb"
(http://en.wikipedia.org/wiki/WorldWideWeb).  Per contract, CERN gets
ownership of the copyright over your Objective C source code, and over
any potential patent rights.

CERN _could_ seek patent rights in various countries, if they could
successfully argue that some techniques demonstrated in your work
were sufficiently innovative (compared to prior hypertext techniques
such as Xanadu, Hypercard, and others) as to be truly new and have no 
prior art elsewhere.  (In the real world, CERN didn't do this.)

CERN's copyright title (ownership) would prevent others from carrying
out unauthorised copying of the WorldWideWeb for NeXTStep application,
or of creating unauthorised derivative works using its codebase.
However, in now way could it prevent others from writing wholly
independent implementations of the same idea, e.g., Mosaic, Cello, and
eventually Netscape Navigator and others.

_If_ it had successfully patented particular of Berners-Lee's techniques
exploited in that browsers, CERN could have forced third-party authors 
to either use different techniques in their implementations, or pay
patent royalties.


Case 2:  Your name is Jim Stockford (or possibly "jim stockford"),
you're in San Francisco, it's 2006, and you're using x86 Linux.  You
write a from-scratch Web browser that you then call NewWideWeb.

As author, you automatically own copyright over your NetWideWeb source
codebase, and any derivatives (including compiled binaries), just by
inventing and writing it down.  (There's an optional step where you
register a copyright at Library of Congress, which merely buys you
stronger enforcement rights.)

You _might_ also try to apply for one or more patent, if you seriously
think that, in 2006, your Web browser is actually doing something
innovative even after 17 years of other people's efforts -- but frankly
this doesn't seem likely, unless it's a really eccentric sort of Web
browser.


> per government code re patent, seems to me in the case of patent
> issues that existing government code would or at least could
> constitute prior art or design and play a part in patentable
> determinations, yes?  no?

_Any_ documented third-party implementation of a specific useful
technique prior to someone's attempt to patent it would constitute prior
art, yes.