Re: [Patents] "Further technical effect"

[James Heald <j.heald@ucl.ac.uk>]

Xavier Drudis Ferran wrote:
> 
> On Fri, May 16, 2003 at 06:29:11PM +0100, James Heald wrote:
> >
[snip]

> 
> > I also think it would be worth explicitly closing the door to make sure
> > that there is no possibility that the argument could ever be produced
> > that a new software technique which allowed a computer to (almost
> > always) solve more quickly in the context of a particular problem might
> > ever be interpreted as a 'new teaching about cause-effect realtionships
> > when using controllable forces of nature'.
> >
> 
> How would you twist it into even looking like such a teaching?.  The
> thing that makes a computer useful is that all cause-effect
> relationships in using controllable forces of nature are already
> solved by the hardware, which presents the programmer with a formal
> model of calculation that is safe to assume that will work, however it
> does. This allows innovation to focus in mathematics, logics,
> organization instead of observation, experimentation and material
> prototyping (therefore innovation is much cheaper), and makes more
> complex systems possible.
> 
> A "software technique" can never provide 'new teaching about
> cause-effect realtionships when using controllable forces of nature',
> in the same way that a poem can never be a weapon to kill someone.

Okay: here's how a judge that really *wanted* to grant a software patent
might parse it:  the computer is a system which is an application of
"controllable forces of nature"; the program is a cause; the result of
the program as mediated by the computer is the effect; and if that
effect was more than just "the program completed", ie if instead it was
"the program completed *more quickly*", this could be seen as an effect
above and beyond the normal functioning of the computer.

Note that I'm not saying this is how the test *should* be interpreted --
it definitely isn't.  But I'm saying it is how the test *just might* be
interpreted, especially by a judge in a national jurisdiction where the
history of the test was not familiar, and especially if that judge
wanted to justify a particular result.

It is heartening to know that under German caselaw an increase in speed
or efficiency is not enough to differentiate a "computer-implemented
invention" apart from a "computer program as such"; but currently under
UK caselaw it is.

One might also (just about) argue, even if the program was known, that
if the exact data was not known in advance, then the efficiency (eg of a
data-compression system), or the speed (of certain numerical
algorithms), might still be a matter for experimentation.

> 
> So I think this is the most clear and straightforward requirement that
> a law can make. Now, if what you mean is that whatever the law says,
> the EPO may plainly ignore it, then I'll have to admit it looks so from
> history. If  you mean that the courts may also ignore anything in the
> law, I'll have a harder time to believe, but who knows. Having  a clear
> law does not solve the problem of ensuring the law  is followed, but it
> must help a little, otherwise we'd have more swpat litigation in Europe
> and nobody would be pushing for this directive.
> 
> In other words, I think the counterproposal in ffii.org is necessary,
> although not sufficent. I tend to think your proposal is not necessary
> nor sufficent, but it is no harmful at all, and not incompatible.



> > So I think both clauses would add something new and worth having to the
> > Echerer et al Amendment 32 and Gebhardt's Amendment 33.  (Though I
> > accept that the first clause could perhaps be rewritten more sharply).
> 
> The new uses of forces of nature stuff is the sharpest way to rewrite
> it that I can think of, but I may be overlooking something, of course.
>

Just to clarify: it was *my* first suggested clause that I admit may not
be particularly good and could stand re-writing, rather than Echerer's.

I do still think that as well as stating the abstract principle
("controllability / forces of nature"), it would be a good and useful
thing to have clauses in the directive recitals which make absolutely
clear what that principle implies for a non-exclusive list of certain
more practical down-to-earth examples.

I would particularly like to see on-screen user interfaces (or the
"methods of interaction between human beings and computers" as the
professors put it) explicitly declared not patentable.