That anti-patent pamphlet I mentioned

Arnoud Galactus Engelfriet galactus at stack.nl
Fri Dec 13 13:35:16 UTC 2002


xdrudis at tinet.org wrote:
> You  are basically saying, to use an analogy from the FFII, I think, 
> 
> 3 EUR + 3 EUR = 6 EUR 
> 
> is not patentable, because it handles no technical data 
>  
> 3 l of fuel + 3 l of fuel = 6 l of fuel 
> 
> is patentable because it is technical. 

Not really, but it comes close. The invention would be something
like reducing the amount of fuel in a motor, which is technical,
and this reduction is achieved by having a program compute the
amounts in two separate fuel tanks or something. 

> This is impossible for me as a programmer to understand since the 
> exact same program perfoms both functions. If you prohibit (monopolize)
> one you prohibit the other, and anything that solves the problem
> would be usable for fuel and therefore infringe. 

The fact that it's done by a program is irrelevant. It would
be the motor with its control means that would be the invention.
And you can't prohibit what's outside the claims, so using
the same solution for money reduction is not infringement.

> The law they pretend to harmonize covers the same matter and territory 
> (or subset of it) than the law the EPO should be following. If the 
> EPO was following the EPC, there would be no need to wellcome
> harmonization of EU law with the EPC. 

The EPO thinks it's following the EPC, but not all national courts
and laws agree with the EPO's thinking. The EU believes the EPO
is doing the right thing and is now trying to harmonize the laws
of the Member States.

> > Well, as far as I understand it, everybody seems to be in agreement
> > that inventions should have technical character. The differences
> > occur over the definition of when something is technical. Nobody
> > wants purely non-technical software to be the subject of patents.
> 
> Sweetman, a british SME manager asked at the EuroParl for an example 
> of a program that could not be constrained to have a technical effect
> under EPO doctrine because he had tried for 6 months to understand it 
> and did not see one, and nobody answered, including 2 EPO officials. 

See EPO BoA decision T 931/95, the pension benefits case.
http://legal.european-patent-office.org/dg3/biblio/t950931ex1.htm
Even though the method claim used software, it was not an
invention.

> The cases you present are not of unpatentable programs, I would say,
> but of applications worded in the wrong way for the EPO to grant the 
> patent. Another wording for the same thign could be patentable.

It is my belief that a clever attorney can always redraft a
claim to incorporate a technical effect, but then the claim
will usually be much more limited in scope. It would be
restricted to a particular implementation for example.

> My interpretation of the EPC is not the only possible, but until 
> someone shows me a _consistent_ intepretation of the EPC that 
> justifies EPO practice, I am entitled to belive it is unjustified 
> by the EPC.

EPO practice is to say that if an invention as a whole exhibits
a technical effect, it is patentable (although you must also
check novelty and inventive step). It is irrelevant whether 
software is part of the claimed invention. However, because then
all attorneys will claim "software interacts with hardware and
so it's always technical", they came up with the further technical
effect that must be more than ordinary interaction.

Basically, if you ignore the fact that a step or feature is
realized with software, and you still have a technical effect
then you have a patentable invention.

> > The Board of Appeals is entitled to provide an interpretation of the
> > EPC. What you seem to say is that their interpretation of the EPC
> > violates the EPC's original intent. This may be so, but since they
> > are the final arbiter, it is hard to point out what they are doing
> > wrong.
> 
> It violates the EPC letter. THey follow the EPC where it says 
> they are to decide, but don't follow it where it says what is patentable.

They seem to think it conforms to the letter.

> > Sure. Personally I have always felt that the only rational way to
> > interpret this provision is that it excludes patents on the computer
> > programs themselves, but not on for example a machine programmed with
> > a particular piece of software. If the behavior exhibited by the
> 
> This sounds absurd to me. A patent prohibits use of an "invention". The
> main use of software is by programming a machine with it. 

Yes, but the patent would prohibit the using or selling of
the programmed machine. Not the act of programming.

> If you can't patent the program but can patent any machine programmed 
> to use it, then nobody can use a machine programmed with the program, 
> so nobody can use the program, so the effect is the same as patenting 
> the program, except for marginal uses (as an expressive work, or 
> as mere transmitted ifnormation). 

Exactly. But since it was believed in 1973 that software is
not a "thing" (just like a blueprint isn't a "thing"), they
made it clear it could not as such be patented. Just like you
can't patent a blueprint, but you can patent a machine built
in accordance with the blueprint.

> If waht you propose is not patenting _any_ machine programmed 
> with the software, then it is only fair to apply novelty and inventive 
> step to the machine, not the software. 

The tests should be applied to the programmed machine, not the
unprogrammed machine. The software *in the machine* makes the
difference.

> > machine is patentable, then it should not make a difference whether
> > the behavior is caused by hardware or software. But making and
> > selling of the program separate from the machine should not be
> > patentable, just like making and selling a blueprint for the machine
> > should not be covered by patents.
> 
> But if the only innovation is in the software, the machine should 
> not be patentable. 

That is where we differ. I believe it is illogical to say "the
innovation is in the software", because using software is not
an innovation. Making a machine do a novel thing is an innovation,
no matter how you make the machine do it.

> I have no problem patenting analogical sound filters 
> but once you discover you can apply a Fourier Trnasform and do that in 
> a DSP, you have already solved a whole range of problems. Maybe the 
> DAC or ADC converter could be patented, but the math and logic you then 
> apply to the signal is no different from business logic, and should 
> not merit a patent. Programming a filter in a DSP is just the normal 
> use of the DSP, and therefore not inventive.

A particularly programmed DSP works differently from an
ordinary DSP. Why is it an invention if the programming is
done with hardware chips, but not if it is done with software?

> > > approach of the 1978 EPO guidelines based on the same EPC (at least
> > > art 52 is unchanged, IIRC)
> > 
> > The EPO Guidelines are binding to no one except examiners. The
> > president can issue new Guidelines whenever he feels like it. I don't
> > think this is a very strong argument.
> 
> My argument is the President can not decide tomorrow that the EPO 
> is going to grant patent of sexual intercourse positions. 

No, but he can instruct examiners that such methods are to
be examined in a particular way. Just like he did with business
methods and software.

> What the 
> president decides must respect the EPC. Since the guidelines of 
> 1978 respected the same EPC as we have now, reinstating them would 
> be consistent with the EPC. More consistent than current guidelines.

Surely it is possible to update your guidelines if you have
a new insight in how the EPC is to be interpreted? Furthermore,
the prez has no choice. If the BoA says it's to be one thing,
the examiners should adhere to that otherwise applicants are not
treated fairly. And so the examiners should be instructed to
do what the BoA said.

> And I think it is important to have the EPO practice, the EPC, the EU,
> and national 
> laws in the same line.

I agree.

>  Now it is no because the EPO does not follow the
> EPC. The solution if changing what the EPO does, not the rest. 

That's one solution. Changing national laws so it follows the
EPO's interpretation of the EPC is another. And it seems
that's what the EU wants.

Arnoud

-- 
Arnoud Engelfriet, Dutch patent attorney - Speaking only for myself
Patents, copyright and IPR explained for techies: http://www.iusmentis.com/



More information about the Discussion mailing list