Re: [Patents] Which companies are against art 2a and 4.3a?

[Jonas Maebe <jonas.maebe@ugent.be>]

On 10 jun 2004, at 17:27, Flemming Bjerke wrote:

> Sorry, I did not expressed myself clearly. I meant that the electronics
> industry to a large extent produces products which transmit, presents, 
> store
> etc. information, e.g. mobile telephones and TV's. Perhaps their 
> reading of
> art 2a and 4.3a is: patenting of software that control the technique in
> mobiles and TVs, is not feasible according to these articles.

That is correct. You cannot patent any software according to those 
articles, regardless of what the software does. On the other hand, some 
invention embedded in a TV or mobile phone does not become unpatentable 
if it's controlled by software (but then you don't get a patent on the 
controlling software, but on the invention controlled by this 
software).

My personal example of this (I may have mentioned it before on this 
list, but I don't remember) is a weaving machine. Suppose you have a 
fully automated weaving machine, completely computer/software 
controlled. You then find out that if the knives of the machine are 
turned at a specific angle relative to the threads, you can cut those 
threads much more easily. However, all you have to do to integrate this 
(possibly patentable) invention in your weaving machine, is change the 
value of a few variables (and maybe add some mathematical formula to 
describe the path that those knives have to follow) in the controlling 
program.

That does not matter according to the EP's version of the directive, 
because your invention is how those knives are positioned relative to 
the thread, not that you use some computer program (or how you use a 
computer program) to do that. However, you do not get a patent on the 
used mathematical algorithm nor on the input parameters to it.

If there is no innovation present outside the mathematical/software 
part, then there is no patentable invention. And there are indeed a lot 
of patents in e.g. the mobile phone industry that correspond to this 
(e.g. on the smart typing of SMS'es).

> Although their
> purpose of using software is control of technique, rather than 
> computing,

A computer is also technique, you can't make meaningful distinctions 
this way. Those things all have an embedded cpu that executes those 
instructions.

What you are arguing for, basically boils down to trying to establish 
what separates the "normal physical interaction between a computer and 
a computer program" from the "further technical effect", and only 
allowing a patent on "inventions" which have this further technical 
effect. It's the current EPO doctrine, which has resulted in all the 
software patents we are stuck with now.

In fact, the EPO itself has argued that this "further technical effect" 
is only confusing, see
http://swpat.ffii.org/letters/cons0406/text/index.en.html#ed24@A2

I still think the only clear doctrine is to consider things described 
in software exactly the same as descriptions in any other form or 
language: if the described stuff would be patentable in described in 
plain English and when not executed by a computer, would it be 
patentable? If yes, great, if no: sorry, but simply translating the 
description of your innovation in a machine-readable form does not 
suddenly make it a patentable invention (the work you did while doing 
that is protected by copyright, and the rest is either a patentable 
invention or not, regardless of its form).

> they see themselved included. Such a reading will push them into the 
> arms of
> Microsoft, IBM, etc. And this enforces the resistance from big business
> against these paragrafs.

In all fairness, I also thought this for a while, but I don't think 
that's the case anymore. It becomes very confusing to argue in such a 
framework ("if I can patent this algorithm when run on a 50MHz cpu with 
256KB ram, then why not when run on a 2GHz cpu with 1GB ram"?) and 
no-one has yet figured out how to prevent "pure software" from being 
patented when using that kind of formulations (the "further technical 
effect" stuff obviously doesn't work, as the EPO has proven).

And then there's also the question what kind of extra investment in 
designing a predictive typing algorithm for a mobile phone warrants a 
patent monopoly, while exactly the same algorithm executed on your 
desktop PC apparently does not warrant it. The same for the deal to the 
economy as a whole: is it really good for the mobile telephony market 
as a whole if a company can monopolise that kind of things, if it's 
considered not to be the case for the "generic PC" market?


Jonas