Re: [Patents] meaning of "program"

[Hartmut Pilch <phm@a2e.de>]

> > Software patents do cover a single program.
> >
> > E.g.
> >
> >     system and method using [ elements of a general-purpose computer
> >     ], characterised by that [ a certain combination of steps is
> >     performed ]
> >
> > This is a computer program, of which many implementations can exist.

> I don't know about legal meanings, but it is very contrary to the
> language I speak as a programmer. For me a program is what I write, If
> the person next door writes some other code that does the same, that's
> another program, and a software patent covers many.

You could also say, the above patent claim describes a class of computer
programs (just like a chemical claim describes a class of possible
chemical reactions -- claims are always about classes).

Are the DOS version and the Unix version also different programs?

Evidently a "program" is not a particular instance running on some
hardware, but an abstraction of such instances. And the art of programming
is the art of abstraction.  Flow charts, algorithms etc are just as much
programs as are final instances on a diskette.

This was very well written in the EPO's examination guidelines of 1978 and
later, in 1985, removed.

> I would like more explanation on this. Why does a patent attorney want
> to make the word program (or any other) more narrow?.

If "program" means "particular instance", then nobody will ever be able to
violate Art 52.2c.  Art 52.2c would then say something like "you man not
get a patent on claims to one instance or to a narrow class of instances,
such as copyright already allows you to own, but you may get a patent on a
huge class of instances, so as to reward an innovation."  That way, Art
52.2c EPC can be made to appear as verbal rubbish.

This is not to say that without the 1978 Guidelines' definition of program
the article is really rubbish.

When you say "He patented that motor", you also don't mean a particular
instance but a class.

What you count as a discrete motor or program which is distinguishable
from "other" motors or programs, i.e. how you define the identity of a
motor or program, is dependent on some implicit rule of forming a class.
Coypright is one, patents are another.  The object of both is the same,
but the scope of what counts as one motor or one program is different.
When talking about patented software, you should think in the scope
defined by patents, not the scope defined by copyright.

-- 
Hartmut Pilch, FFII & Eurolinux Alliance              tel. +49-89-12789608
Protecting Innovation against Patent Inflation       http://swpat.ffii.org/
140,000 votes 400 firms against software patents     http://noepatents.org/