From xuan--swpat@baldauf.org Fri, 04 Aug 2000 12:27:23 +0200
Date: Fri, 04 Aug 2000 12:27:23 +0200
From: Xuan Baldauf xuan--swpat@baldauf.org
Subject: [Patents] Re: Freesw Re: Patents DeCSS, source code and software
 patents


Philippe.AIGRAIN@cec.eu.int wrote:

> Xuan,
>
> Indeed, if the reference is allowed to be a "generic computer"
> or if even worse one considers that "a software represents its
> execution on a generic computer", then the danger you are
> highlighting is well real.
> But one the other hand, it was classically accepted that if
> you specify a precise device or physical process as an
> execution environment (including all input/output devices
> and contents) and if the patent protection scope is granted
> only to that device or process and does not extend for
> instance to the execution of the same software in another
> execution environment, then patentability can be reasonable.
> What do you think?

I think that this won't work. For example, if the EPO chooses to not let you
patent the MP3 algorithm, but to patent it running in a Java Virtual Machine on a
hardware MP3 player you can carry with you (like a Walkman), the patent would be
worth much or nothing, depending on how far the interpretation of "execution
environment" is. Some competitor might build also a Walkman-like MP3 player, but
not internally building on "Embedded Java", but "Embedded Linux". Now you have
another execution environment. Does the clone infringe the patent of the original?

You could say yes, because the MP3 algorithm is the same. You could say no,
because the "execution environment" is another one. But you also could say yes,
because despite of having another micro operation system, you could use exactly
the same hardware. You could say no, because despite of the fact that every MP3
algorithm has the same principle in mind, they are likely to differ in detail.

And "execution environments" if you mean something like a virtual machine with a
defined command|API set, are transformable into each other. Once upon a time, you
could emulate MS-DOS on an Amiga computer, and now you can emulate an Amiga on a
generic personal computer (using DOS or Windows or Linux...). For the many JavaVMs
out there, it's the same. If you have some Java software and you patent it along
with the JavaVM and Windows98 and a PentiumIII Processor, does an implementation
on Athlon infringe the patent, despite of the changed execution environment?

Because every general purpose "execution environment" is transformable to every
other general purpose "execution environment", you are giving the patent judges a
wide choice between "A transformation does not infringe the patent." and "Every
possible transformation is a patent infringement". The former would be the better
one, but useless for people who patent algorithms (which are capable of being run
on a general purpose computer), the latter would lead to software patents. This
choice is to broad to be handled by the judiciary, because this would worsen both
sides: Software patenters would not be certain wether their patents are usable,
and software writers would not be certain wether the patents they possibly
infringe are can do any harm to them or not.

>
>
> Note that as early as 1948, in his famous conference given
> to a congress of psychiatrists, John von Neumann pointed out
> that the notion of an universal computer is not a practical
> one, that in the realm of real technology (that is material
> machines, not abstract ones like the Turing machine),
> there are no universal machines but only particular ones,
> and that technology, including the von Neumann's model
> of computers is about building the RIGHT particular ones.

There might be a lack of an universal|general machine, but there are many general
_purpose_ computers|machines out there, and the latter are software-technically
transformable into each other.

>
>
> Philippe

Xun.

>
> > PILCH Hartmut wrote:
> >
> > > [...]
> > > But one could for reasons of political expediency allow material
> > > embodiments like firmware or mp3 players to be patented, as long as it is
> > > clear that the software embodiment can not become the target of any
> > > claims.
> > >
> > > There are some areas in industry where people have become so used to
> > > patenting software inventions (such as a motor with an ABS control system)
> > > that they might be unhappy if these inventions are no longer patentable.
> > > And patenting them doesn't do much harm, as long as only the whole
> > > device and not the information is targeted.
> >
> > You know that this can open a can of worms? If software "as such" cannot
> > infringe a
> > patent: fine. But software "as such" is always useless for us. We need a
> > computer
> > processing the software to make use of it. If someone says "Gee, I do not
> > patent
> > some software, I patent some software in combination with a general purpose
> > computer". So reimplementations may be written, but not used, because their
> > usage
> > will always be in combination with a computer. This would be unbearable.
> >
> > Xun.
> >
> >
> >
> >
> > _______________________________________________
> > Freesw mailing list
> > Freesw@conecta.it
> > http://mail.conecta.it/mailman/listinfo/freesw
>
> _______________________________________________
> Patents maillist  -  Patents@liberte.aful.org
> http://liberte.aful.org/mailman/listinfo/patents







