[Patents] Third Paradigm between Patent and Copyright is the 1844(!) french patent reinvented by Hatrmut

[PILCH Hartmut phm@a2e.de]

> I rewrote the Questionnaires for French industry A (small company) and
> answered

thank you. =20

Are you speaking on behalf of Novatek, Estelfree or rather some imaginary
non-software company?
=20
> How would patents affect your company?
> medium
>    1. Have you ever been approached or attacked by a patent owner for
>        infringement of his patents?
>  yes, american or German competitor
>     2. How many of the 400000 European patents could be relevant

OK, I should have written "of the >20,000 European software patents"

>        for your field?
> say 2000
>     3. How many of these do you own?
> say100
>     4. How much money do you expect to spend on
>          1. patent research
>  15000 FRF per year
>          2. patent applications
> 50000 FRF per year
>          3. patent license fees (subtract the fees you expect to earn
>             yourself)
>   5000 FRF per year
>          4. patent infringement lawsuits
>  250000 FRF per ten  years if I am the patentee
> less if I am the infringer because I modify the infringing object for non
> infringement
>        5. How many patent specialists are working in your company now?
> zero

> 2. Whether and HOW computer programs should be ENFORCED depends mainly on
>        the effects this would have on the European economy and society.
> yes , the States have a lot of compulsory licences schemes in their natio=
nal
> law for that

I agree that the idea of compulsory licensing should be revived.
But it is hard to formulate clear rules for compulsory licensing, and
there has been a tendency to do away with this system or use it only
in big cases where anti-competitive behavior comes into play.
Such behavior is however very common, almost daily routine, in the
software industry.  Using compulsory licenses to guarantee
interoperability would require too much bureaucracy.=20
Currently there seems to be no convenient way for anyone to obtain a
compulsory license. =20
=20
>        These effects have not been sufficiently studied.
> N/A
>  Europe should not change the law this year
> yes
>  but allow itself a 2 year moratorium to further discuss the issue.
> NO not before a solution exists for a correct examination

That is also a good point.
=20
>     3. Computer programs should not be patentable
>          1. Software copyright is quite sufficient to protect the
>             legitimate interests of software companies
>     I do not know

Many people in the software business say that.

The reason is that most abstract ideas are easy to find but hard to
implement.  Thus it is copyright which protects the difficult part of the
process.

>      2. Copyright alone isn't sufficient either. A third paradigm
>             property scheme is needed to better reward innovation
>=20
>   that is only another kind of patents with a special license scheme
> the word paradigm brings no difference

As an object oriented programmer, I would also tend to have one parent
class (paradigm), from which the specialties are inherited by the child
classes (sub paradigms).

It is possible to integrate patent and logilege into one general class.
But still there is a need for two subclasses:  that of concrete-physical
(technical) and that of abstract-logical innovations.   Rights that would
work for the former will cause trouble when applied to the latter.

In the abstract-logical world we also have copyright.  What would be the
parallel of copyright inthe concrete-physical world?  Perhaps ownership of
physical objects?

If taken that way, we get one single paradigm for everything:

=09=09concept=09=09implementation
physical=09patent =09=09posession=09=09
logical=09=09logilege=09copyright

> > > > Soft Patents for Harware Logics
> > > >
> > > >    Innovative processes for whose implementation not more than a ne=
w
> > > >    computer program running on known hardware is needed, ARE ALREAD=
Y>
> > >    monopolised in those cases, in which they do not run on a universa=
l
> > > >    computer.
>  Thus, if a car maker does not choose to lay out his motor
> > > >    control device for customer-side reprogramming (making it more
> costly
> > > >    and less safe), he would probably prefer to pay moderate license
> fees
> > > >    to a patent owner.
> What is  "moderate" ?

defined by the weaker bargaining position

> Why should the patentee not monopolise the programmed apparatus?

He should only not monopolise it, if it is a universal computer that can
be reprogrammed.   Otherwise he would not be monopolising an apparatus but
the program as such.
=20
> > Legacy "software patents" could be enforced against hardware
> > implementations.
> please explain the "Legacy" "software patents"

Legacy: "granted by the EPO against the written law but no longer wanted no=
w".
software patents:  put in quotation marks, because the term is misleading.
  the real issue is patents on abstract-logical concepts rather than on
  solutions that involve a new model of natural forces to solve a problem
  posed by physical reality.
  Much "software" can be implemented as hardware.

> If the same mechanism is realised as software that runs
> > on a universal computer (including a user-reprogrammable automobile
> > engine) the patent would no longer be enforcable.
> yes, d=E9j=E0 comme =E7a
> this is the enforceability exception of "private or experimental
> use"(article L613-5 CPI)

no.
I mean non-enforcable against whatever use.

> > This kind of rule would be useful for dealing with legacy patents from =
the
> > EPO.  I don't believe it has much merit beyond its usefulness as a
> > compromise solution.
> yes, d=E9j=E0 comme =E7a
> but the EPO has today nopower regarding enforceability

You mean that French courts allow enforcement of software patents (I mean
patents on abstract-logical schemes which can be implemented as computer
programs on a known device, i.e. the innovative nucleus lies in the
computer program) against hardware devices (e.g. DRAM chips) but not
against computer programs?
=20
> > > >    The disclosed logilege description comprises a reference
> > > >    implementation
> > >
> > > as used to be the case for patents, if I remember correctly
> >
> > patents should contain implementation examples, but in the case of
> > software patents, this rule is not applied.
> do not understand

I have never seen a software patent that contained in the "implementation
examples" section a full computer program source text.  This is contrary
to the spirit of the law, which requires the implementation section to
fully enable the man skilled in the art to rework the process.
=20
>   It seems that the patent
> > offices are afraid of being consistent, because consistency in this poi=
nt
> > would expose the basic inconsistency of patenting software.
> >
> do not understand "the basic inconsistency "  and how it is measured

If you fully enable the software engineer to rework the process, then you
must supply a reference implementation.  But by doing that, you already
provide the "invention" itself, and this violates the patent claims, at
least if a "computer program" or "computer program product" is claimed, as
is the case with EPO patents since 1997.
=20
> > > I think the patent system should work like that, possibly with longer
> > > validity.
> yes, d=E9j=E0 comme =E7a

non, pas du tout.
e.g. no reward is paid for a successful opposition.
almost none of the characteristics I listed is comme ca in the patent
system.

> > The following logilege features may not be applicable to patents:
> > (a) 3 year validity

> for free ??

Yes.  The privilege owner should pay those who falsify his claims.
The longer he keeps the claims, the greater the risk that someone comes up
and falsifies them.
It is wrong if the proprietors of good claims are forced to finance an
examination system, and it is even more wrong to create incentives for
a big bureaucracy to inflate the patent system.
The logilege doesn't even need a formal examination.
Court fees would be paid by the loser.=20

> 20 year for patents but you pay each year an increasing tax

This only means that you are forced to commercially exploit your idea.
But abstract ideas often cannot be commercially exploited.
So only owners of huge portfolios can play poker with such items.
=20
> > (b) scope of claims must be narrow enough to enclose neither prior art
> yes, d=E9j=E0 comme =E7a for patents
>=20
> >     nor possible future innovations
> this is the problem
> what can you do to enforce the logilege in that case ?
> the infringer is "only a future innovator" all the time

Why is that?

The typical infringers whom you want to keep away is the one who just
imitates your innovation without improving or enriching it.

You would have to either disclose full information on all ideas needed to
make your claims work, or you would have to narrow down your claims to
suit the little information that you are ready to provide.

This kind of provision is helpful for avoiding large claims like that of
the IBM spellcheck algorithm, which is really a generic class of possible
algorithms which all require great inventive activity to be put to work.
IBM on the other hand side wasn't particularly inventive.  Only their
patent lawyer was, because he managed to derive an unusual super-class
from a particular algorithm.

The claims I am referring to can be found in

=09http://swpat.ffii.org/vreji/papri/bpatg17w6998de.html

in German.

This kind of problem is typical for abstract-logical innovations.  The
purpose of such innovations often lies in the area of abstraction.  One
way to avoid overly broad claims is to require the applicant to disclose
ideas needed to put those claims into practise.
=20
> > (c) limitation by right to publish and right of interoperability
>=20
> yes, d=E9j=E0 comme =E7a for patents
=20
Unclear.

Patent lawyers regularly claim that the publication of a program in the
Internet already infringes on their patent, and they even write their
claims like that.

If a program is a patentable object like any other, then making it
available to others constitutes an infringement.

If it is not, then there is no reason to change Art 52.
=20
> > (d) inadmissibility of posterior narrowing of claims
> yes, d=E9j=E0 comme =E7a for patents

patent claims are frequently narrowed.
Most proceedings end with the acceptance of a set of auxiliary claims by
the court. =20

> but will probably change under american pressure

I'd be interested in more details.
Is France different?

> > Of course it could be wise to apply b-d to the whole patent system.
> yes, 75% d=E9j=E0 comme =E7a for patents
>=20
> > There is no reason not to apply c except that it normally is not a
> > problem with traditional patents.
>=20
> yes, d=E9j=E0 comme =E7a for patents
> there is a french ruling of court applying c

I'd be interested in details.=20
=20
> > Then there would no longer be a third paradigm but only an improved pat=
ent
> > system with a reduced validity period for abstract-logical innovations.
> >
> there is no problem because drugs enjoy an extended validity period

TRIPS Art 33 says that it must be 20 years or more for all patents.
=20
> > But reforming the patent system in this way has been made impossible by
> > TRIPS, and after all it.
> then use compulsory licensing
> the States have a lot of compulsory licences schemes in their national la=
w
> for that

Also to circumvent Art 33 of TRIPS?
=20
>  And even if everything is integrated into one
> > system, the fact that there is a recognized borderline between
> > concrete-physical and abstract-logical as well as between publication a=
nd
> > use of information already makes for a third paradigm.
>=20
> publication has never been infringing

Microsoft has forbidden the developpers of VirtualDub to publish their
software.  The Ghostscript source contains no LZW handling algorithm
because the programmers are afraid.  Perhaps they are just overly afraid.
We would probably have to fight quite a few exemplary cases in court
before we can be sure of that.

> > One more thing to do would be an inventivity competition:  the inventor
> > first publishes the problem and tenders for solutions.
>=20
> inventivity competition is a competiton and enjoys no right

It should, because inventivity is a vague concept which cannot be
determined by formal criteria to be handled by some bureaucrat (examiner).
Btw, this is also the problem with the compulsory license:  it leaves
to much room for the judgment of a bureaucrat based on difficult criteria.

> > For three months anybody can submit a solution.  The submitted
> > solutions

> what happens if the invention has been stolen?

We already have this problem in the patent system.  It is one reason why
the USA have a "first-to-invent" principle and a grace period.  Europe, on
the other hand, requires the applicant to exercise discipline and take the
responsibility of keeping his invention secret until the day of the
application.
=20
> > against which the logilege description, which is
> > then published, must be measured.
>=20
> what happens if the invention has been stolen?

The inventor sues the thief.

> the inventor "first(?)" publishes the problem and has to pay his thief ?
> very fair !

He publishes only the problem.
Every patent description has a "problem" section.
This could also be connected to a "prior art" section.
The innovator describes what had been known up to now and what was the new
problem for which he had to come up with something that prior art didn't
provide.

He would publish this section one month ahead of the publication of the
solution.  This one month period could be used by others to submit
solutions, either from prior art or self-invented.

When the Fraunhofer Gesellschaft came up with audio compression algorithms
that went into MP3, they might have applied for a logilege by specifying
the problem of "finding a way to compress audio data by a rate of 10 with
a quality loss of less than 1%".  It is very unlikely that anyone would
have found the FhG algorithm within one month.  However, it is almost
certain that thousands of people would have found the above-cited IBM
"invention" within one day.

This huge difference cannot be handled by the conventional criteria of the
patent system, but it can be handled by a competition game, where the
public replaces the examiner and a reward is paid to the winner.

The reason why it doesn't work otherwise is that innovation is cheap and
poorly documented in the abstract-logical realm.

-phm=20