[Patents] "fork vs leaves" or "worlds 1, 2, 3"

[orre@nada.kth.se (Roland Orre)]

On Thu, 10 May 2001, PILCH Hartmut wrote:
> > It seems as you have not really understood about the tree
> > concept yet, I will try to make that somewhat clearer.
>
> I am not sure whether I understood it.

Yes,  I think we still have some slight discrepancy. I have tested
this structure on a set of collegues, mathematicians mostly,
today, and most of them grasped it almost directly. Tonight I'll
discuss with a set of computer science collegues. The model still
need some refinement. The goal is to make it as crisp as possible so
anyone can understand it.

> Patents of true inventions (i.e. innovations in applied natural sciences,
> involving new insights in physical causality, pertaining to material
> embodiments) also often come in genealogies that look like tree
> structures.

Yes, I agree, the "hardware" and the "software" tree are both trees
as I see it, but the fork sprouting is very much different due to
the software world is a "perfect world". In the software world
you can easily create as many abstraction levels (like odontological
levels) as you need, which is also standard for software designers.
When you deal with the physical world, you can not work on several
abstraction levels easily.
(As a simple illustrative example: you could e.g. simulate a bouncing
 boll on an analogue computer which would correspond to two levels
 of abstraction. The physical simulation will usually be rather
 rude, but in the software perfect world, this is trivial and
 standard)

> You have one "basic patent" and many patents on some further advances
> around the basic patent.  A common clichee is that the forks (basic
> patents) will belong to western companies, while the leaves (or subforks)
> will be the result of a Japanese counter-strategy that allows the Japanese
> to get free or cheap licenses on the basic patents.
> In the above-stated clichee case, the patent on the fork actually did not
> prevent development of leaves but rather incited it.

This can only be done with physical designs. The design principle
still allows for wide rearranging of the actual physical structure,
making many leaf patents possible from a specific design.

> This doesn't mean that this is a great mechanism for inciting innovation.
> Machlup has some good arguments on why this is wasteful seen from an
> overview perspective.  But I don't think that these arguments allow you to
> set up a rule saying that fork patents are particularly harmful while leaf
> patents are ok.

Yes, but what I wanted to illustrate is that for software patents
they are even more severe as they actually cut off that fork. One
software design can not be the basis for other software designs
without including the "basic fork".

What need more ellaboration here is how sub forks grows. I will
discuss that tonight.

> Many of the fork patents (basic patents) actually have a particular R&D
> merit behind them.  I was recently dealing with a base patent that had
> disclosed a new magnetic material with a magnetic density of a few 100
> times that of anything known before.  It had taken quite a lot of
> experimenting to find that, and the ensuing leaf patents were
> comparatively trivial.

My view is also that hardware leaf patents are often trivial when
you have the basic design but considering software to make a
subfork is even obvious, that's what software designers and
programmers do all the time, by combining previously known concepts.
I will ellaborate more on how these obvious forks are generated
later. For me it's obvious, but intuitive. I have to analyze how
these subforks are generated, but I think I would have no problem
writing an algorithm for that. I'm actually working on a draft
of such a principle.

> Moreover, I have the impression that the "fork vs leaf"  argument applies
> to software copyright as well as software patents.  By making software
> free, you enable it to become a fork instead of only a leaf.  The "open
> source" can become a "living source", while proprietary software is
> crippled and will usually die immediately if the marketing department
> decides so.  Information has the potential of being a living social
> structure, while neither material embodiments nor natural laws have that
> potential.  Both are determined dead ends.

I agree!

> In Karl Popper's terms you would say that information evolves in "world 3"
> while material embodiments and natural laws belong to "world 1".  Patent
> law has traditionally excluded worlds 2 and 3.  In the terminology of
> the German Federal Court of 1969-81 they were formulated as
>
> world 1: world of things (Welt der Dinge)
>          controllable natural forces (beherrschbare Naturkraefte)
> world 2: interspersed activity of the human intellect
>        (zwischengeschaltete menschliche Verstandestaetigkeit)
> world 3: rules of calculation and organisation
>        (Rechen- und Organisationsregeln)
>
> worlds 2 and 3 were also lumped together as "world of the mind" (Welt des
> Geistes).

Yes, this is exactly how I see it.

> Today, patent inflationists lump worlds 1 and 3 together as a world of
> objectivised solutions, i.e. "practical and repeatable solutions"
>
>       Mark Schar 1998 "What is Technical?"
>       http://swpat.ffii.org/vreji/papri/grur-schar98/

I didn't find that page, was it possibly
http://swpat.ffii.org/vreji/papri/grur-schoelch01/
you were referring to?
A page I considered to be absurd, even though I didn't understood
every word.

> They fail to see that world 3, although objectivised and no longer part of
> the subjective human mind, is even more alive and faster moving than world
> 2.  It makes no sense to just exclude world 2 and patent everything around
> it.

Yes, and with the idea generation algorithm I have drafted, this
world (3) can actually start moving so fast that the idea of patenting
something from this world would just be impossible.

But any software patent would severly restrict the number
of possibily outcomes. Any software patent would severly cripple
the set of ideas generated.

It's actually so severe that even just one single software patent,
if it's the correct one, used in the most harmful way, could make
the whole idea impossible to proceed with.

So, the whole idea of software patents is thererfore insanely absurd.

This is clear for most computer scientists and mathematicians,
but as the concept is so abstract it need to be expressed as
crisp as possible for other people.

        Best regards
        Roland Orre