Subject: Re: Examples needed against Soft Patents
From: Ben Tilly <>
Date: Thu, 30 Dec 2004 15:24:00 -0800

On Wed, 29 Dec 2004 22:27:26 -0500, Seth Johnson
<> wrote:
> Russell McOrmond wrote:
> >
> > On Tue, 28 Dec 2004, Ben Tilly wrote:
> > >
> > > Furthermore there is nothing intrinsic in the reasoning for why we
> > > have patents that suggests it should only apply to hardware.  The
> > > fact that it does can be seen as just an accident of history.
> No, it was an essential precondition for the establishment of
> patents in the first place.  Though putting it in the relative
> terms of "hard-" vs "soft-" ware is not properly informative with
> respect to what the distinction really is.
> >   It isn't about only applying to hardware, but the need to check for each
> > subject matter class whether the first principles that justified patent
> > policy still hold.
> >
> >   a) will it provide incentives for innovation, or stifle innovation.
> >   b) is the embarrassment of a patent greater or lesser than the potential
> > loss of an "invention" due to secrecy or other lack of disclosure.
> This is analysis that we use in discussion with policymakers
> until they can be convinced to return to more basic first
> principles that used to be ordinary common sense.  In actual
> fact, software is in a pure category that is appropriately simply
> excluded from patentability at the outset, not according to
> economic analyses or tests of novelty, etc.

That is a strong claim.  Can you document it?

More, can you prove that the categories would have been
drawn like you want if software had existed historically?

> Yes, but it's not the response that people are listening for.
> The issue is not "hard" vs. "soft" -- it's concrete vs. purely
> abstract.
> It's important to understand the meaning of "pure" -- pure
> abstractions are abstractions that hold regardless of empirical
> instantiation.  Logic and math are both pure bodies of knowledge
> that are utterly independent of empirical instantiation.
> A patentable process may represent certain physical law, but it's
> the discovery of a new concrete way to use those physical laws
> that's covered, not the physical laws themselves.
> The point about software, however, is about *pure* abstraction.
> How do you address this with respect to what Ben has asserted?
> It is somewhat simple, once you think it through.
> Software is concrete in the wrong sense for what we're
> considering here.  It is concrete like written words are
> concrete, not like a natural causal process is concrete.
> Software, or more precisely, a logical algorithm, is not concrete
> in itself.  Logical algorithms just happen to be expressed in
> concrete form, just as words on a page represent ideas that are
> not concrete, in a concrete written form.  That is, software is a
> representation of pure abstraction, just as written
> representations of logical and mathematical principles are.

This is a nice theory.  But I don't see it supported by history.

For a random instance, US patent 1,647 issued in 1840 seems to
be about something that's pretty abstract.  (That's the patent
on Morse code.)

> This is why I think a simple policy solution for the problem of
> software patents would be a simple requirement that patent
> applications include a statement that the process claims do not
> include abstract instructions provided to a generic logic device,
> and if they do, those abstract instructions are not covered.
> That would do the trick pretty handily.  This would make it
> really easy for examiners, and would provide simple means for
> correcting any mistakes that occur, by stating up front that
> anything that can be shown to be instructions for a generic logic
> device are not patentable.  That sort of policy would result in
> us not having courts trying to draw distinctions without
> sufficient principles expressed in the law, and feeling compelled
> to address matters largely with respect to the interests of the
> direct parties in a suit.

I'd like that law.

> > > I do not believe that I started or ended with the conclusion that
> > > patent lawyers wanted a new type of monopoly.
> < SNIP >
> > > It is true that the line of reasoning was intended to result in a
> > > specific conclusion (that if you accept patents, there is no obvious
> > > reason to say that you shouldn't have software patents), but the
> > > intention has no bearing on whether the reasoning is valid.
> Ben, you have accepted a change whereby patent lawyers have
> started asserting numerous misguided ideas that basically come
> down to the notion that what's novel, nonobvious, etc. must be
> patentable.  But in fact certain things were excluded at the
> outset, regardless of how innovative they might be.  If we had
> not excluded pure abstraction, patents would never have been
> established in the first place.  A king giving one person an
> exclusive right to a pure abstraction would have been such an
> affront to plain reasonableness, that that would have been just
> the most direct demonstration to all those privileged enough to
> be engaged in any form of science, of that king's tyrannical
> character.

I'm amused at this characterization of history.

FYI, both patent law and copyright law started with various
kinds of nobility handing out various kinds of patents.  Many of
these patents were on things that were quite abstract, like the
right to publish a particular book for a period of time.

As for whether giving patents on abstract things would be
cause for rebellion, there's a lack of perspective.  This is a
period of history where nobility could (and did) execute people
just because they disagreed on, say, religion.  And many kings
(and queens!) did.  Review why, for instance, bloody Mary got
her nickname.

Compared to that, a few unfair patents were small potatos.