Re: [Patents] more on information is physical, software technical

[James Heald <j.heald@ucl.ac.uk>]

Possibly cotributions to this sort of discussion are best left in the 
ghetto of alt.physics.loonies or wherever, but against all better 
judgment, here's twopence more that I can't resist adding.

I'm not sure it's particularly relevant to the question of patentability 
(as I'll come back to at the end), and which has been taken up in other 
threads.

But I'm concerned that with all this talk of "physicality of 
information", Greg is straying dangerously close to what E. T. Jaynes 
called "the Mind Projection Fallacy"  -- the mistake of locating 
probabilities, and functions of probabilities such as entropies or 
information (Sum p log p), as entities in their own right, somehow 
having an independent existence in the real world, rather than where 
they properly belong, namely in the mind of the analyst.

Facing up to the analyst-dependent Bayesian nature of probabilities, 
which Ed Jaynes did, as the first person to systematically present 
thermodynamic entropy as just a particular example of information 
theory, back in the 1950s, makes the whole intellectual structure a lot 
cleaner, easier and more coherent, just as the Bayesian approach does in 
statistics.

Rather than sweep issues about the nature of probability under the 
carpet, the Bayesian viewpoint makes it a feature - and thereby 
genuinely robust.

The point about the information, entropy and probabilities is that they 
are defined by the analyst's knowledge of, and model for, the system -- 
they are not an intrinsic part of it.


Quantum mechanics.

The probabilities in quantum mechanics are also most usefully seen as 
mental, rather than as part of underlying reality -- in this case, in 
addition to uncertainty about the true quantum state, probabilities also 
arise because of an insistence on using a model which has more freedom 
(a larger information content) than the true reality.  True reality 
doesn't have enough information to fix a state in the (false) classical 
 model of phase-space, so an unavoidable entropy remains: the result of 
the analyst's decision to use an over-complete model.


Patentability.

So, should the fact that in doing physics we use subjective probability 
and information theory to model the real world cause FFII any headaches?

I can't see why it should.

FFII has consistently put forward the proposition that a useful line (ie 
one with a high degree of legal predictability) can be drawn between 
innovations which exhibit "plan-conformant control of the forces of 
nature" on the one hand, and those which are purely abstract on the other.

A further line can be drawn where the "control of the forces of nature" 
 is limited to merely a data processing device carrying out its normal 
function of processing data, where the only innovation is in the choice 
of instructions being executed.  (With whether the said choice of data 
processing instructions either has a practical utility, or increases the 
speed of processing or reduces the heat dissipated, or not, being quite 
moot).

As I understand it, FFII's position is that such a position makes for a 
very clear line, which courts could interpret quite predictably.

The notion of "plan-conformant control" seems to virtually require some 
model of reality -- if such a model uses probability, entropy or 
information theory, it seems to me this is neither here nor there.

On the other hand, if there is no control of nature beyond the 
calculation, then it clearly falls into the second class.

I can't see that there is anything here to blur the line at all
-- what's your point, Greg ?


Cheers,

   James Heald.







Hartmut Pilch wrote:
> On Mon, 5 Jul 2004, Gregory Aharonian wrote:
>
>
> >    From May 5th to 9th, a conference of physicists and philosophers (i.e.
> > people who understand the "physical" better than politicians) held a
> > conference in Minnesota to debate the foundations of quantum mechanics.
> > (An account of the conference is in the 25 June 2004 issue of Science,
> > page 1896-7).  Quantum mechanics (and its decohered classical mechanics)
> > underlies everything that is physical.  Applications of these mechanics
> > is as technical as you can get.  Descriptions of these applications is
> > as patentable as you can get.
> >
> >    So what was one of the main topics of this group of scientists?  I
> > quote from the article:
> >
> >    ..., participants at this year's Seven Pines Symposium are
> >    taking a different tack entirely.  Spurred by recent efforts
> >    to develop quantum information technologies, thinkers such
> >    as Jeffrey Bub, a philosopher at the University of Maryland,
> >    hope to prove that QUANTUM MECHANICS GROWS NATURALLY OUT OF
> >    THE PRINCIPLES OF INFORMATION THEORY.
> >
> > That's right, information is the fundamentual basis of the physical.
> >
> >    Last year Bub and colleagues showed that three restrictions
> >    on how information can be transmitted and copied can reproduce
> >    from scratch the basic features of quantum mechanics.  They
> >    assume that the world can be described by a type of mathematics
> >    known as a C* algebra.  Then they lay down the laws: no
> >    instantaneous transmission of information, no perfect copying
> >    of information, and no sharing of information in a particular
> >    way called "unconditional bit committment".  Those restrictions
> >    produce a mathematical formalism that looks like quantum
> >    mechanics, Bub says.  The result suggests that THE BASIC STUFF
> >    OF THE UNIVERSE IS INFORMATION, he says.
> >
> > That's right, information is the fundamentual basis of the physical. That
> > information is physical is seen in the effects of changes of information
> > in microprocessors.  Current microprocessors are so fast that transistor
> > state changes due to information manipulation lead to excessive heat being
> > generated - as much as five times the heat per square inch a stove produces.
> > The July 2nd San Jose Mercury News has a nice article on this heat problem:
> > www.mercurynews.com/mld/mercurynews/business/9063464.htm
> >
> >    "What's it mean to say that the universe is just information?",
> >    says Philip Stamp, a physicist at the University of British
> >    Columbia in Vancouver.  "To me information has to be information
> >    about something."
> >
> > Interesting question.  How about EPO Boards trying to answer the question,
> > instead of fostering the semantic fiction that "technical" can't be
> > defined.
> >
> >    Physicist Christopher Fuchs of Lucent Technologies has an answer:
> >    The wave function is an observer's information about the
> >    objective world.  A measurement instantly refines that information,
> >    which is all in the observer's head, giving the illusion that the
> >    wave function collapses.  In Fuch's view, the instantaneous
> >    collapse is no more mysterious than the clarity that suddenly
> >    emerges when you remember that you left your checkbook beside the
> >    toaster.  "Quantum measurement is the refinement of expectations,
> >    period.", Fuch says.  "The only mystery in quantum mechanics is
> >    why there is a restriction" on the completeness of knowledge.
> >
> > If transforming the physical is the requirement to be technical, and
> > descriptions of such transformations patentable, then transforming
> > information is technical, and the descriptions of such transformations,
> > hardware circuits and software listings, are patentable.  Indeed, you
> > see this equivalence logic being used in patent specifications.  For
> > example, from U.S. Patent Application 20040117276 for a system for
> > electronic shopping, paragraph 44 reads:
> >
> >      [0044]  The system described herein is not limited to the
> >      implementations described above; it may find applicability
> >      in any computing or processing environment.  The system
> >      may be implemented in hardware, software, or a combination
> >      of the two.  For example, the system may be implemented
> >      using circuitry, such as one or more of programmable logic
> >      (e.g. an ASIC), logic gates, a processor, and a memory.
> >
> > One of the gurus of modern computer science, Peter Denning, in the July
> > 2004 issue of Communications of the ACM, in an article titled "The Field
> > of Programmers Myth", writes in the first paragraph:
> >
> >      What's more, many people [falsely] believe computer science
> >      is only a technology field without much science and
> >      engineering of its own; the word "science" in our title
> >      is undeserved. ....
> >
> >      The great principles framework is a new portrayal of our field
> >      that emphasizes or scientific and engineering principles and
> >      our four core practices.  Our fundamental principles are in
> >      design and in the mechanics of computation, coordination,
> >      recollection and automation.   ...  Our four core practices
> >      are programming, engineering of systems, modeling and innovation.
> >
> > All this technical language about the world of software, and yet software
> > is deemed not to achieve a technical effect?  What scientific and
> > semantic nonsense.
> >
> > Not to say that the physicality of information, the technicality of
> > information processing, is not debatable (and subject to much physics
> > research).  But it is much much less debatable than maintaining the
> > nonsense that "techincal" can't be more precisely defined and that
> > software is not technical.  Please stop this betrayal of physics, and
> > focus on the real task at hand - getting Patent Offices to improve the
> > quality of issued patents.
> >
> > Greg Aharonian
> > Internet Patent News Service
> > ----------------------------------------------------------------------
> > This message was distributed by the EPO mailing list.
> > Information on how to leave this mailinglist can be found at:
> > http://www.european-patent-office.org/mail.htm
> > ----------------------------------------------------------------------