|
|
|
| Authored by: pem on Tuesday, January 01 2013 @ 05:44 PM EST |
1) If the patent is disallowed solely because the
computer is involved, then you have a result where purpose-built hardware could
be patented, yet repurposing a generic computer to achieve the same task could
not be. This would discourage reuse and hardly be in the spirit of advancing the
useful arts.
This is conflating the algorithm with the device
with computes it. It is like conflating a novel with a stack of paper covered
with ink.
I assume you mean that the disallowance because of the
use of the computer would be this conflation. Otherwise, I am not sure what you
are getting at. If this is the case, I don't see where our disagreement is; if
not, you'll have to explain better.
I am perfectly fine with
patenting a new a nonobvious circuit for carry out the calculation as long a the
circuit is patented.
But that non-obvious circuit will, these
days, most likely be expressed in Verilog or VHDL. An expression which is
easily translated into a programming language, even automatically. (In fact,
VHDL is really just a variant of ADA.) So should the non-obvious resultant
program also have patent protection?
The algorithm is not the
circuit and people should be free to do the calculation through different
means.
But the algorithm is also not the program. So what is
really being patented? Where is the bright line?
Also programming
a computer doesn't make a new machine. This is a legal fiction.
I
suppose that depends on your definition of machine. Certainly programmers
engage in the same fiction, if that is what it is, when they talk of virtual
machines.
Occam's Razor would seem to indicate that
(3) is the right answer, so no, you don't actually have an invention "outside of
the calculations" especially since equivalent calculations would be required
even without a computer.
This conclusion doesn't follow from
point 3. The calculations are part of the inventions. They are not the whole
invention.
We are in violent agreement here, except, perhaps that
I would go farther and state that, in some cases there is no invention (nothing
new or patentable) _without_ new calculations.
The calculations
taken in isolation are [not] the invention.
Sure, but the rest of
the machine that cures the rubber is not the invention either. The calculations
are a necessary component. Even back then (especially back then, before the
patent office went wild) you would probably have a hard time patenting a
"process" with a bunch of vague diagrams about how a computer controls valves
and heaters and pumps to produce a desired product. You have to be somewhat
specific about how your new process or machine works.
In the
rubber curing example, you still have to cure the rubber. In the space shuttle
example you still have to fly it.
And, again, for many practical
purposes, that makes it a moot point whether you can actually patent the
software or not. If you can patent the machine containing the software or the
process using the software, then from a commercial perspective, you might as
well have patented the software.
Addition of two
known numbers is normally considered applied math, not abstract math, regardless
of the size of the numbers
This is wrong.
We'll
simply have to disagree about that.
Besides the patents in Benson
and Flook were patents on mathematical algorithms where the calculation is
actually carried out. So you can have abstract math in the Surpreme Court sense
when actual calculations are done.
Which the Supreme Court
backtracked on in Diamond (foreshadowed by the dissent in
Flook).
As a general note, my discussion on the real-life
limitations is to clarify what is a mathematical algorithm in the sense
mathematicians give to the word. This knowledge helps knowing when a claim is
written to a mathematical algorithm as opposed to an application of the
algorithms.
On the one hand, if that approach is adopted, it will
make claims construction more exact. On the other hand, it will probably lead
to more impenetrable patents, as the lawyers discuss every possible application
of any algorithm.
The difference between an abstract idea and the
application of an idea is precisely the topic of the debate. The insistence that
actually carrying out the application is the application of mathematics is one
of the errors that leads to so many silly software patents.
I
don't think so. If you insist on driving this distinction home, then one likely
outcome is that just like "on a computer" became magical faerie patent pixie
dust earlier, we could easily wind up with similar magical faerie patent pixie
dust that reads something like "and for any application that could make use of a
similar calculation."
This is because it takes more than some
physical effects to have an application of mathematics.
Does it?
Either math underlies everything or it doesn't...
Actually
carrying out the calculation involves a physical effect but this is not
sufficient to have an application of mathematics. We need to take into
consideration that mathematics is a language. We need to distinguish a
manipulation of the language from an application of the
language.
Good luck with that. Many of the most productive
applications of computer programs involve manipulation of other programs.
(Compilers, for example.)
To illustrate this point, we may write a
parody of the "programming a new computer makes a new machine" using a printing
press. A printing press configured to print a book is a different printing press
than one without such a configuration. One can lay out ink on paper to make a
book while the other cannot. A book with ink laid out in different manner are
physically different books. And printing presses configured to print different
books perform different functions, they make different articles of
manufacture.
Yes, but the similarities are striking. The law
deals with this with things like the Doctrine of Equivalents, and you would not
get a patent on a machine that prints a different book, or a process for
printing said different book, absent an actual invention, such as something that
prints braille instead of ink.
The printing process is useful, it
makes a physical article of manufacture. A novel is the list of letters that
specifies how the ink must be laid out. A patentee is free to describe his
invention that way he chooses. He may use the text of a novel to do
so.
But the point of patents is to protect new, non-obvious
inventions. To the extent that a book doesn't use any new process, there is
nothing protectible about rearranging the type for a different
book.
I guess you get the point. In this example the invention is
the novel. It is not a book, a printing press or a printing process. But we may
play with words and use the fact we need a physical support to record the novel
to try to claim it.
The law protects what goes into a novel with
copyright. It is explicitly excluded from patent protection. There is no
"invention" as described by the law in a new novel.
If you want a better
analogy, try an injection molding machine. Like a printing press, this can
churn out a zillion identical copies, then be easily reconfigured to churn out a
zillion identical copies of something else. Each of the designs being printed
out could be patented; in fact, could be protected by multiple
patents.
The problem is the Federal Circuit has not recognized the
role of the semantics of symbols (bits) in software. One need to recognizes the
role of semantics to be able to draw a meaningful line.
I do not
find the analogy of printing press and book vs. general purpose computer and
program at all useful, and I am sure there are many others with similar
viewpoints. If you want to make this analogy work, you first have to
distinguish the printing press from the injection molding machine, and explain
why the computer is more like the printing press than it is like the injection
molding machine. Personally, I find a general purpose computer to be more like
an erector set. You can build a lot of different things with it, but it remains
an integral part of what you build.
Back to the shuttle example,
the engineers have invented something other than a programmed computer. There is
a shuttle flying. But if the act of computing itself is considered a patentable
physical effect than how is that different from patenting a novel in the manner
I have described?
Injection molding is more like printing than
computing is, and the product produced by injection molding is patentable. So
work on that distinction first.
But in any case, another practical difference
in the process required to print one novel vs. another is negligible, and
requires no effort that any ordinary citizen would consider to be "invention"
but the practical difference between writing a word processor and writing a
space shuttle control program is rightly perceived to be huge.
I
know the law makes a difference, but my question is why? This seems an arbitrary
decision to me.
Sure, the law is full of arbitrary decisions.
Novels are copyrightable, not patentable, by statute and even, depending on who
you ask, the constitution.
[ Reply to This | Parent | # ]
|
|
|
|
|
