|
Authored by: Anonymous on Monday, June 11 2012 @ 03:29 PM EDT |
What does "from a practical perspective" even mean? It means
nothing.
Computer science is a mathematical discipline. Theory IS practice. There is no
difference.
There is NO DIFFERENCE between a mathematician carrying out a calculation with
pen and paper, and a computer carrying out that calculation with little
electrical signals. Both are physical machines performing an abstract
mathematical calculation that manipulates symbols.
A general-purpose computer DOES math. It computes. That's the ONLY thing it
ever does. Putting new software into it does not make a new machine, that's
ridiculous! Forgive me, but it's the kind of ridiculous nonsense that only a
patent lawyer could believe. It just shows how out of touch with reality they
are.
Software patents are bad for a whole host of reasons. They have almost
exclusively negative effects on everybody involved (except for patent lawyers,
and very large companies which can use them to crush their smaller competitors).
They are unconstitutional. They claim ownership of specific mathematical
ideas. They prevent other implementors from using those same *ideas* in
practice. They have huge chilling effects, putting all innovators at severe
litigation risk in exchange for virtually no societal benefit. Thousands of
them are granted every year, nearly ALL of which should be invalid even under
the current rules. (Under the current circumstances, the presumption of
validity is ridiculous, and highly prejudicial to anyone accused of
infringement).
Please help us build a saner world. Ban software patents. And while you're at
it, please feed all patent lawyers to the lions. Help make our world a better
place.
--programmer of 20 years.[ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Monday, June 11 2012 @ 03:34 PM EDT |
Setting virtual machines aside, no doubt the hardware is the machine and it was
already capable. The fact the software merely turn a feature on or off does not
change the machine. And whatever process there is the software is carrying out
is immediately obvious because it follows from the machine spec (for example, if
the antenna can't be turn off in hardware, no software will be able to do it).[ Reply to This | Parent | # ]
|
|
Authored by: PolR on Monday, June 11 2012 @ 04:36 PM EDT |
I don't think your credibility is on the line. You are the legal expert and this
discussion is much needed. I am happy to see you here.
I think the following
might be the heart of the dispute.
I am not understanding the
fixation on a new machines for two reasons: a. Processes are patentable, so new
software in a machine means we can use it for a process we could not before, and
that has nothing to do with this "new machine" distinction; b. Patent law has
never required that every patent cover a new machine. Instead, it has required
that the capabilities of machines be new. That's why we have a gazillion patents
on plows, flour hoppers, and other such stuff. Plows may have all the same
pieces, but if they are made if different materials, made in different ways, or
have the parts configured slightly differently, they can be patentable. Software
is being treated the same way - in practice.
Technical people
think in terms of causality. If attorneys are going to tell us that new
functions are evidence that a new machine is being made, we want to see what is
done to the machine in structural terms to implement the new function. Otherwise
we are in a regime where it is assumed that function can never be implemented
without physically some new machine and this is technically incorrect.
Here
is what we see in this practical approach to what is a new machine. We
are in a regime where:
- Every physical changes potentially make a new
machine.
- Physically identical changes sometimes make a new machine and
sometimes they don't depending on legal context.
- The causal relationships
as understood in principles of computer science and electronic engineering are
not considered relevant in this determination and are ignored.
- The actual
machine structure is not even looked at. Function alone is evidence that the new
machine is made.
- The possibility of implementing new function without
actually making a corresponding physical structure which causes this function is
ruled out by legal fiat.
- The existing capabilities of the machine which have
been built-in by the designer are ignored when determining whether the machine
has a new capability. Only prior art counts in this analysis.
I don't see
how this is different from a license for attorney to call a machine whatever
they want, leaving prior art as the sole test to determine of whether this
machine is new. We may as well read the word machine out of the statute and say
new functions are what is patentable.
Software is very different from plows.
These mechanical devices are not programmable. Sure there are commonalities from
a plow to another but the differences in design are permanent features of the
plow and they have a causal relationship with the improve function of the plow.
In software the decision that a new machine is made disregards causal
relationships and ignore the applicable scientific principles.
This view
also disregards that it is possible to make sure the so-called structure which
has a causal relationship with the function to be transient, that the structure
in memory doesn't stay in place for the entire execution of the software. How
could this circumstances be deemed irrelevant in practice? It suggests that the
invention is not a machine at all, old or new.
This last paragraph is a
major sticking point. We can't have a rational discussion on whether the patent
reads on an algorithm with someone who thinks the invention is a machine even
when no machine is actually made. This notion obscures what the invention is. In
fact, it prevents any analysis of what the invention is because the invention is
deemed to be something else by legal fiat.
This also obscures the
discussion on what the process claimed in software is. The nature of this
process is not the activity of transistors. But people who think software is a
machine have this conception in their mind. Again this makes impossible rational
discussion and analysis of what the invention actually is.
I observe that
your article omits that it is mathematical algorithms that are not
patentable. The reference to mathematics is important. But an analysis of
whether the invention is the algorithm cannot be done correctly by someone who
thinks the invention is a machine, or the activity of transistor. Having the
facts of technology right is the first step. If you work from facts that are
wrong, analysis is indeed impossible. But if you work from facts which are
correct the analysis will be much easier.
I think that your approach is
equivalent to saying function is patentable regardless of what is actually
implemented and therefore analysis of implementation is not required.
[ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Monday, June 11 2012 @ 04:46 PM EDT |
If I replace my phone software with software that disables the cell
antenna, it is practically and functionally a different machine. And if I add
navigation software to my phone, it also becomes a navigation device and it
wasn't before. Yes, all the parts are there, but the use of the machine is far
different.
By this reasoning turning on the headlights in a car
turns it into a new machine. Just like flipping switches to turn on the
headlights and turn off four wheel drive, all software is just the flipping of
switches (i.e. the bits in memory). In older mainframes, a human would
physically flip these switches. Now days, the switches are electronic and are
set electronically.
Would you propose that every setting of the switches on
an Enigma machine turns it into a new machine worthy of individual patent? Does
flipping the dip-switches on a circuit board turn it into a new patentable
device? The only differences between dip-switches and memory is how you set the
switches (physically vs electronically) and scale (dozens vs billions).
Part
of the difference of view may be that you (like many users) seem to view a
computer in terms of tasks (e.g. web-browsing vs word processing). In that
view, a computer looks like a different machine when being used for different
tasks. However, the beauty of the computer is that it is like a chameleon and
can appear like many different machines merely by flipping a few switches
(activating pigmentation cells?). A new program does not turn a computer into a
different machine, it merely makes it look like a different machine.
I am
specifically not addressing the patentability of processes with this comment or
whether software patents are a net positive for society. My point is that the
courts' idea that reprogramming a computer makes it a new machine does not
reflect the engineering reality, and I doubt (hope?) courts would make similar
determinations on other devices (e.g. car headlights, Enigma machines,
dip-switches, etc.). (A case on dip-switch setting could be a particularly
interesting and relevant precedent.)
[ Reply to This | Parent | # ]
|
- Well, since my credibility is on the line... - Authored by: mrisch on Monday, June 11 2012 @ 05:26 PM EDT
- Well, since my credibility is on the line... - Authored by: Anonymous on Monday, June 11 2012 @ 05:57 PM EDT
- Well, since my credibility is on the line... - Authored by: PolR on Monday, June 11 2012 @ 06:51 PM EDT
- Well, since my credibility is on the line... - Authored by: Anonymous on Tuesday, June 12 2012 @ 09:42 AM EDT
- Well, since my credibility is on the line... - Authored by: mipmip on Tuesday, June 12 2012 @ 12:49 PM EDT
- Well, since my credibility is on the line... - Authored by: rcsteiner on Tuesday, June 12 2012 @ 04:36 PM EDT
|
Authored by: nsomos on Monday, June 11 2012 @ 05:56 PM EDT |
By that logic, depending on what route you drive,
and what cargo you might carry, the family minivan
becomes a different machine. What mighty magic is this?
If I carry computers for delivery in there, can I get
a patent on that. Maybe if I carry something in it,
that no one has ever carried before, I could patent
that use of that vehicle.[ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Wednesday, June 13 2012 @ 06:46 AM EDT |
Thanks for continuing to participate, prof Risch, but you have ducked the main
point: that current US law on software patents relies on fictional computer
science.
I agree with PolR that you are the legal expert, and am happy to defer to you on
legal matters. But I maintain that unless you show the same courtesy to the
computer science experts on Groklaw, your credibility remains that of a patent
lawyer.
Please respond to PolR's comments about how computing machines really work and
why the current US law is factually in error.[ Reply to This | Parent | # ]
|
|
|
|
|