|
|
|
| Authored by: PJ on Monday, June 11 2012 @ 06:25 PM EDT |
It's closer to the 500 individuals than 5 100 times.
I'm sure we all thank you for listening to what
everyone had to say.
And personally what I would hope you would think about
is that these are the voices of those skilled in the
art. We recognize that corporate entities value
patents, and they get their employees to file for
patents, but that doesn't mean that the employees
like it. At Sun, for example, they used to have a
contest to see who could get the stupidest patent,
and they got them too.
So what I would hope you would do is this: think about
the benefit to harm ratio in this picture. There *has*
to be a reason that programmers hate software patents
so, so much. And then look at who loves them. Then
look at how the system is currently playing out.
In other words, think not from the perspective of
the happy patent owner making a bundle only. Think
about the public interest. There is no balance
currently. A happy few make money, and the rest of
the world of software is suffering, and over time,
that has a huge impact on innovation. Judge Posner
recently recognized the importance of considering
patents from the standpoint of the public interest,
and he tossed out an Apple-Motorola case, because
he said neither side could prove damages, but even
it they could, it was contrary to the public interest
to issue injunctions.
When judges start to realize there is real damage
to the public, it's time for lawyers to start to
think about that too.[ Reply to This | Parent | # ]
|
| |
| Authored by: Anonymous on Monday, June 11 2012 @ 06:54 PM EDT |
I think maybe there is still a lack of understanding regarding some of the
points raised (by both sides). For one example, I think you haven't quite
absorbed the "software is math" arguments. Its very fundamental.
Software is as abstract as something can possibly get. Writing it and executing
it are both purely mental activities. Software is a symbolic representation of
a mathematical algorithm. Executing it is carrying out a mathematical algorithm
(a mental process).
People who say "software isn't math", either don't really understand
what software is, or don't really understand what math is, or both.
Anyhoo!
I'm very glad you stirred up this debate and took the time to read and respond
to so many of the comments. So thank you for that!
It usually seems like the legal community completely ignores us, and just makes
up their own "facts" about how computers work which are totally
irreconcilable with reality. :P[ Reply to This | Parent | # ]
|
| |
| Authored by: Anonymous on Monday, June 11 2012 @ 07:09 PM EDT |
| a. Most software patents are not like the Swype patent. They're more like this patent which Oracle just
tried to use against Google in a court case. It purports to be about some
kind of "machine" but its really just a patent on an algorithm (or on using an
algorithm to solve a certain kind of problem), and its so hopelessly broad that
Google just spent several million dollars in a multi-week jury trial just to
show that their clearly very different implementation, did not
infringe.
b. ["Even if software is all really math in theory, I'm
talking
about applied math."] Software is math, both in theory and in practice.
Whether you apply it to something has no effect on this. Maybe the
process of applying it to something should be patentable subject matter,
but software itself should not.
b. ["We've had applied math as
patentable in
the past - some of you have said the rubber
process in Diehr. There's also the
MacKay Radio case - a
mathematical equation applied to an antenna, etc."] Okay,
the thing to understand is that the process of curing the rubber, or the
physical antenna, may be patentable subject matter. The piece of math by itself
is not. The piece of math has some semantics associated with it that
have something to do with curing rubber (temperatures and times, etc.) Those
semantics might be part of the process; they're certainly NOT part of the math.
The math is just an abstract manipulation of symbols. Assigning meaning to the
symbols is done by the humans, outside of the math part, so they can use the
math part to do something practical. Crucially -- the *very same math* might be
useful to do something very different, if you just assign a different meaning to
the symbols in the math.
c. ["I have read hundreds and hundreds of computer
science
journal articles dating from the 1980's until now."] Cool, so have I.
All of those articles are written by people who are trying to solve a certain
problem. They describe their problem, and the solution they came up with.
Although it isn't usually described this way, their solutions are a combination
of two things: (1) a bunch of abstract mathematics (algorithms), and (2) a
semantics which assign some sort of meaning to the symbols used by those
mathematics. Part (2) is what allows us to do useful real-world things using
software as a tool (just like anything in mathematics).
The argument is not
"software is simply doing mathematics", although in a literal sense that
is also true.
The argument is that "software IS mathematics". Full
stop. It should simply not be patentable. [ Reply to This | Parent | # ]
|
| |
| Authored by: PolR on Monday, June 11 2012 @ 07:51 PM EDT |
| I think this summary reflects what has been debated. I am happy to see you here.
Even when we disagree I think it is valuable to discuss like this.
I wish to
clarify a bit about software is math. Let's consider:
Even if
software is all really math in theory, I'm
talking about applied math. We've had
applied math as
patentable in the past - some of you have said the
rubber
process in Diehr. There's also the MacKay Radio case - a
mathematical
equation applied to an antenna, etc.
The compare
with:
There has been a lot of talk about software creating a
"new
machine."
. . .
As I note in that post, structure is irrelevant for the
most
part, capabilities are what matter. As I also note, I think
the focus is
misplaced since a process is patentable whether
you create a "new" machine or
not.
I think these two points are connected. Your examples of
applied math are physical things. The Diehr patent is on an industrial process
where the rubber is actually cured. The MacKay Radio case is on an antenna which
is actually built. What would happen to corresponding patents which claim just
the calculations and say this is a patentable process on the grounds that the
formulas have a practical use? In these patents the rubber is not cured. We just
compute the curing time. In the antenna patent the antenna is not built. We just
compute the numbers. Would this be patent on math? I think this would. These are
claims on the formulas with a field of use limitation.
When we say software
is math we are often told that the mathematical algorithm describe a physical
invention, therefore software is not math unless we say the rubber curing and
antenna patents are patents on math. This is the notion that software makes a
new machine in a different form. The software patent is treated like an applied
math patent like Diehr and MacKay Radio because there is something physical
being invented.
Having been told that, I examined the structure of the
computer to find out what is it this applied math invention which is being
claimed. If the math of software describe something physical, then something
physical must be found that match the math. And I have found that this is
nowhere in the computer structure. I have reported this to you. I am now being
told that structure is irrelevant. It is capabilities that matters.
I must
ask again. Where is the application of this applied math? What is being
described by the application of math? Is it capabilities disjointed from machine
structure? How is that different from a patent on the rubber curing formula
without curing the rubber? How is that different from a patent on the
computation for the antenna without making the antenna?
This is why I focus
on the "make a new machine" thing. It tells us what the applied math is, or more
exactly it tells us that it is nothing physical. Unless there is something in
the claim like physically curing the rubber there is only a mathematical
computation with a field of use limitation. What you call a machine capability
is often a field of use limitation applied to an underlying algorithm which is
not expressly mentioned. But still the algorithm is there because otherwise the
computer wouldn't compute.
When we say software is math, we say software is
not an application of math. Software is the math itself. We say that unless we
have something like the MacKay Radio patent in front of us, the application is
not the subject matter of the claim. The algorithm is the subject matter. And we
also say it is not hard to find which mathematical algorithm is involved, even
when it is not clearly mentioned in the words of the patent. We just have to
take a good look at computation theory and look at what are the operating
principles of the bare computer. The algorithm is there like the electricity is
there whether or not the claim mentions them.
[ Reply to This | Parent | # ]
|
- Thanks a lot, and a clarification - Authored by: mrisch on Monday, June 11 2012 @ 08:16 PM EDT
- Thanks a lot, and a clarification - Authored by: Anonymous on Monday, June 11 2012 @ 08:27 PM EDT
- Thanks a lot, and a clarification - Authored by: Anonymous on Monday, June 11 2012 @ 08:35 PM EDT
- Thanks a lot, and a clarification - Authored by: PolR on Monday, June 11 2012 @ 09:15 PM EDT
- And even on those processors that do - Authored by: jesse on Tuesday, June 12 2012 @ 08:18 AM EDT
- Thanks a lot, and a clarification - Authored by: mrisch on Tuesday, June 12 2012 @ 08:28 AM EDT
- disgusting. - Authored by: Anonymous on Tuesday, June 12 2012 @ 08:45 AM EDT
- Can do vs. Will do - Authored by: Anonymous on Tuesday, June 12 2012 @ 08:58 AM EDT
- Ha. You make the complaint of every mathematician makes - Authored by: jesse on Tuesday, June 12 2012 @ 09:04 AM EDT
- Your last paragraph is funny. - Authored by: jesse on Tuesday, June 12 2012 @ 09:18 AM EDT
- General-purpose computer - Authored by: Anonymous on Tuesday, June 12 2012 @ 09:24 AM EDT
- Thanks a lot, and a clarification - Authored by: PolR on Tuesday, June 12 2012 @ 09:24 AM EDT
- "Capable"? - Authored by: Anonymous on Tuesday, June 12 2012 @ 09:57 AM EDT
- Patent on problem specification? - Authored by: Anonymous on Tuesday, June 12 2012 @ 10:03 AM EDT
- Thanks a lot, and a clarification - Authored by: PJ on Tuesday, June 12 2012 @ 11:10 AM EDT
- Atoms - Authored by: PolR on Tuesday, June 12 2012 @ 11:12 AM EDT
- People stuck in the wrong environment - Authored by: Anonymous on Tuesday, June 12 2012 @ 11:54 AM EDT
- Mrisch, from a patent lawyer's point of view, your program was SUGGESTED - Authored by: Anonymous on Tuesday, June 12 2012 @ 10:19 AM EDT
- Software Patents ne Process Patents - Authored by: RMAC9.5 on Tuesday, June 12 2012 @ 04:34 AM EDT
- Thanks a lot, and a clarification - Authored by: Anonymous on Tuesday, June 12 2012 @ 10:16 AM EDT
| |
| Authored by: jonathon on Monday, June 11 2012 @ 10:07 PM EDT |
> I wonder whether many readers think some patents are broader than they
are.
That is the precise problem with patents. Not just software patents, but others
as well.
Even though a patent has at clause 4 "Eat a Babe Ruth Bar", and when
you come to clause 4, you "Drink a Jolt", with software patents, you
have infringed, precisely because of the patent nonsense that courts have
allowed patent trolls to get away with.
[ Reply to This | Parent | # ]
|
| |
| Authored by: Anonymous on Tuesday, June 12 2012 @ 01:49 AM EDT |
If the argument is over what the law SHOULD be, then one must make an argument
in terms of justice and net benefit to society. And in those terms software
patents are absolutely indefensible.
If the argument is over what the law actually IS, then why are you starting with
the presumption that software is patentable? Software is copyrightable. For what
reason should it also be patentable? One can argue strongly that software is
mathematics and so falls within one of the existing exceptions to patentable
matter. There is no legal reason why the law should not be construed so that the
prohibition against patenting mathematics is interpreted broadly enough to
capture software in its penumbra.
The world is full of patent lawyers being paid to argue that everything should
be patentable. There is a desperate need for people to put the other side of the
argument on behalf of the public at large. I don't understand why you, who are
not specifically being paid by a patent holder, should choose to start from the
extreme position that everything is patentable and the exclusion for mathematics
should be so narrowly defined that things like
(x^a)^b = x (mod pq) where ab = 1 (mod phi(pq) )
are allowed to be patented. In case you didn't recognise it, this is the guts of
the RSA encryption algorithm. It quacks like mathematics to me.[ Reply to This | Parent | # ]
|
| |
| Authored by: mjscud on Tuesday, June 12 2012 @ 01:54 AM EDT |
You have argued against subject matter distinctions, and I think the Supreme
Court was right to reject that argument.
You agree that Software is
Math, but you don't think that means that software should therefore not be
patentable. Thus I think I need to argue why Math should not be patentable, as a
subject matter. As I argued at D
rawing a clear and well defined line - Turing Equivalence comment, I think
Software as Math can be a bright line for distinguishing patentability. I think
trying to do this on the basis of usability is an unnatural extension of that
(good!) criteria to fill a need where a better criteria already
exists.
Math by nature and practice is an abstracting and progressive
activity. Granting monopolies is inimical to its progress. It is a natural
mathematical step to consider all numbers, all algorithms, all data structures,
and to use them all at once as well as individually. Accreditation and
appreciation, as well as the experience of beauty, not monopolization, is the
accepted and time tested way in which it rewards is practitioners. Diametrically
opposed to the monopolies granted by patents, the best reward for a new idea in
math is for it to be appropriated and utilized by as many and as freely as
possible.
---
Even a fool, when he keeps silent, is considered
wise. Proverbs 17:28 [ Reply to This | Parent | # ]
|
| |
| Authored by: Anonymous on Tuesday, June 12 2012 @ 10:08 AM EDT |
including the Swype patent, which patents the combination of unpatentable
software, someone else's patented touchscreen, and a general-purpose computer.
It doesn't create a new process to solve any real-world problem (like
rubber-curing) either; it solves a *business methods* problem.
It's unpatentable in many, many ways.[ Reply to This | Parent | # ]
|
| |
| Authored by: Anonymous on Tuesday, June 12 2012 @ 10:34 AM EDT |
Returning your own comment with which you started your article, if you think
that software patents are a good thing, it’s unlikely that I will convince you
otherwise.[ Reply to This | Parent | # ]
|
| |
| Authored by: mrisch on Tuesday, June 12 2012 @ 01:59 PM EDT |
OK, all, I think we're about at the point of diminishing
returns in the discussion. We aren't hearing much new from
each other in the comments at this point. A few more points
to consider:
1. I think the analogy of the pianola (or even guitar) tune
to the software is an important one. I think that there is
an important practical distinction between a computer being
instructed to do some things and a musical instrument being
instructed to do some things, even if in theory it all
really boils down to a mathematical model. I realize that
most of you disagree.
2. I'm still not sure we are all talking about the same
things when we talk about software. I think more precision
between "pure" software (e.g. calculations) and software as
part of some physical activity would be helpful, including
how physical the activity must be (printing on a display v.
moving a robotic arm).
3. I think it would be interesting to have some sort of
experiment on obviousness. I don't know how to do it -
coming up with the list of previously unsolved problems and
then seeing whether people can solve it. I hear you all
saying that there is no problem that can't be easily solved,
and I would like to put that to the test.
4. Relatedly, there is a lot more to be done on the theory
of obviousness - work I haven't done. Historically, the big
idea that leads to a solution to a problem has been judged
separately from the implementation, which might be obvious
once you've had the big idea. I hear many of you saying that
obviousness should work that way. I'm on the fence, quite
frankly. I think there is value in the big idea, but I also
worry that too many small ideas are being treated as big
ideas. The debate over the Swype patent (which I think was
probably obvious) is a great example of this phenomenon.
5. I think that many readers here put much more emphasis on
the "new machine" concept than patent lawyers, legal
academics, courts, and the PTO do. For my part, I'm happy to
focus on the process and just never have new apparatus
patents. The good (and bad) of the system won't change a lot
with that distinction. Se spent a lot of space here arguing
about whether a new machine is made (in practice if not in
structure), but that's a sideshow for the things I'm
thinking about given that patentees can easily sidestep the
issue with magic words of claiming.
That said, I'm about to read a draft of a new article about
functional claiming - so this is definitely a live issue.
Thanks again for all your comments.[ Reply to This | Parent | # ]
|
| |
| Authored by: Anonymous on Tuesday, June 12 2012 @ 08:05 PM EDT |
it's patently obvious that you DON'T get it.
(see what I did there?)
you just don't get it. and what's more you don't want to get it. your whole
fragile world clings to these false beliefs.
whether it's software, or math, or applied math or existential math or euclidean
physics or algebra or even calculus.
it's all still math, and it cannot and should not be patentable.
software is math, not matter what "math" you want to call it. it's
still math and still not patentable.
i don't care if its the most amazing "applied" math you've ever seen
in your life, it's STILL *JUST* math.
and yknow what, you still don't get it, so forget I said anything.... [ Reply to This | Parent | # ]
|
| |
| Authored by: Anonymous on Wednesday, June 13 2012 @ 03:37 AM EDT |
Michael, I fear that unless you consciously and intentionally revise your
premise that it's money that matters in this area, you will never be able to
reach a full understanding of why you are so out of step with software
practioners skilled in the art.
Although you do understand software, your outlook is not that of a skilled
software practitioner, because you do not share their values. As a result, any
conclusions you reach will always fail the test of domain insight, because your
insight is swayed by a totally different domain, that of business.
There is no possible escape from this quandry. You currently believe that you
are right and that we simply haven't analyzed the situation well enough to agree
with you.
But you are wrong in that. We simply belong to different domains, and what you
hold as an axiom is to us false.[ Reply to This | Parent | # ]
|
| |
| Authored by: Anonymous on Wednesday, June 13 2012 @ 09:00 AM EDT |
I hope you found this discussion at least partially fruitful and civil.
I couldn't read all of the comments, so I don't have any idea if I'm restating
something that's been said.
Seems to me, a conceptual problem about patenting software, is that software on
it's own, doesn't do anything. Like you said hardware matters. It has to part
of something else. Software could be incidental to an invention, or a process,
but then the whole system is the thing that should be patented, if it qualifies,
and not the software in and of itself. Software patents don't deserve it's own
special category. Patents should be narrow and precise, rather than broad and
sweeping.
Maybe you are right, in that it might be easier to fix a large part of the
problems with some minor tweaks or even just a proper implementation and
enforcement of the existing system, or adding incentives and penalties, rather
than attempting to pass a complete rejection of software patents into law.
Rather than commenting on your solutions, I'd like to re-emphasize some of the
problems:
1) Why are so many bad patents passed? Why do so many patents fail validation
under re-examination? Maybe the solution is to force the main elements of the
re-examination process before the patent is ever granted, giving a chance for
other interested parties to submit prior art, or otherwise challenge patent
validity. We need the PTO office to apply more field expertise in the
evaluations than they are doing before bad patents can do any damage. Maybe we
need to find a more efficient way to penalize submitters of frivolous patents
without discouraging real inventors.
2) Litigation is expensive. I guess I didn't understand your suggestion about
being more efficient to solve the problem at litigation. Litigation is costly,
starting at millions of dollars, and out of scope for individuals and small
businesses. Not to mention the harm of wrongful injunctions. Patents, and
software development, then become a game only the rich can play.
3) How do you do due diligence? Supposed you honestly want to follow the rules,
and develop an application that infringes no patent? Is it possible to scan the
universe of applicable patents? What about patents pending? The patent system
becomes a game of Russian roulette for any individual, startup, or small
business.
4) The real measure of success is the progression of the useful arts. If
patents only serve to delay a technology, it's having an opposite effect.[ Reply to This | Parent | # ]
|
| |
| Authored by: mrisch on Wednesday, June 13 2012 @ 02:51 PM EDT |
All - I'm out of time on this post, unfortunately, as I've
got other deadlines
and kids who want to play in the
summertime (they joys of being on a school
schedule). Thus,
I likely won't be able to respond to responses to this post
(and I'm sure there will be some). :) Thank you all for your
ideas again.
I
have read many of the follow-up sum-up posts, and leave
you with these
thoughts:
1. Please remember that I started my post with an agreement
that
software patents are a problem. I understand that
patents - and especially
overbroad and otherwise bad patents
- can limit what you do and expose you to
risk.
2. Some of the debate in the posts is a simple disconnect
about how
patents are drafted and how patent claims work,
etc. I made a mistake by not
putting more of that background
into my post, as that might have helped
structure the
discussion. Sorry about that.
3. The much debated issue of a
"new machine" is a good one
and deserves more explanation. I'm persuaded that
changing
the series of switches inside a general purpose computer is
not
structurally a new machine. Of course, that was easy to
see. The disconnect we
have had (as noted in some threads
above), is that patent claiming has allowed
what we call
"functional" claiming, such as "means for doing x". What
this
means is that the law treats the programmed computer as
having functionally new
structure even if there is no new
physical structure. Judge Rich wrote Alappat.
He wrote the
patent law. He sat on the court of appeals for 50+ years. I
assure you that he knew that changing the software did not
change the
structure of the device. It's just that the law
did not really require that the
physical structure be
different - just the specifically implemented
functionality.
I think it is worth looking at whether that view is wrong,
and
I know a couple articles are being worked on that
discuss this. Here's a good
one to start:
http://www.bannerwitcoff.com/_docs/library/articles/FUNCTION
AL%20CLAIMING%20AN
D%20FUNCTIONAL%20DISCLOSURE%20Wright%202.1
1.pdf But note that he looks at
things a lot like I do -
looking at other patentability criteria like written
description, definiteness, etc.
4. At the same time, my suggestion to you as a
community is
that you focus a little more on process claims, which are
much
harder to pick on than "new machine" claims. The
statute makes clear that a new
use of a machine (or anything
else) is patentable. This is very broad. If I
develop a drug
to treat blood pressure, and later discover that it cures
erectile dysfunction, I can get a patent on the use of that
drug for ED even
though the structure of the medicine has
not changed (as we see with
[redacted]). These claims will
likely survive Prometheus - they are
specifically mentioned
as NOT being covered by the opinion.
So, if I discover
a better way to use a telephone to
transmit electronic communications or a new
way to use a
cell phone for navigation, then I've created a process that
is
patentable even if I didn't change the physical structure
of the device. The
fact that it is a general purpose
computer is irrelevant - if no one used the
computer for
that purpose before, then it's a new use of the machine.
The
problem (for me, at least) is that it is not enough from
a subject matter point
of view to say that there is software
involved. However, if it is ONLY software
for pushing
numbers/symbols around, (see, e.g. Gottschalk v. Benson)
then
there is an argument. You would say it is abstract. I
would say it has no
practical utility because it doesn't do
anything. But the end result is the
same.
5. Finally, we all need to think hard about just where the
abstractness
ends. Yes, if it is for "software" only (like
Sun's patent), then it's an easy
case. Some of you have
argued that most software patents are like that. I think
the
opposite, in part because I've had discussions with folks
who have read
tens of thousands of software patents (and
I've read quite a few myself).
Patent lawyers are smarter
than that - they know that a "pure" software patent
has a
bigger chance of getting bounced. Even the RSA patent, which
is as close
to pure math as you can get, claims physical
devices used to encode and decode.
Yes, there's software and
math that makes it go, but software and math makes
everything go.
To solve this problem, courts have said we should ignore
"insignificant post solution activity" ignore the physical
if the solution is
really just abstract. This works for some
- it may work for many readers here.
I tend to think it is
an untenable line.
First, it blurs the lines of
novelty, obviousness, and
subject matter, but with none of the rigor of novelty
or
obviousness. Courts can do what they want. That's great if
they invalidate
the patent, but not so great if they don't.
Second, as I noted in my original
post, hardware matters.
Maybe not in theory, but in practice. Some solutions
may be
harder to achieve in the hardware of the day, and
determining which
ones are "insignificant" in a particular
hardware is difficult.
Third, I think
the hardware line is not clear. Most
commenters don't seem to have a problem
with Diehr. That's
fine - the computer drives the rubber mold and timer. Now
what about a robotic arm. Still physical. What about machine
vision? That
seem's physical, but is more abstract. What
about touchscreen apps? That's
handling input from physical
devices. What about monitors? The bits sent to the
monitor
are no different in theory than those sent to the robotic
arm or the
rubber mold machine, so why should they be
considered any less physical?
I
realize in practice that stuff on the screen affects more
of you than stuff in
rubber molds and robot arms, but I'm
looking for consistent, principled ways to
make the
separation, because somebody will always be affected.
So, those are
some thoughts. I hope they are helpful. I
learned a lot here, and I appreciate
your input.[ Reply to This | Parent | # ]
|
|
|
|
|
