|
|
|
| Authored by: Anonymous on Sunday, June 10 2012 @ 07:41 PM EDT |
Copyright should be the only protector of software.
It is a PURE FACT that Patents, meaning methods and
concepts, when you look at software, you see what it is.
Math, period.
From the original machine to add, it is all the same, in a
binary world. And really, all software is in part derived
from the allowed math that is processed by the chips. The
real invention, is the chips and the design of those that
allow for the software to work. The pro-software-patent
crowd divorces themselves from the fact that software is
attached by the logic of an umbilical cord of allowed,
totally dependent on the machine and the design of the
electronics of the machine (chip).
In reality, Intel Corp, could lock up all uses of it's chip
by filing for one patent on the "derived-from" uses of all
the allowed math (or software) that can be processed by
their chips. So, no one else could ever file for a patent,
as the "great-grand-daddy" of all patents would be ruled by
the chip designer, and it's master patent. It's math, so
if a master software patent were allowed by the "inventor"
then, game over. Lowest common denominator, is what it is,
you can't dream it into being something else.
The chip maker being in control of all software written to
work on that chip, is in itself, proof that software is
math. If you go down all the logic rabbit holes (and I will
defer to all the Groklaw posts and articles), you arrive at
a lowest common denominator... that software is Math,
nothing else.
PURE math, so no patents please.
Copyright yes, but only on Original works, not simply
something that someone made better by taking it from analog
use to digital and thus then written in code (same story,
but to work on a machine), a story that has been in the
"analog" world, used for a long time before. Meaning, just
making the process run faster on a machine, that existed
before without a machine, is not something you can protect
in any way at all. Prior art, obvious.
Software is math, and stories are stories (not a lot new
there, ever compare some Beatles tunes with Mozart).
Supreme court in Bilski was WRONG. And progress is
suffering because of that ruling. Not logical, absent all
use of the Socratic Method to arrive at their result. Maybe
they were influenced by the myth that digital is really new?
The tribal medicine man won in that ruling (history will
look back on it as a mistake).
[ Reply to This | Parent | # ]
|
| |
| Authored by: Anonymous on Sunday, June 10 2012 @ 10:27 PM EDT |
You tell me all software is math, and, at some low level, it is...
but that's a lot like telling me life is "just chemistry"...we can
reason from there and see how long it takes to justify your favorite form of
random mayhem....
Now, let's get onto a math example of current, practical utility: In the Flame
virus, reports are that someone clearly has an algorithm to get a given,
arbitrary MD5 checksum by changing a few "don't care" bytes in a
message, and that they used this to forge the Microsoft Update site.
Question: should the "prefix attack" on the MD5 algorithm be
patentable, given its obvious and specific utility? Yes, it's very mathematical,
but it is also quite concrete, clearly novel, and definitely useful.
[ Reply to This | Parent | # ]
|
| |
| Authored by: Anonymous on Sunday, June 10 2012 @ 10:46 PM EDT |
>>there's no getting around it: all software IS math.
No, software is something Brand New, it is much more than mere 'Maths'.
It isn't just Maths or Algorithms, it isn't (just) a recipe or design for
building/making something and it's a lot more than the fruits of labour.
I think there is a progression through the 3 sorts of I.P. protections we have
now:
- Things we make, artefacts.
[Trade Marks and Registered Designs]
- a Creative Expression, potentially copyable, more than simple artefacts.
[Copyright]
- Processes, Designs and Methods for creating 'stuff'.
[Patents]
It's worth noting that each successive step can embody the outputs of
previous steps. A patent for a musical device creates both Artefacts and
creative expressions of music.
But Software is a lot more than this:
it is the precise and complete embodied thinking steps to perform a
cognitive task.
If you remember the 3 basic building blocks of Software you can see the
difference to Maths:
- sequential statements
- repetitions and loops
- conditionals
Alan Kay described writing software as "building a convincing proof".
But it is more than just proving a hypothesis, its a result that goes further:
It combines Logic, Actions and Algorithms.
But there's more...
Software has the potential to embody any and all cognitive tasks of a human
mind that we can precisely describe. Is there any reason it can't Learn and
Adapt, fully emulating a human brain?
Already the best chess players are embodied in software.
There are symbolic solvers for maths and logic problems.
And automated reasoning solving Formal Proofs of problems (like Software
itself).
The central question here is:
Are Computers "Thinking Machines"?
If so, programs are precisely and fully defined sets of small-t thinking steps.
They only make sense when they are executed.
If not, then what are they? And does that create its own class of Intellectual
Property?
The Turning machine tells us what matters is not the engine, not commands
needed with an instruction set, nor even the human-readable input in
whatever way expressed:
its what executes, when translated, on an idealised platform.
Turing intended to only address 'computability', but he tangentially showed
that all machines of an architecture (eg. von Newman) are equivalent. It's a
harder problem to show that implementations in different architectures are
equivalent.
The world of AI is catching up to its promise of the sixties. It has the
aural/oral skills of a 2-3 year-old and the beginnings of visual skills.
Importantly, how anything is implemented in Silicon is very different than with
neurones, but the functionality is equivalent.
But is there any question that AI can one day implement all human cognitive
processes, given we can discover/define them? Are neurones (and hence
animal brains) a proper computing device and ultimately representable by a
Turing Machine? I think not.
Can we patent thoughts and patterns of thinking??
For me, this is what Software Patents are:
an attempt to gain a limited monopoly on a set of cognitive steps or
processes.
Now the contentious bit:
- there has to be payback for commercial efforts to encourage blind
investment,
- easily copied innovations shouldn't be ripped off,
- but trivial work shouldn't be rewarded, either.
Trying to extract rental for 20 years on mediocre software is monopolistic
business practice at its worst.
But forcing really great minds to give up all claims on breakthroughs seems
draconian to me as well. I couldn't have invented Postscript or the Adobe
graphics engines - they, IMO, deserve a commercial benefit.
Though a 20 year monopoly is way too long for a field that is barely 50 years
old.
Because Software is a series of cognitive steps, not a recipe/design for
anything, it requires a new name: "patent" does not cover this new
category of
Intellectual Property.
I think the question should be:
- what "degree of difficulty" of invention should attract any
protection?
- What maximum protection time is appropriate? 3-5 years, IMO.
- Can something like "Trade Marks" be available to prevent binary
artefacts
from being ripped off or translated? ["Want it? they write your own"
applies]
Something like this needs a large expert panel to adjudicate on, not a lone
examiner not even 'skilled in the arts', let alone an expert in the field.
Coca Cola has demonstrated its possible to keep Trade Secrets for many,
many decades. With the advent of the Gas Chromatograph, how can it still be
a secret? But it is...
Software houses that have really strong inventions should need only rely on
Trade Secret and licensing deals. With the ability to sue for unlicensed use.
If others can re-implement your software, especially quickly and easily, then it
wasn't that hard or inventive and not worthy of a Patent.
The current system has a gigantic hole, as far as I'm concerned:
if you want to lodge a Software Patent, then you shoul have to lodge the
Software... Ie. Everything needed to produce Running Code. Not snippets, but
a complete, running and usable implementation. Compile the code and see
the full product...
This is effectively what we see in Open Source Projects.
The Real Deal, not some vague 'back of envelope' descriptions, potentially
incomplete and certainly devoid of 'full disclosure'.
[Steve Jenkin, Canberra. 11-Jun-2012][ Reply to This | Parent | # ]
|
| |
| Authored by: Anonymous on Monday, June 11 2012 @ 12:19 AM EDT |
Everything software does is mathematically describeable.
However, the same holds true of the planetary gear. Does
that make the planetary gear pure math? Did it make the
planetary gear unpatentable?
I presume it's obvious that both may be answered in the
negative.
So the fact that it can be described mathematically does not
make it pure math or unpatentable automatically. (However,
I'm sure that a number of software patents could be
classified as pure math...)
However, note that the planetary gear is *not* the same as
the crank it was intended to replace (due to the patenting
of the latter). Both provided roughly the same funtionality.
Similarly, patents must be specific to an approach to
implementation.
Many software patents should not have been approved, because
they
1) Describe a solution in terms broad enough to cover
dissimilar approaches (1 patent for cranks and planetary
gears) to solving one problem.
AND/OR
2) Use terms so abstract that they describe a mathemematical
approach or algorithm instead of an implementation (patenting
the math behind the planetary gear)
However, this doesn't mean that all patents fall into one of
these traps. If they don't, let them be.[ Reply to This | Parent | # ]
|
| |
| Authored by: Anonymous on Monday, June 11 2012 @ 09:58 AM EDT |
Math, short for Mathematics, is fundamentally a conceptual modeling system.
While software does include a conceptual model, the part you refer to, it has
one requisite that is absent in math, functionality.
Without a mathematician, engineer or scientist to give volition, implementation
and/or interpretation to a mathematical model, math is void of functionality.[ Reply to This | Parent | # ]
|
| |
| Authored by: Anonymous on Monday, June 11 2012 @ 09:08 PM EDT |
First, I think Mr. Risch is onto something with his idea of utility as a key
to controlling patents (generally, and software patents in
particular).
Second, I do understand the argument that “software is
mathematics”. I even agree with it. I simply think that it doesn’t
matter.
As the noted legal commentator Lawrence P. Berra observed, “in
theory there is no difference between theory and practice. In practice there
is.” The law aspires to be a consistent system of high principle and public
stewardship. It turns out to be an ad-hoc, patchwork scrum that frustrates and
infuriates, yet still manages to keep most of us from bashing each other over
the head with cudgels and axes, most of the time.
The law is, in
practice, whatever generations of Philadelphia lawyers have gotten judges to
sign off on, and K Street lobbyists to get Congresscritters to vote in. And I
say that without cynicism because they’re just doing their jobs -- it is
their canon duty to get the best deal for their clients.
I don’t mind
that, because the law is also whatever we can get our Congresscritters
and our fellow citizens to enact. If you don’t believe that you can effect
change on this issue, stop reading and go do something more entertaining; the
rest of us have work to do. That’s what Mr. Risch is doing. And if you don’t
think he goes far enough, well, once we get him up to speed, let’s don’t stop
pushing, nu?
Mathematics is an idea (or a collection of them), an
abstraction. At its highest level, mathematicians take delight that it is
content-free. And the letter of the law says that an idea (alone) cannot be
patented. (Nor copyrighted. Trade secret is about all you can do with
one.)
Yet ideas are patented every day -- in software, in business
methods (which are not math, or at least are less obviously math). For that
matter, any invention incorporates an idea. If you disallow patenting
ideas, then you disallow patents. (Which may be your preference, but really?
ain’t gonna happen.)
The distinction seems to be that the idea is
attached to something tangible (for values of “tangible”) -- and further, that
the attachment is necessary to the invention, not merely incidental (or
contrived for the legal purpose of making the abstraction patentable). We need a
good definition of that distinction, and then we need to push it -- to Congress,
to whoever the Nimmer of patents is, to the courts -- against the dogged
opposition of those whose job it is to build a system that they can game. (It
is, by definition, not cheating if they get the rules written their way. So
let’s don’t let them.)
It will take years.
I got
years.
I very much like, as Mr. Risch does, pem’s notion of
crowd-sourced obviousness and prior-art search. I would love to see that written
into law. I do note that an inventor, simply by publishing a problem, thereby
calls to the public attention that it is a problem worth solving. That might
push an inventor not to start the patent process at all, but to rely instead on
the much more limited protection of trade secrecy.
As to obviousness
(and continuing the baseball flavor), Mr. Berra's fictional counterpart Lou
Brown once stated the principle that distinguishes the new from the nothing:
"That's a hell of an idea." If a court or jury can't say that of the patent, it
ain't.
Finally, going back to first principles:
To
promote the Progress of Science and useful Arts, by securing for
limited Times to Authors and Inventors the exclusive Right to their
respective Writings and Discoveries.
-- Article I, Section 8, Clause 8, U.S.
Constitution (emphases added)
I would favor requiring patent
holders to show, within seven years of grant, that their patent has produced
salable products, employment, or otherwise contributed to the economy, or their
patent would expire early. Merely using a patent to block others from coming to
market would not qualify.
Well, it’s a start!
[ Reply to This | Parent | # ]
|
|
|
|
|
