Slashdot Mirror
Courts May Revisit Software Patents
An anonymous reader writes "It looks like the courts may finally be gearing up to overturn the ruling that opened the floodgates for both software and business model patents. It's been nearly ten years since the US courts decided that business methods were patentable and that most software could be patentable — and we've all seen what's happened since then. With all the efforts to fix the patent system lately, it appears that the court that originally made that decision may be regretting it, and has agreed to hear a new case that could overturn that ruling and restore some sanity to the patent system."
Think of the people that patents DO benefit... I'm thinking any lawyer fighting to abolish patents won't exactly be pushing themselves to win the case...
Code, Hardware, stuff like that.
I just can't see the CAFC reversing themselves to any significant degree here. Most entities involved have way too much to lose.
I predict this will get quietly swept under the rug...again.
____
~ |rip/\/\aster /\/\onkey
While I may not agree with software patents I can't agree with your argument, it seems analogous to "If you can't patent chemical elements then why should you be able to patent devices, as they are nothing more then chemical elements."
Hmmm, what exactly is NOT maths in this world? :)
"Give me six lines of C++ code written by the most competent programmer, and I will find enough in there to hang him."
It's called greedy reductionism, or "nothing buttery". It's the first line of defense for reactionary or fanatical Slashdot trolls.
However, a lot of the software patents seem to be based on algorithms, and not a whole lot more.
-mrxak
Onions Will Kill You
patent on devices ok, patent on the maths not ok. most of the software patents I've seen are on the maths not on the actual specific device. If you start putting patents on a general class of devices using the maths then your probably just patenting the maths.
thank God the internet isn't a human right.
Hmmm, what exactly is NOT maths in this world? :)
Patents are meant to cover a particular implementation in physical terms of a theoretical idea, and right now, they are often being used to try and cover the theory as well.
For example, my own pet hobby is working on a new way to factor large numbers. Let's say that my redneck republican self gets insanely lucky and bumbles into an algorithm that actually factors something in polynomial time, or even close enough to it so that RSA and the like are untenable. Since my approach depends on treating factor as a decision problem, it follows that if I did get really lucky and struck gold, that, it would be applicable to a wide range of other problems. Under today's law, patenting that would basically give me the right to apply that mathematical breakthrough for my own ends, when clearly, its in the interest of society that as many people should be allowed to exploit it. Basically, I would be allowed to charge money for any sort of an implementation of a combinatorial problem, which is absurd. Yes, I might theoretically build a billion dollar enterprise to milk this concept for all its worth, I would ultimately though screw everyone else with whom such a breakthrough might be useful, and damage the overall economy that many millions of times more.
Really, the dividing line is one of information and knowledge versus an actual real world device. As Jefferson so adroitly pointed out, information does not lose its value when it is copied. If I know something, and give that information to you, we both know something, and that doesn't hurt me that you know it. It does mean that I can't build some sort of an empire at your expense, but, given that we already went through the Catholic attempt, and then the various State attempts, to monopolize information, with disasterous results all the way around (and not a single success in 2000 years!), it is obvious that a social framework which allows information monopolies works to the disadvantage of mankind.
Quite ironically, those people whose livelihoods depend on information having value are the ones most arguing that information ought to be free. Patents are, in theory, today, supposed to protect IT workers and their inventions, but most GOOD IT workers these days remember that computer science as a field advanced even more before today's patent nuttiness. If we did anything, it would be to allow the shared discovery and utilitization of new techniques, but protect, if desired, commercial and open implementations. So, for example, if Microsoft invents a new GUI dongle, or on the flip side, someone invents a browser plugin, then, it would be better for everyone if you simply could not be sued for making your own implementation of that idea. That gives us a world where everyone's products can advance, we have IT for our customers and ourselves and leave the lawyers out, and everyone is happy.
It is really only the idiots at Wall Street, that have handed us the internet boom mess, the present mortgage mess and the previous S & L mess, that want to maximize every asset as much as possible with silly things like patents and create yet another bubble that will burst and screw the rest of us up. But really, Windows doesn't need any patents any more than Linux does. The value of both of those products is predicated on their overall customer experience, not some silly mining like claim staked out in Washington DC!
This is my sig.
that's the problem! They aren't! One-click isn't an algorithm, it's an obvious idea.
-1 not first post
A better analogy would be "if you can't patent chemical compounds, you shouldn't be able to patent drugs, as drugs are nothing more than chemical compounds".
...
Software is just a subset of math, just as drugs is a subset of chemical compounds. Your analogy make it sound like math is at a totally different level of abstraction than software. It isn't.
Or maybe we need a car analogy to make it perfectly clear
It's been nearly ten years since the US courts decided that business methods were patentable and that most software could be patentable -- and we've all seen what's happened since then.
Since then we've seen the emergence of Google as a powerful challenger to Microsoft. This is one example among many of a company whose entire existence, much less its massive success, is dependent upon a patent (# 6285999, in Google's case). If Google had not been able to patent its major innovation, then Microsoft could easily have co-opted the idea, and it would have dominated search as well as operating systems and office suites. Recall that most of Microsoft's meteoric rise took place during a time when software was not patentable. The absence of software patents is precisely what makes "embrace, extend, extinguish" possible. Software patents give the original innovator the power to stop that strategy in its tracks.
I believe that what Slashdot readers truly dislike are bad patents, not software patents per se. Software just has more bad patents than it should because of the way the PTO treated them. Until the courts basically forced the PTO to accept software patents, the PTO did not hire computer scientists as patent examiners. Even now, the PTO has a massive backlog of software patent applications, and as a result computer related applications have by far the longest median time to issuance (roughly 44 months!). The PTO tends to err on the side of issuance, and so we end up with a flood of terrible software patents.
How to fix this? The simplest way is to eliminate the presumption that patents are valid, which requires a patent challenger to prove invalidity by clear and convincing evidence (a standard almost as high as beyond a reasonable doubt). Instead, we should recognize that many patents are not valid and end the presumption of validity. That way, bad patents can be more easily challenged, and patent trolls will think twice before bringing spurious suits.
Why wouldn't the lawyers fight hard on both sides of the case? That's what they are paid to do. Court cases that shift around the boundaries of patent law do not hurt or help "the lawyers". It just changes where the fighting is but there will still be fighting.
For example, the recent KSR v. Teleflex case changed the rules on obviousness. That created lots of work. Suddenly there were new arguments to make in seeking new patents. There were also new arguments to make in invalidating old patents. Lots and lots of work.
The only way us patent lawyers would actually loose is if patents were outlawed.
I am a lawyer, but not yours. Anything I tell you might be a total lie intended to benefit my clients at your expense.
The problem is that nearly all algorithms are obvious to one with ordinary skill in the art once you tell them what the inputs and outputs are. That's why patents on software are stupid. The only exceptions are those in which ordinary skill in the general art of computing are insufficient to understand what is meant by the description of the input or output.
For example, someone with ordinary skill in the art won't understand what a discrete cosine transform is, so describing a JPEG decoder as "an algorithm that generates a pixel array from a compressed image consisting of a series of XnY blocks (where X and Y are usually 8 or 16) in the form of a discrete cosine transform block compressed with Huffman codes" will mostly result in blank stares. However given that description and a notion of the order of these blocks, someone with average skill in the more specific art of image compression could readily write a JPEG decompression algorithm. It would take a while because the individual components are pretty complex, but the simplest complete description of the inputs and outputs is sufficient to completely define the entire algorithm.
For simpler software patents, this goes without saying. "An algorithm that takes as its inputs a click on a particular item in an HTML page and a stored representation of the user's credit card information on a server and generates as a result an order for the displayed product charged to that user's credit card" is a perfect example. Anybody who has ever written the simplest web app could do this in five minutes. Maybe they couldn't do it securely in five minutes, but the amount of thought that goes into taking those inputs and generating the output is negligible.
The only patents that should be allowed are those in which the a complete description of the inputs and outputs is insufficient to give someone with typical skill in the very narrow field of writing similar software enough information to write the software. For example, as someone who is getting annoyed at camcorders that do image stabilization based on what percentage of the image changes, I thought about "an algorithm that takes information from accelerometers and an image from a CCD and generates a motion-smoothed output". That patent might be acceptable under these rules because the process for manipulating the incoming data into a reasonable set of output data is not obvious, and indeed, there are many different smoothing techniques that one could come up with to muck with the accelerometer data and distinguish between jitter and a pan/tilt. On the other hand, it should only cover the use of a specific smoothing algorithm for that purpose, not the general process of smoothing, nor the definition of that smoothing algorithm for other purposes.
Check out my sci-fi/humor trilogy at PatriotsBooks.
I guess that many programmers will be anti patent for a number of reasons:
You left out the biggest reasons to oppose software patents, they stifle innovation and they lockout others who come up with something innovative.
FalconShould there be a Law?