Microsoft's Marshall Phelps On Patents And Linux 282
An anonymous reader writes "Microsoft's Marshall Phelps says he is running 'a licensing shop, not a litigation shop.' Bill Gates's intellectual property guru talks to Brad Stone about Redmond's new emphasis on patents, why he can't license Microsoft IP to distributors of open source software -- and why he shouldn't be feared."
So can somebody explain me this? (Score:3, Interesting)
Re:Seems on the level. (Score:5, Interesting)
I don't think they'll be losing much ground with standard Joe Consumer any day soon. Even if people do hear about how MS is or isn't bad, it doesn't effect their choice. People are brought up on MS, and companies like Dell and Gateway forcefeed it to their customers.
(I am aware this is only one portion of the market, I am not talking about corporate use.)
A great way to deal with the issue. (Score:5, Interesting)
MSNBC slammiing Microsoft (Score:5, Interesting)
From the article: The overtaxed U.S. Patent and Trademark Office often grants absurdly broad patents that reflect little actual innovation. (For example, Microsoft owns a patent for activating a program on a handheld device by holding a button down for several seconds).
Is it just me, or is someone at MSNBC got a hate on for Microsoft? First the reccomend Mozilla and now they slam them for patents. I am guessing there is some behind the scene tension there.
So, just licencing IP then, no lawsuits? (Score:5, Interesting)
"Marshall Phelps tries to dispel the notion that Microsoft is preparing a patent assault on open-source software. He notes that at IBM he never initiated a single lawsuit"
Will this mean that the 27 seven unidentified patents that Open Source Risk Management [osriskmanagement.com] have found can just remain then? Maybe not. If those 27 patents are identified under a contract of no disclosure to key kernel developers, will the community be able to implement workarounds before the suits starts to rain?
I would be very interested in knowing who else owns a patent that yet have not been tested in court that covers code in the Linux 2.4 and 2.6 kernel-series. Without knowing if they have been tested in court or not we will just have to stick with the numbers made public then: IBM has 60 patents, Microsoft has the 27, 20 has HP and 11 goes to Intel. That leaves us with just above 160 other unidentified patents to deal with.
Microsofts 27 patents are most likely the worst here, but have they been tested in court? Things like their silly patent for "activating a program on a handheld device by holding a button down for several seconds" will most likely not pass any court, or I for sure hope not.
I think that the SCO-case will set a precedence so that other stupid claims are thrown out without years of countersuits and motions in all directions. Lets atleast pray it does. I will.
Re:Seems on the level. (Score:5, Interesting)
then if the company a can point to the Open Source infringer and say the response would be along the lines of
Most large patent holders dont want to rock the boat too much in case they give the anti-sw-patent lobby too much ammunition. Software patents are a goldmine for large companies. If patent battles start descending into lots of legal wars, Congress will have to take action, even if the Senetor's Corparate owners dont want it.
Besides, Microsoft have already managed to dodge anti-trust issues in the US, I suspect they will wait a few years before they judge it safe to try to kill a competitor in such an obvious manner again.
Comment removed (Score:4, Interesting)
sigh. (Score:5, Interesting)
Microsoft, even granting the somewhat tenuous proposition that they're doing this for good, will soon be a different set of people. You know. Darl McBride junior and his buddies.
So okay, Marshall, I'm sure you're a good guy and all, but I'm not giving you kudos for playing along with an evil, broken system "for good." I'd Microsoft spent its seven hundred godzillion dollars helping bring about reform in the patent system and changing its abuse-of-monopoly behaviours.
No offense.
Patent Suit Defense Fund? (Score:5, Interesting)
While reading the article, two phrases struck me: a research group is prepared to list 283 patents violated by Linux; and half of all patents are defeated in courts. (I believe the latter is half of all challenged patents.)
Would it be difficult for an organization, say the EFF or the GNU foundation, to set up a specific fund for collecting donations to be used only to defend patent law suits?
Imagine the war chest available if half of the Linux users donated $10 to this fund... And it'd be tax deductable in the States, too!
Grab the money and run. (Score:5, Interesting)
Re:Seems on the level. (Score:2, Interesting)
Re:Seems on the level. (Score:1, Interesting)
MS: "You're infringing our patent, pay us $x"
Company A: "But they're infringing too, and they're not paying?"
MS: "I know. But we've required them to pay a redistribution fee in order to redistribute the patented code. That fee is the release of their code under the GPL. Feel free to pay the same cost"
THAT is what patented code in GPL works means. You are free to use it as long as you obey the GPL. There are so many companies that are afraid of the GPL they would *gladly* pay money rather than *possibly* have to have their code GPL.
Re:Prior art database (Score:3, Interesting)
History Lesson Time (Score:5, Interesting)
Now one way that you can make money out of your invention is to persuade a backer to lend you enough money so you can afford the tools and materials to start building it. But in order to do that, you need to convince your sponsor -- who is in all probability a banker or financier, not a scientist or an engineer -- that you can earn enough money by selling your invention to eventually pay them back, and that requires either an unparallelled degree of chutzpah or some kind of official document stating the worth of your invention. Another way is to get someone else to build your invention for you. But in order to do that, you will have to tell them how to make it -- and once they know that, they can cut you out of the loop. They have the invention, they can afford to make and sell it, why should you get anything?
The patent system was set up to solve both these problems. You demonstrate your invention to a trustworthy body, and say that you are prepared to share it with the world at large; and in return, you are given a copy of an official letter which describes it in full and states that you are the true inventor. The original is held in a library where anyone can look at it. This letter also grants you, for a limited time, exclusive control over the commercial application of your idea. Now you can seek assistance, confident that you will be able to earn the true worth of your invention: a group of experts have attested to the fact that it really works (so nobody can think you are trying to rip them off with vapourware), and nobody else can claim it as their invention and rip you off. Then, after you have had a fair chance to get rich off your invention, it gets formally released so everyone can have a bite of the cherry -- which is your little way of saying "thank you" for all the inventions and discoveries which have come before and from which you have already benefitted; such as fire, tools, agriculture, sanitation, electricity, and so forth.
If your invention is a piece of software, which is something which can be reproduced at no cost, then you are by definition not too poor to make it. But, additionally, some things should never be patentable. Mathematical processes, for one. What if integration were the subject of a patent claim? Integration is a mathematical concept that crops up time and time again in the real world. Would you have to pay a royalty every time you poured a liquid from one container into another? What if subtracting one were patented, and the patent holder refused absolutely to licence it? Would adding one and subtracting two be permitted as a work-around? Software is just a formalisation of a mathematical process.
The most fundamental thing wrong with the US patent office today is that patent applications are not being properly tested. I believe firstly that the requirement to produce a working prototype should be reinstated. A patent application not supported by a prototype is nothing more than a work of science fiction -- and in any case, if you are not good enough to build a prototype, then perhaps you do not deserve to be recognised as an inventor. The question of licencing needs to be addressed -- I firmly believe in non-discriminatory licencing, in other words that a patent should be licenced to everybody, and everybody for the same price, or nobody. Additionally, a procedure needs to be created for verifying that an invention is original -- and for dealing with exceptions. Since this is an example of civil rather than criminal law, the terms "innocent" and "guilty" do not really apply, so the question of burden of proof is a thorny one. Finally, there need to be clear and unambiguous rules about what can and what cannot be patented; and, for the inevitable case of an invention which is so new that none of the existing rules can be applied to it, why.
Re:So can somebody explain me this? (Score:2, Interesting)
Basically, the competition looks at where MS violates their patents, sees where they violate MS' patents, and sign a cross-license to make the whole problem go away. If both sides have enough patents, they can force a stalemate.
Re:Prior art database (Score:3, Interesting)
Article is written by Brad, not Phelps (Score:3, Interesting)
Re:History Lesson Time (Score:2, Interesting)
Re:Seems on the level. (Score:4, Interesting)
But either way.... they do get to act like rulers of the PC world, because they are.
Re:Patents are not licenses (Score:3, Interesting)
I know that sounds crazy... but sometimes you have to use a system to make it break, or just defend yourself before it crumbles.
I would consider patenting and licensing to anyone that follows the GPL. If a competitor comes knocking, I have patents to defend myself. Heck, imagine a hundred FL/OSS companies with GPL-patents all counter-attacking any commercial company stupid enough to attack any of its members. It couldn't be done at first for blatant infringement, but more power would allow such an alliance to challenge the very system.
A couple more behaviours that would be interesting would be to license to any commercial company that wanted to use them (again for free) as long as they either don't have patents or make theirs available for free to the GPL-world.
Dunno, maybe this is crazy...
Re:Seems on the level. (Score:3, Interesting)
The obvious answer is, "They don't have any money; you do.". Look at Fraunhoffer, LAME and MP3.
Re:Patents are not licenses (Score:2, Interesting)
What I mean is that opensource has a lot easier to prove prior art than most companies that has records that often aren't as accurate or trustworthy.
Re:Slight error (Score:3, Interesting)
Besides, the GPL doesn't lay down the required mechanism for obtaining the source code. I could publish a binary version and say, "for the source code, you must hand-deliver twelve 5.2" floppy diskettes to an address in New Zealand. It does meet the criteria listed. There are plenty of ways to abuse this that are far worse than simply not publishing your source code until the final release when you've had time to delete the swear words from the comments.... :-)
Re:Why not the EFF - Electronic Frontier Foundatio (Score:3, Interesting)
Doesn't an ordinary practitionar of the art coming up with a patented solution independent of the patent refute word-for-word the non-obvious requirement of a patent?
If not, why not?