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."
Re:That is logical from MS' point of view (Score:5, Informative)
When you have the whole market, you can only grow at the growth rate of the market. Microsoft investors and Wall Street are asking MS to look for new avenues of growth, and patent licensing is one of them (XBox is another).
It's extremely hard for MS to find any new areas that can get significant enough to impact their bottom line.
So expect a lot of litigation, or at least behind-the-scenes dealmaking.
Re:Seems on the level. (Score:5, Informative)
No they don't, that's Trademarks. Trademarks can be renewed, and have to be defended.
Patents cannot be renewed, are valid for 21 years after issuing (in the US), and can be selectively defended without any fear of losing the patent.
Copyright also is valid for a time period (but that keeps on getting extended by Disney&Co's hired congresscritters) and also does not *have* to be defended to remain valid.
The most important defenses against patents are more patents (big companies give each other rights to use each other's patents, something Open Source cannot do) or finding some reason for a patent to be declared invalid, either by finding Prior Art or by showing that a Patent is "obvious to a person skilled in the field".
No matter what, getting a patent declared invalid is not cheap.
Re:So can somebody explain me this? (Score:1, Informative)
Re:Slight error (Score:5, Informative)
Apart from that, the GPL only addresses copyright. If there are patent issues, you have to take care of those separately before you can use the code.
Re:Prior art database (Score:4, Informative)
Patents are not licenses (Score:5, Informative)
The patent holder can license their implementation however they choose. It is perfectly legal to expressly say that your code may be used by GPL applications as a shared library, the same way that LGPL code can be used by commercial source.
Otherwise you'd best get rid of the NVidia drivers, the commercial databases, and every other non-GPL product that runs on Linux. That's not the intent of the GPL, never was, and it's nothing but Microsoft FUD bolstered by uncertain IP laws which suggests otherwise.
The GPL is about ensuring that if you use GPL code for your work then it must be licensed under the GPL, not that the GPL can't use any libaries which aren't expressly GPL'd. Contrary to any fantasies the full-bore GPL zealot might have, there is no legal way you could do that, and it would be against the very spirit of freedom that led to the creation of the GPL.
Also note they are not talking about refusing to license the IP for specific projects, but to the distributor of the source. In other words, they've decided to play their patent portfolio as a market-blocking threat the same way they yanked all the incentives from hardware manufacturers who dared ship something other than Windows.
As far as I'm concerned, Microsoft has just expressly stated their intent to maintain monopoly profits and control through IP barratry and by locking out any company which dares support or distribute GPL software.
If that's not a RICO violation, I don't know what is, and it's well past the time we stopped putting up with this crap as an industry. I greatly look forward to Sun, IBM, and every other IP holder banding together and teaching Bill that he does not rule the world, no matter how much he'd like to.
Re:Slight error (Score:4, Informative)
I've got loads of GPL'ed stuff I've not given away to anyone. You don't have to give it away even if you're a developer - you're perfectly entitled to keep it in-house. However, if you *do* give it away, you have to make the source available.
Shouldn't that last line read "However, if you *do* publish the binary, you have to make the source available." ?? IIRC You don't have to give away (gratis) the product to be tied to making the source available. If the binary is not for internal use only (i.e. you're selling it, licensing it or giving it away) then the source must be also made available, and for no more than the costs incurred in distributing the source.
US = first to invent not first to file (Score:2, Informative)
With first to file, the first to file for a patent with the patent office gets the patent. It's simpler and it's what the rest of the world does.
(No doubt I'll be corrected if I'm wrong...)
Re:That is logical from MS' point of view (Score:3, Informative)
It takes even a good lawyer years just to understand what the patent actually says. Combining legalese with technobabble is an excellent way of generating truly incomprehensible (and therefore useless) documents. "Patently obvious" long ago became an oxymoron.
Re:That is logical from MS' point of view (Score:3, Informative)
Re:Patents are not licenses (Score:5, Informative)
Don't hold your breath waiting for Sun to join that fight. They executed a patent cross-license deal with MS about a month or so ago. Sun agreed to drop a lawsuit they had filed, and MS gave Sun aboot $2B.
Patents were intended to promote innovation. Unfortunately, they are more and more frequently being used to stifle innovation instead.
Re:Seems on the level. (Score:3, Informative)
Not necessarily. Non-discriminatory licensing is not actually required. I can license my patent to MS for $10 per copy, and to Linux for free, and to my little brother for $1000, if that floats my boat. If I had a patent.
Not that I am trying to imply that MS will do "the Right Thing" (whatever that is) with its patent arsenal.
Re:So can somebody explain me this? (Score:3, Informative)
Re:new trend in IP (Score:2, Informative)
Submarine patents are a nasty and relatively recent development in the intangible property arena. With the sloppiness of the USPTO, it benefits corporations to spend the relatively small amount of money to maintain a technology in the application process until favorable market trends develop (by filing continuations, etc.).
There's a guy named Lemelson who would be surprised to hear that submarine patents are a relatively recent development (actually not, since he's dead). Try googling that name and patents. Hint: 1950s.
There have been changes to the law since 1995 that make it extremely difficult to obtain anything resembling a classic "submarine patent". In most cases, a patent application is published 18 months after filing (it is possible to avoid this). In all cases, patents filed after 1995 (original applications and continuations) have a terms that last for 20 years from the FILING date, which means the longer an application is pending before the USPTO, the shorter the time available to assert the claims against alleged infringers. The USPTO has also made it very difficult to keep an application pending before the office for long periods of time (more than a handful of office actions).
Then, when the application is finally approved, the "inventor" receives priority status because our system, for better or for worse, favors the first to file.
This is "the Earth is flat" wrong. The US has a first to invent system. It always has, see 35 U.S.C. 102(g). Plus your point ignores that if the invention is publicly disclosed, demonstrated or sold, it's prior art and you don't even have to argue about first to file/first to invent.
Once this occurs, said corporations can then leverage their patent portfolio... to obtain compliance from those unwillingly infringing.
No matter how you read this, it's a screed against the idea of patents, and centuries of legal history and multitudes of countries clearly believe that you're wrong.
Unlike copyrights or trade secrets, the essence of patents involves disclosure of the idea to the public. Unlike copyright or trade secrets, you can copy directly from the patent so long as you don't infringe the claims. That's the bargain, live with it.
The only "unwilling" infringer is the infringer that does not know the full extent of the state of their art. Since nobody has perfect knowledge, infringement merely means that you're liable for damages equal to lost profits or a reasonable royalty (remember, in theory the inventor is supposed to be compensated for disclosing their invention). A "willful" infringer is someone who has actual knowledge of the patent and choses to ignore it. They can be punitively punished by so-called "triple damages".
Unlike trade secret law, or copyright, there is no innocent infringer or independent creation defense, so an infringer is faced with a situation where they are forced to pay lost profits and/or treble damages in addition to having their inventory siezed.
The screed continues, with damages hopelessly muddled into a FUDdy tangle of authoritative words. If someone is genuinely innocent, they can negotiate the lost profits or reasonable royalty and dispose of the problem at a relatively low cost.
Because a patent litigation suit averages about $3 million, only the hardiest of defendants can afford to challenge a patent's validity.
Luckily, the USPTO provides a process known as reexamination that costs between $2500-$8500 to file, plus the costs of whatever time you care to invest (if you chose inter partes reexamination), and the USPTO will happily use your allegedly killer prior art to limit or invalidate the patent owner's claims.
Costs a factor of 1000 less. You conveniently forgot that one.
Economically, it is often a better option to simply succumb to licensing fees.
Or design around
Re:new trend in IP (Score:3, Informative)
My father is an engineer who has been involved in a number of patent applications going back to the 70's. He once mentioned that using non-industry-standard language is, and long has been, the common practice, because it increases the likelihood that competitors will not notice the patent and start infringing. Intentionally abusing the pitiful review process may be a recent development, but submarine patents are standard operating procedure.
Re:Patents are not licenses (Score:1, Informative)
Ah, that clarifies your point quite a bit.
The truth of the matter is that software patents are not loved within the USPTO, either. The USPTO doesn't make the laws or court decisions, however, and the patent attorneys will be happy to ruin an examiner's career if it gets his client a patent. This is why "software" is not patentable in the US, but "executable instructions on a computer readable medium" is patentable. Some lawyer dragged it through court and now the USPTO follows those court decisions.
I haven't read the "one click purchasing system" patent myself, but I am aware of it. As a patent examiner myself, I'd like to point out that examiners do not enjoy issuing patents for ideas they hate. Examiners don't get to say, "This will be a bad patent, therefore I won't grant it." The applicant has a lawyer and that lawyer will get the patent if it is allowable under the law. The USPTO doesn't make that law. If you're upset about these software patents, look to the lawyers, not to the USPTO.
That said, sure, examiners are only human and sometimes bad judgements are made, but the foundation of the examination process in software is a bunch of computer geeks who do their best to reject applications day in and day out. The USPTO doesn't really grant patents, it's more like they fail to reject them, and the only reason they fail is because the applicant's lawyer missed his rabies shots.
And I'll make this one anonymous because the USPTO does have rules about what an examiner can say publicly about the office and I don't feel like getting extremely familiar with them.