Why Consider Linux Kernel Patent Risks?

chromatic writes "After the hoopla about OSRM's study of patent risks in the Linux kernel, I talked to Dan Ravicher, the patent attorney and free software afficionado who conducted the study. Contrary to my initial reaction, I've come to believe that the study is actually very valuable. Linux and Patent Risks on the O'Reilly Network explains why."

Comments

Linus and patents

[blogtim]

Linux founder and leader Linus Torvalds has taken that approach. "Finding patent infringement has always been a responsibility of the patent holders," he said in a 2003 interview. "It is a fact that I do not encourage engineers to look up patent information

Actually, I read an article in Wired maybe a year ago quoting Torvalds as saying that he personally avoids looking at patents because willful violation leads to triple damages. Better just to close your eyes and do whatever...

Re:Linus and patents

[Bruce Perens]

Imagine if my lawyer does some research and finds out that my technology potentially violates several patents. And then I tell him, don't tell me, I want to be protected.

This would be covered under attorney-client privilege. Neither you nor your attorney can be made to testify about what your attorney told you. However, what I tell you as a consultant rather than an attorney is not similarly protected. I have indeed told some customers that they should not see my findings because my findings were not favorable for a court case they were considering. What they didn't see, they would not have to testify about. They agreed, thanked me for being so careful regarding their interest, and happily paid me for a report I never delivered!

In this case, if you know you are infringing, the amount of damages you may have to pay goes up by more than three times. That's how the law works. It sure isn't OSRM's fault.

Bruce

Re:Linus.

[MarsDefenseMinister]

Linus owns the trademark to the kernel, but not the copyright, except for the portions he wrote himself. As far as I know, Linux does not require copyright assignment as a condition for code inclusion.

Copyright != Patent != Trademark.

Re:Software Patents vs. Free Speech

[OwnedByTwoCats]

It seems to me that an individual with ordinary skill in the art of developing software, coming up with a patented solution to a problem, poses as much a problem for the owner of a patent as it does for the individual developer.

One of the requirements for granting a patent is that it describe an invention or process that is nonobvious to one skilled in the art...

Re:Would it be illegal?

[ron_ivi]

" First you would have to have the source code. I suppose it could be de-compiled? I don't know how that works. Are there laws against that?"

Even proprietary vendors make their source code available to important customers "to conduct security reviews of the products" - why shouldn't they do the same for Intellectual Property Rights?

Furthermore, you could have third-party auditing companies provide that service in the same way they audit other confidential information such as company finances.

Re:implied patent license

[killmenow]

Also, thanks to Groklaw, I learned about this thing called Promissory Estoppel that would pretty much stop them from bringing a patent suit at this point.

Re:So....

[spacefrog]

• If somebody decides to sue Linux for copyright infringement, who defends it?

IBM, Apparently.

Re:implied patent license

[Bruce Perens]

To a great extent IBM has avoided distributing GPL software directly. We think this is because they don't like the patent terms of the GPL.

Bruce

Re:It's a MAD World!!!

[Bruce Perens]

It does not help that IBM, HP, and MS violate each others patents, because they are already cross-licensed with each other.

They can't enforce their patents against each other except in unusual cases. MS contract with HP allows MS to enforce against HP in regard to specific free software, as we learned in the infamous memo, but that memo is all that we know of the MS-HP cross license. We do not know of any similar carve-out that would allow HP to enforce against Microsoft.

Bruce

Re:patents free for private use ?

[Meowing]

Yes, private use and research often qualify for royalty-free use of a patent, but beware that these exemptions are very narrow. More information here.

Also, the international treaties covering patents are still evolving (there's a bunch of 'em -- not just WIPO but PLT, PCT, TRIPS and more I'm sure I'm forgetting), so it's still necessary to check out the local laws.

Re:Has anyone audited non-free OS's for patents?

[Bruce Perens]

In Microsoft's case, we know that they have offered to indemnify their customers regarding patent risk in their software. So, even if MS has something to hide, the customer is theoretically off of the hook. In practice, the customer is potentially in the position of having to sue Microsoft in order to get them to make good on their indemnity. Even the Federal Government doesn't have much luck at suing MS, so the customer can only hope that MS pays without a fight.

Most other companies could not offer to pay their own indemnities, and many of them believe this is covered by their liability insurance when that may actually not be the case.

So, I think it still turns out that code that is open for all to view is better.Bruce

Re:So....

[Bruce Perens]

Well, in the U.S. it isn't necessary to figure out who is responsible for the infringement before you decide who to sue, because the patent title says that use is one of the infringing acts you can sue for. This is why there is even a question that users can be sued, we would just love to have a court rule that they can't be.Bruce