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

implied patent license

[bollow (a) NoLockIn]

From the article

Ravicher discovered that open-source-friendly companies (including IBM and HP) hold about 100 of those patents. Again, the likelihood that such a company would bring suit against someone using or distributing Linux is small, especially since those companies often distribute Linux themselves. (Legally, a company probably could, but it goes against the spirit of open source.)

The assertion "legally, a company probably could" in the above statement is false. Even though it's primarily a copyright license, the GPL contains an implied patent license.

Re:implied patent license

[Richard_at_work]

This is an interesting one. Reading section 7 of the GPL on gnu.org seems to indicate that having a patent issue on the code DOES NOT excuse you from the obligations under the GPL (namely distribution of source code). Indeed, what it does indicate is that you STILL have to fulfil the source obligations under the GPL to those who you distribute it to (and the permission propagates down the chain).

But reading it carefully, it seems to indicate that you are not giving a carte blanche license TO THAT CODE. The 'implied patent license' only applies to those YOU HAVE DISTRIBUTED TO (and on down the chain). This means that the linux kernel IBM has for download may be free from patent issues, but the copy on kernel.org and in redhats distribution may not be. I dont think that was an intention of this paragraph but after running it past a friendly lawyer (hey, dont hear that often do you? :) ) who agrees with me, it might be something that needs looking at.

Also, another problem with the GPL is that it attempts to 'trump' any rulings by the legal courts. Paragraph 7 extends itself outside of patent issues and INCLUDES COURT JUDGEMENTS when it says that they do not excuse you from complying with the license. Personally, I think a court ruling that I cannot distribute the code (for example, my offices and actions are frozen by the court during a federal investigation) would superceed the lowly GPL in that case (indeed, as carrying out the obligation would land me in contempt of court and potentially with a criminal record for such).

Just some food for thought.

But, if they do find someone to attack

[Trizor]

If the point is negledgable they could get a compotent opinion form a patent lawyer that says they think the patent is bogus. Then they are an unknowing offender and only have to pay royalties. There are no royalties with OSS. Problem solved. But then again, they can't even find a target. Linux isn't one face, its many, unassociated, different corporations and organizations. Sue this!

Big F'in Blue

[Hanna's Goblin Toys]

Worse yet, though a project may have pedigreed and documented prior art that could easily convince a court to overturn a patent, the cost of such an action is out of reach for most developers -- and many companies.

The flipside to this is that frivolous, non-innovative patents can be attacked without an infringment case - out of the blue - as restraint of trade, provided you have the cash to do so. If someone were to patent something obvious, let's say ordering a list of files by date, anyone can file a restraint of trade suit agianst the patent holder.

Linux's new backer, Big Blue, does have the lawyer army required to attack frivolous patents. The patent war may mean the true end of non-business-backed Linux. It also may mean the doom of companies that are senselessly patenting other people's inventions, and future security for Linux.

Re:So....

[InternationalCow]

I find this a very intriguing question although you might want to phrase it differently: whom to sue if a patent is infringed upon by the Linux kernel? In the cooperative development environment that is Linux, no-one really *owns* the kernel if I understand the GPL correctly. I think that is what has the microsofts of this world so worried - no single entity to attack and subdue. I would be most interested to see what a patent lawyer would come up with. I note that this issue is not mentioned in the article. Anyone?

Simple question

[Anonymous Coward]

If, hypothetically speaking, software XYZ violates a patent, can the users of the program be held liable, or just the program's developers or vendors?

I am pretty sure that copyrights only apply to copying, not use, but does anyone clearly know what the case is with respect to patents?

An observation based on the SCO stupidity...

[pandrijeczko]

Correct me if I am wrong but all the SCO vs Linux stuff has now died away.

The Linux kernel can mount devices formatted under a whole heap of filesystems including System V, UNIXWare and other UNIX-type partitions. Presumably, since SCO holds the UNIX name, SCO also holds the patents to some of these filesystems.

So why didn't SCO carry out legal action based on patent infringement within the Linux kernel for building in that filesystem support?

Isn't this exactly the same type of attack we would expect from Microsoft for FAT, NTFS & Joliet support in the kernel?

Couldn't this be a sign that attacking Linux on patents might therefore simply not be worth it based on the money & time that would be involved?

Any comments?

Re:An observation based on the SCO stupidity...

[13Echo]

"The Open Group" holds the UNIX name, not SCO. SCO owns some UNIX code (System V) but does not own the UNIX name.

SCO could potentially have patents on some of their filesystem stuff, just as Microsoft has patents on certain parts of FAT. FAT's patent was granted in 1996, but there are at least three instances of prior art that have come up that seriously put the patent into question. Before Microsoft's recent attempts, they have never demanded any license over the use of FAT. On the other hand, it's very simple to exclude support for any filesystem from a built Linux kernel (and thus from the source). It's not really something that I think that most people have to worry about. If it becomes an issue, then it can be removed... It's a shame, really, because people will need a common filesystem between Windows and other operating systems. It doesn't matter too much though, because people are making stuff like EXT2/3 support drivers for Windows anyway. It might make Linux migration more difficult for some, but it's just something we'll have to accept.

Re:Maybe less dangerous than we think...

[killjoe]

This is the wrong way to look at it IMHO. We need to plan for what WE are going to do WHEN ms starts suing people.

We need to organize now so we can start boycotting MS products and services, protesting at MS locations, programming our servers to reject traffic from MS, and fighting their FUD with counter FUD (yes FUD we have to fight fire with fire), filing countersuits or what have you.

Plans need to be made, people need to get organized. Would a daily protest in redmond work? Would countersuits by individuals and small companies work? Would the press publish articles that critize MS? Can we really organize a boycott of MS hardware, software and services on a wide scale? Where do we discuss these and other strategies?

I don't think we can presume all these large companies are going to come to the aid of OSS. They might (so far they have) but you have to plan for the eventuality that they won't.

What are YOU going to do when MS starts their war for real? That's the question.

Re:Linus.

[Anonymous Coward]

The Chief Penguin was well aware of the risk (and temptation). He conciousley did not request copyright assignment, so the Linux Kernel is a derivative work of by now hundreds and possibly thousands of contributors. There is no overall copyright holder for the Linux Kernel.

When the FSF was having one of its funny turns and was proposing changes to the GPL that Linus was unhappy with, he stripped the "or at your option any later version" phrase from the licence on his contribution to the Linux kernel, and requested other contributors did the same. In general this request seems to have been followed.

In practical terms the Linux Kernel's copyright is tangled beyond any unscrambling. Licencing it other than via the GPL is untenable.

In many ways the Chief Penguin is very wise....

Why proprietary software is better...

[russotto]

...for the patent issue: Open Source's usual strength is its weakness here. Since anyone can look at the source, anyone -- including hostile IP companies -- can fairly easily find infringing code. With a proprietary OS, the patent-holder first has to reverse-engineer the code to find the infringement. Which isn't so easy.

As for the claim above that no one has sued an end user: nonsense, of course. SCO has, for one. And one company was even successfully sued for a program which did not embody or use a patented process or device, but simply wrote one bit of data that, when inserted into another device, would cause it to execute the patented process in a manner violating the license granted to the manufacturer of the second device.

Re:linux == can of worms == MAD

[Doomdark]

Also, patents are invalid if the idea was published ANYWHERE prior to being patented.

Nope. In US there's the 1 year time period after publishing during which one can file patent applications; obviously as long as applicant published the invention (built something that implements the invention).

There really should be a Slashdot Patent FAQ; this same claim has been refuted a number of times.

For purposes of Open Sourcing things this doesn't have much effect, except that the author can both Open Source something AND patent couple of "inventions" contained therein... :-)

So what are the 283 patents?

[patmc]

Let's say that I know I will never be contributing to the Linux kernel, but I would like to make my own evaluation of this list of potentially infringing patents.

Does OSRM provide the list of 283 patents to those that want to see it?

Pat