Red Hat Makes a GPL-Compatible Patent Deal

Bruce Perens writes "Red Hat has settled patent suits with Firestar Software, Inc., Amphion, and Datatern on a patent covering the Object-Relational Database Model, which those companies asserted was used in the jBoss Hibernate package — not in Red Hat Linux. The settlement is said to protect upstream developers and derivative works of the upstream software, thus protecting the overall Open Source community. Full terms of the settlement and patent licenses are not available at this time."

Reader Koohoolinn adds a link to RedHat's own report of the settlement and adds that the deal "is GPLv2 and even GPLv3-compatible." Koohoolinn also points out commentary on Groklaw that this deal "means that those who claim the GPL isolates itself from standards bodies' IP pledges are wrong. It is possible to come up with language that satisfies the GPL and still acknowledges patents, and this is the proof. That means Microsoft could do it for OOXML if it wanted to. So who is isolating whom?"

Sure this protects anyone other than Red Hat, Inc?

[xxxJonBoyxxx]

Are you sure this protects anyone other than Red Hat, Inc.?

"The settlement is said to protect upstream developers and derivative works of the upstream software, thus protecting the overall Open Source community. Full terms of the settlement and patent licenses are not available at this time."

In other words..."trust us, you little people wouldn't understand the details?"

Re:Not hard

[mapsjanhere]

The issue is that such a license can badly backfire if it deals with software, since software is so nicely reusable.
Using so all so popular car allegory, lets say I make trucks, and I have a patent for a new suspension. Now, since I'm not interested in cars, I license the patent to car makers under GPL. And gosh, some clever truck manufacturer will find a way to claim the design out of a GPLed car to fit into his truck, and point to my GPLed patent.
So, is the truck axle a protected "derivative work" of the GPLed product or a clear violation of the patent? This will probably not be defined in this context other then by the courts.
Until the details of this deal are known I think its too early to call the patent-GPLv3 issue solved.

Re:Legitimacy?

[chunk08]

I just read through part of Red Hat's release, and they seem to claim that they did not admit to infringement. So, it is possible that Firestar and Datatern were afraid of a countersuit as it became clear that they had little standing, and Red Hat decided to settle to quickly get the terms they wanted.

Re:Legitimacy?

[Rhapsody Scarlet]

The GPLv3 itself gives software patents (more than a few of which are ridiculous) a level of legitimacy, by implying that they are a threat to free software and aren't going anywhere any time soon. As much as I'd like to, we can't just ignore patent trolls. They have to be dealt with in some manner, whether that's by challenging the patent or by paying them enough to get acceptable licensing terms.

If you really want to solve the problem once and for all, there's only one way to do it. Get the patent system reformed so it's more difficult to make such patents. The only way that's going to start is with grassroots action, so get to it.

Good for Linux, bad for OSS

[Klync]

While this deal might be good news for RH and anyone up/down stream using the code base, I fear it's counter to the spirit of the F/OSS movement. Yes, I'm actually one of those people who thinks RMS and the GNU Manifesto make sense and raise important considerations. The problem is this: as an intermediate coder (I'm a sysadmin, not a programmer, fwiw), I will now hesitate to look at the source of some "open source" code for ideas and methods. If I do, I might cherry-pick a method which is encumbered, even though its implementation is open. Will the GPL eventually be as useful as MSFT's "shared source" license? I hope not.

Re:Good for Linux, bad for OSS

[setagllib]

At least the GPLv3 protects you in this case. If I remember correctly, by distributing to you a GPLv3 implementation of their patent, a company explicitly protects you from suits based on that patent. GPLv3 has guys like you in mind. That's why it's an even more aggressive little-guy protector than GPLv2.

Re:Not hard

[mapsjanhere]

Sure, in your case that's how the GPL is intended. In my case the truck manufacturer has no intention to license his truck patent under the GPL for use in trucks. He is trying to license it for the use in cars under the GPL. Then someone is taking the car part to break the truck patent.
In software terms, as an engineer I might have a patent on a method to model the aging of engineered structures. Now some academic approaches me to use my program to look at erosion in Triassic fossil fields. Not seeing a conflict or commercial value, I give him a little GPLed program for his application. Now some other engineer modifies that code to compete in my field, claiming GPL protection from my patent since it's "derivative work" of a piece of GPL code.
This is not about me using GPL code and not passing it on, it's about me potentially NOT being able to give the code away fearing for my legitimate patent (which protects my work in a specialized application, not something a patent troll has coughed up to sue).

Re:Sure this protects anyone other than Red Hat, I

[Rolman]

Read carefully:

"The settlement is said to protect upstream developers and derivative works of the upstream software". It means that any implementation of the same code will be protected by this deal, whether is made by Red Hat or not.

That also means it'd be quite the same for you to say about the kernel: "Trust Torvalds, Molnar, Cox and many others, you little people wouldn't understand the kernel internals' details".

How much did this freedom cost?

[theendlessnow]

I now that Red Hat is shooting for GPL high fives... but normally when a company settles, there's money involved. Did this Red Hat settlement result in money being paid to GET the community off the hook? Isn't that patent blood money as well? Just curious.

Re:Sure this protects anyone other than Red Hat, I

[Anonymous Coward]

It seems like this would be simple enough to work around: You have something totally new and you want it covered? Find something Redhat distributes, pick a couple of highly generic classes out of their code, stick them in your code, now it's a derivative and your covered. No?

Re:Not hard

[bobsledbob]

If you choose to release your work under the GPLv3, then you are obligated not to sue the downstream "other engineer" for patent infringement, because his code is a derivative work.

But, this doesn't exclude him from having to follow the requirements of the GPL as well. Meaning, because he is a derivative work, his code is now "infected" and upon distribution, he too will have to provide the source code to his product as well.

In essence, he is bound to release his product as GPLv3, and because of this, he is also not able to sue anyone downstream. He is, in fact, now unable to defend his patent as he's waived the rights to suing by including your original work.

So, call this a competitive stale-mate if you will. You can't sue him, but he can't release something proprietary based on your original work. This is the way it's supposed to work.