|
|
Log in / Subscribe / Register

Patents and the open-source community

By Nathan Willis
June 2, 2016

OSCON

At OSCON 2016 in Austin, a panel of invited experts debated the always-thorny subject of how open-source software projects deal with patents. The panel was packed, featuring representatives from the free-software world, commerce, and the legal community, so there was scarcely enough time to move through the prepared topics in the time allotted, much less to take questions from the audience. But the discussion was able to highlight a number of current issues, including patent abolition, implicit patent licenses, and where the open-source community should focus its efforts to improve matters.

Defining the problem

Jim Jagielski from the Apache Software Foundation (ASF) served as moderator. The panelists were Bradley Kuhn of the Software Freedom [Jim Jagielski] Conservancy (SFC), Heather Meeker from the law firm of O'Melveny & Myers, Rabin Bhattacharya from Capital One, Keith Bergelt of the Open Invention Network (OIN), Mishi Choudhary of the Software Freedom Law Center (SFLC) and SFLC India, and Eben Moglen of the SFLC. After introducing everyone, Jagielski posed the first question: are software patents inherently evil, or are they just implemented wrongly in the legal system?

Kuhn replied first, saying that he has a long history of saying "eliminate them all." In fact, he said, "I'm probably the only patent abolitionist here." Since only Congressional legislation could abolish software patents in the US, he said, the free-software community uses other means to mitigate them—such as copyright. Cloudhary also called herself a patent abolitionist, though, noting that SFLC has worked through the courts (such as filing amicus briefs) to combat software patents. In India, those efforts have been a success: India's Patent Act does not cover software. Yet there is still resistance, with companies in India acting as though software patents are allowed (such as by offering to file software patent applications in India on a client's behalf). Consequently, she said, the patent office has had to release guidelines on several occasions clarifying that software patents are not permitted.

Bergelt replied next, saying that the patent system has means for self-correction, which the open-source community must use—like OIN's Defensive Publications program. But the changes the community wants must be implemented on top of everything that came before, he said. "We're writing on a dirty slate," he said, and there is a long way to go. But he said that the community does not want to get in the way of actual innovation, which is what justifies OIN's "nonaggression" approach. Moglen then identified himself as not just a software-patent abolitionist, but as someone opposed to "government-granted monopolies" of any kind. "I think we could do without all patents," he said, "so I'm more abolitionist than anyone else." As a lawyer, he said, his job is to do what he can for his clients and, where software patents are concerned, that means he is preoccupied with defensive measures.

Bhattacharya said he was "somewhere in the middle." On one hand, there are clearly software patents that are bad. But, on the other, he said, recent changes have made it much more difficult to get a new software patent issued. Research indicates that there is a correlation between a start-up company getting software patents granted and that company getting venture capital. So the industry believes that there is a financial upside. And, he said, he finds many philosophical objections to software patents unconvincing. "Think about how much hardware is emulated in software these days. The inventions that people make as hardware tweaks are patentable, but when the same thing is implemented in software it isn't. Why?"

In contrast, Meeker reported that all of her clients, which include "the biggest technology companies in the world," tell her that they hate software patents. "Who's being well-served by software patents? I have one idea: the hotels and restaurants in Marshall, Texas. They're getting a good deal." To everyone else, she said, software patents are a menace they are afraid of but are afraid to let go of, too. "I'm disappointed that we don't hold the government more accountable to fix the problem," she said.

Strategy and tactics

The next subject Jagielski asked about was how open-source software licenses ought to address patents. Other than a select few (such as the Apache license), most open-source licenses ignore patents, he said. But there is a theory circulating that all software licenses include an implicit grant to use any patents on the licensed software. Jagielski asked the panelists what they made of the theory and whether licenses should be updated to reflect it.

Moglen replied first. Under US patent law, he said, the theory that Jagielski referred to is known as "exhaustion." It holds that a seller cannot sue its own customers for patents in the product that they have sold: the seller's right to exclusive control over the patent is exhausted once the article in question is sold. But in many other jurisdictions, such as the UK, no such doctrine exists. And that is why, he said, the GPLv3 includes its "patent non-aggression" clause. "It's a single-license version of what the OIN pool does. It says that sharing is the rule and sharing is compulsory." Under permissive licenses like the X, MIT, and BSD licenses, the user is at the mercy of whatever the local patent system says. "That varies and it might change. The Kazakh or Chinese patent system may come after you some day."

The Apache license, he added, relies on "defensive suspension" of patent grants whenever someone initiates patent-infringement litigation. That clause is a "poison pill." Bergelt then commented that he believes there are some open-source project communities (such as the Bitcoin and Hadoop communities) that have grown large enough that they may be interested in a "complementary" solution to the Apache license's defensive-suspension approach, although he did not speculate on what such a solution would look like.

Meeker added that if one reads the case law about implied licensing, there is not much to stand on. First of all, she said, the terms "exhaustion," "estoppel," and "implicit license" are all used interchangeably. While the three terms supposedly refer to separate doctrines that would prevent a patent holder from suing a customer for infringement when there is no explicit license, case law does not appear to clearly define or differentiate between them. Consequently, she said, "no lawyer will write you an opinion taking a stance" on implicit patent licenses.

[Patent panel]

Jagielski then asked the panel what else the open-source community should be doing to take action about software patents. Choudhary said she tells clients to join OIN, because "you get so much for free." In other jurisdictions, she added, the software community has started to see the benefit of OIN-style patent pools and defensive publications. "You have to do those things in parallel with working to abolish software patents," she said.

Moglen advised separating the past from the future when discussing tactics. For the future, he said, "we should prevent people from getting patents." But, in the past, the problem is old patents "rising up and smiting" projects, harming innovation. The work that Choudhary does trying to abolish patents, he said, covered how to protect people in the future—and he fully expects SFLC to continue that work, arguing in front of the Supreme Court about the "design patent" case between Apple and Samsung. Furthermore, he said, "open-source software is an immense repository of prior art. Free software can help by educating people on all that we've invented and that you therefore cannot reinvent and patent."

Kuhn urged free-software developers who work at for-profit companies to refuse to file for patents on their work, even if that means losing bonuses and, possibly, promotions. "It might be good for your career to get patents, but taking a stand is good for all of software freedom." Bergelt suggested that developers look at Twitter's Innovator's Patent Agreement (IPA), which he called a "middle ground" approach. The IPA is a pledge that Twitter will only use patents on employee inventions defensively, unless they get the employee's consent. Meeker recommended the License On Transfer (LOT) Network, which seeks to prevent patent trolls from acquiring patents from companies in desperate financial situations. "Cooperative approaches are great," she said, "but they do not affect patent trolls. My clients are way more afraid of trolls than they are of their competitors."

Choudhary then added that interested developers could join the work going on in India. "It's the only jurisdiction where we're the incumbent, but we still face lots of pressure." She also advocated taking a stand on international trade agreements, which she noted are usually written in secret, a fact that companies try to take advantage of. Bhattacharya added that recent changes at the US Patent and Trademark Office (USPTO) allow the public to submit prior-art examples on patent applications. "If you see one and know that it shouldn't be granted, you can anonymously submit up to three pieces of prior art." Kuhn replied that he sees one downside to Bhattacharya's suggestion: by fighting some patent applications, any that get granted will be perceived as stronger. Kuhn remembers the RSA patent, he said, which many developers might be too young to recall. "It set the progress of encryption back twenty years, and something like that can happen again."

Fear and participation

Next, Jagielski asked panelists about patent-owning corporations' fears of open source. "People are using open-source software all over the place; some companies worry that using it means giving away the 'keys to the kingdom.' Is that fear real?" While those in the open-source community may find such fears hard to imagine, Meeker reported that they are still quite common. "I should have a standard line item for 'talking clients off the ledge over concern about using open-source software,'" she said. Nevertheless, she said, once she sits down with clients and looks at the licenses involved, they get over their fears. "I have never had a client throw up their hands and say 'no, we can't do this, because we have patents.'" Bhattacharya added that there is "a lot of daylight" between the patent policies most companies have and the norms of the open-source community, but that understanding the rights granted in the software licenses and knowing what code you are contributing makes all the difference.

Kuhn said that, in the long term, "it is very hard to be an authentic open-source software participant and not contribute." Free-software communities are rather self-organizing, he said, so if a company wants to lead a project, it will have to contribute code. Moglen responded that corporate fears about open source and patents depend on the culture. In the US, companies rely on lawyers to "talk them off the ledge." But China is completely different; the Communist Party's view is that all patents are government property, he said, so companies there operate without the fear of open source that one sees in the US. Indian companies, he added, "haven't yet grasped how good their situation is" regarding software patents.

As time ran out in the session, Jagielski posed one final question to the panelists in reference to the then-in-progress Oracle-v-Google lawsuit—in which Oracle asserted copyright protection over Java's API. "We think we understand patent law," he said, "but if other things we think we understand, like copyright law, are as uncertain and malleable as they sound today, what does that change for open source?" Moglen was the only panelist to respond, saying "I think we don't know patent law." The "design patent" lawsuit between Apple and Samsung is an example, he said. "I'm very glad that our friends at Google are going to find out what 'fair use' is. It'd be good to know that for design patents, because that clarity will improve enormously our ability to innovate in our world."

With that, the session broke up, although the discussion about how free and open-source software projects work with the patent system will, no doubt, continue for some time to come.

Index entries for this article
ConferenceOSCON/2016


to post comments

Patents and the open-source community

Posted Jun 2, 2016 23:46 UTC (Thu) by smoogen (subscriber, #97) [Link] (8 responses)

Back in the early 90's when software patents were becoming big, and it looked like the courts would uphold them, I asked RMS why we didn't have a GNU Patent License that hacked Patents like they did Copyright. You could use the patent all you wanted but only if the code you implemented was under a GPL license. The argument that I got back was that they should just be abolished which was true but not likely.

The bigger problem with my idea was that patents are not 'freely' given like copyright. It takes a lot of work and money to get a patent filed so there needed to be a backing source that was willing to 'give' up money for something that it wasn't getting a monetary return on. That never appeared so a GNU Patent License was never going to be possible.

Patents and the open-source community

Posted Jun 4, 2016 20:02 UTC (Sat) by jcm (subscriber, #18262) [Link] (7 responses)

Well, it did happen in the GPLv3. My own personal opinion is that the GPLv3 was incredibly damaging as a result. It didn't actually kill the patent system so much as force all of the large corporations (especially those new to FLOSS) dealing with Open Source to impose extremely draconian (yet understandable) measures to prevent the undermining of their patents. What should be happening is the reform of patent legislation on the books, not attempts to hack the system with licenses that simply serve to push good folks away. This is one of the reasons I killed my FSF membership a while back. They no longer represent the kinds of views I hold.

Patents and the open-source community

Posted Jun 4, 2016 23:52 UTC (Sat) by smoogen (subscriber, #97) [Link]

I was looking at them to be separate documents where you put code that followed the patent into the GPL universe versus tieing them together in one license. [It was more of a 'well if they are going to make it so I can't code certain things because its patented.. then maybe they shouldn't be able to code certain things without paying me the way I want.] However I was also much much younger and more naive about how patents, licensing and commercial companies worked.

Patents and the open-source community

Posted Jun 5, 2016 11:49 UTC (Sun) by pboddie (guest, #50784) [Link]

Are "good folks" ones that offer works to others but who reserve the right to prevent those others from taking advantage of the promises made to them when receiving those works? Because that's what patents effectively do to copyright licences, and given the presence of patent-related clauses in other licences, it isn't just the FSF who has noticed.

There's also a train of thought that GPLv2 actually compels patent-holders to license their patents to recipients, too. So when companies like Nokia who offered patent-related promises around Linux did so, it probably wasn't corporate generosity on display: the highly-specific version information accompanying such promises should have given the game away there.

Patents and the open-source community

Posted Jun 5, 2016 16:15 UTC (Sun) by flussence (guest, #85566) [Link] (2 responses)

I can understand people saying the GPL3 was damaging, and it definitely didn't cause more openness as was the intent. But on the other hand, it did a very good job of weeding out the corporations that were just exploiting GPL2'ed code as a free lunch and never reciprocating.

(And on the subject of exploitative corporations, I could've sworn KHTML was LGPLed... how did that get diluted to BSD-only in Webkit and derivatives?)

Patents and the open-source community

Posted Jun 5, 2016 17:19 UTC (Sun) by rahulsundaram (subscriber, #21946) [Link] (1 responses)

> And on the subject of exploitative corporations, I could've sworn KHTML was >LGPLed... how did that get diluted to BSD-only in Webkit and derivatives?

What makes you think so?

https://webkit.org/licensing-webkit/

Patents and the open-source community

Posted Jun 5, 2016 21:18 UTC (Sun) by Cyberax (✭ supporter ✭, #52523) [Link]

Currently most of the code in WebKit/Blink is BSD-licensed. All the new contributions are also BSD and it's entirely possible to build it without a single line of LGPL code being present.

Now, there's still a question of whether it will be considered a derived work of the original KHTML and the answer is probably: "Yes, for now".

Patents and the open-source community

Posted Jun 6, 2016 19:32 UTC (Mon) by atai (subscriber, #10977) [Link] (1 responses)

That was kind of strange because the FSF was not the leader in putting patent protection clauses into FOSS licenses. Mozilla or Apache seemed to be the first one mentioning patents in copyright licenses. You should blame these first, right?

Patents and the open-source community

Posted Jun 6, 2016 21:29 UTC (Mon) by andresfreund (subscriber, #69562) [Link]

I don't think those are comparable, due to the commonly used "or later" clause in GPL licensing allowing a lot of existing projects to migrate to GPLv3 effortlessly.

Patents and the open-source community

Posted Jun 13, 2016 17:00 UTC (Mon) by murukesh (subscriber, #97031) [Link]

I'm very interested in what the situation with software patents in India is, and why they think it's so good. IP issues rarely occupy news space in India, and rarely do foreign sites deal with Indian issues. Any clues?


Copyright © 2016, Eklektix, Inc.
This article may be redistributed under the terms of the Creative Commons CC BY-SA 4.0 license
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds