Does interoperability violate the GPL?

Does interoperability violate the GPL?

Summary: If you want to bring suit based on an interpretation of the GPL at odds with the understanding of the Linux Foundation go ahead. But so far no one has taken them up on the offer.


VMWare boxes from its logoI got an e-mail this morning, tickling me to look into the idea that VMWare is violating the GPL.

This idea has been around for some time and Big Money Matt has covered it beautifully.

(Matt Asay's writing is first-rate and his sources top-notch. If he ever decided to become a full-time reporter I'd hire him in a New York minute.)

Apparently VMWare loads a module of Linux in order to provide its function, which is virtualization.

In a response to one of Matt's pieces Bruce Perens of the Software Freedom Law Center wrote in to say that nVidia does the same thing, but that the Linux Foundation has done nothing to pursue the matter in court.

"This legal theory will have to be decided with another GPL kernel, not Linux," he concluded.

The reasonable follow-on question becomes, why? And the answer I return with is interoperability.

One of the key Internet values is that things should work together. By following the Internet Protocol while online Linux and Windows machines do just that.

But the concept of interoperability goes beyond this, and it can run into a brick wall with absolutists on both sides. If you can't touch the code you can't interoperate. You can't virtualize.

Long ago, Linus Torvalds took a practical approach to this problem. The Linux kernel has never been subject to FOSS absolutism, the requirement that in order to use it you have to give everything you do back to the community.

As to the questions raised by VMWare, it would seem, Torvalds brings up the loadable module exception. It's a gray area, he writes. He doesn't want to get into it. Better to write code.

This is part of what I call the Fourth Freedom, the freedom to get back what's added to the code, the tip of the spear separating FOSS from open source.

FOSS considers this to be bedrock, open source clay. Open source, as a business model, is not rigid on this point. FOSS, especially through the GPL V. 3, attempts to be rigid on this point. Linux is still under the GPL V. 2.

VMWare has been, since August, a member in good standing of the Linux Foundation. The question my PR friend was asking, in other words, has been asked and answered.

If you want to bring suit based on an interpretation of the GPL at odds with the understanding of the Linux Foundation go ahead. But so far no one has taken them up on the offer.

Should they?

Topics: VMware, Linux, Open Source, Operating Systems, Software

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Open source, development method or moral philosophy?

    I'm not certain that efforts to separate the two will always be successful. Many of the people who are supplying the work that allows reduced staff expenses are probably doing so because they believe to some degree in open source principles.

    What happens when they realize that the work they do to reduce the monetary value of software is actually being exploited to provide substantial money to distributors?

    Interoperability is unlikely to be the issue that makes the issue unavoidable. Cooperation like that between Microsoft and Novell appears useful to both sides. Novell because the ubiquity of Microsoft ceases to be a hindrance and instead becomes a sales point. While Microsoft receives some PR benefit and increased opportunities to displace Linux in the future.

    No, the trigger incident would be something like news reports of Red Hat executives receiving salaries of $10's of millions annually for selling what devotees have given them free. And without the slightest effect on proprietary software, which continues to employ as many people as before.

    That could generate outrage expressing itself in a Court battle.
    Anton Philidor
    • Software is NOT religion!

      You sound like a programmer getting paid good money to write software is a bad thing! Why is it important that "proprietary software, which continues to employ as many people as before".

      Proprietary software still profitable and good for programmers, what's wrong with that? Capitalism still works, what's wrong with that?
    • Don't Mix the Two

      Open source is a development method, free software is a belief system.
    • Nothing wrong with

      working for free if you do not mind
      • Exactlly

        [i]Nothing wrong with working for free if you do not mind[/i]

        Its funny that many people will point out that this developer or that developer is a fool for "working for free",yet wouldn't think twice about someone donating thier time to work for free on something like a Habitat for Humanity house, or voluntering their time to clean up a park with an orginization.

        I agree, if it makes you happy and it doesn't bother you to "work for free" then what's the problem?
        John Zern
  • I'm not with SFLC

    Dana, I'm not with SFLC!!! Please fix this story right away. I am an expert on Open Source licensing and creator of the Open Source Definition.
    • Corection made -- sorry about the brain cramp

      Thanks for staying on top of me. Glad to be on your
      RSS feed. Made the correction.

      But you're still important, and your opinion is still
      important in this area. You told Matt that it won't be
      Linux that will try to make this legal point.

      Whom might it be?
      • Since Linux is basically just a trademark

        couldn't he just take the entire code base and replace all instances of the trademark "Linux" with "Brucix" and have his new kernel to proceed with?

        Or would he have to suck VMWare (or nVidia) into providing support for it before he has a case? Or would he have to contribute something to the code base (either fork) to stake a claim? Or is the trademark completely irrelevant in this discussion?
        Michael Kelly
        • The trademark has other purposes

          You may recall that Linus put together a trademark and
          a foundation for it. But its aim is mainly the
          policing of use of the term. It's not the legal
          framework under which the operating system runs.
  • Of course not

    Since the GPL precludes the use of patents to restrict copying or usage of the software, people are free to write their own implementations of whatever algorithms the software employs. Proprietary developers are free to study the code all they want, as long as they don't copy it into proprietary software (but they can still write interoperable code).
    John L. Ries
    • It's not about interoperable code.

      It's about actually using the code in your product. That is what the article is about. Modules being used with proprietary modules. Not the interoperability of the modules. So in effect they have copied the open source code into their proprietary code.
      • How else do you do a virtualizer....

        ...except by using some of the operating system code
        in what you're virtualizing. Microsoft has made this
        point over and over when people tried virtualizers and
        emulators on Windows. Samba. Wine. They eventually
        gave in, under antitrust pressure, and allowed for

        Linux has indicated they won't get in the way, even i
        the case of VMware, which as you indicate has some of
        its code in there.

        So who will? Will anyone? Is it right if no one will?
        Is this a legal point worth making? Or is this a fight
        worth avoiding?
        • Protocols, Interfaces, and Code

          Maybe I'm jumping in on the wrong issue, but is there a problem with interoperability among the free operating systems and applications? I'm small potatoes but it seems like my home FreeBSD server and my Linux systems get along hunky dory.

          Or have I made a dumb observation because one of the tenets of open source development is to not rewrite what already works? That seems to me to mean that interoperability is in open source's DNA.

          As to virtualizers, if we know the public interface do we need to know the implementation details? I don't think so. The only time one requires implementation details is when the behavior is different then documented or documentation is sparse. In the first case, I think the implementation is broken, though then one gets into the thicket of de facto standards.

          Sun, by the way, has always made the opposite point to Microsoft's. If one implements the publicly documented interfaces and protocols, then Sun's stuff talks with you as an equal (for example, NFS).
  • So where is the controversy?

    The binary blob of VMware would have to contain code copied from Linux (or other GPL code) for a GPL violation to occur. I didn't see anywhere in the article (or links) that unambiguously stated that the binary blob contained Linux code, only that it interfaced/interacted with the Linux code. Did I miss something?
    Even if the binary blob requires a customized Linux kernel to function properly, as long as the source for the customized kernel is made available under the GPL, the blob itself can be proprietary. No GPL violation.
    Something designed to work with something else is not a derivative. To be a derivative the new item has to include significant pieces of the original item.
    A cheap solution that doesn't work is neither,
    Say What?
    • Sorry about the formating...

      didn't use the correct tag delimiters. Oh why can't everyone use the same conventions?

      Say What?