Re and Expanding on Jaron Lanier on Closed Source Software

by on January 3, 2008 · 16 comments

(This is in response to a thoughtful post by Adam, who beat me to the punch, and to a controversial recent article (“Long Live Closed-Source Software!“) by the free-thinking Jaron Lanier that Adam discusses.)

No one needs to say “Long live open-source software,” because it is what it is and isn’t going anywhere. Think of it as the ground beneath our feet.

As Lanier explains, closed source is the font of nearly all paradigmatic innovation–the great revolutionary leaps. Open source contributes iterative innovation, such as the best kernel scheduler for variable workloads–this is a problem it is possible to work out slowly, with small changes over a period of years. It is also a problem that doesn’t matter at all to most users–good enough is good enough, though better is, of course, better.

Where the two forms of development come together most interestingly is the use of open source as a stepping stone for closed-source radical innovation. Asus, for example, didn’t have to step forward and create its own OS from the ground up for its EeePC. Even though the thing sports an interface unlike those in most Linux distributions, the underlying guts are the same. Would something like the EeePC even be possible without open source? Could a manufacturer afford to undertake the great expense, and gamble, of working out an OS for itself? Free software lets businesses take chances on projects that would otherwise be too expensive to devise and products that would otherwise be too expensive to market.


Apple’s OS X is another example, built on the same core (well, nearly) as the open-source FreeBSD OS. Running FreeBSD and OS X are completely different experiences, but again, the guts are the same. And if Apple (or NeXT before it) had had to develop the whole system itself, it probably never would have been built–at least, not in anything like the form we see it today. (Would some of us be running, god help us, Copland?)

My point, which I think jibes with Lanier’s, is that open source establishes a technological baseline that facilitates innovation in more interesting areas. Very few of those innovations, though, are actually done in open source, at least not until they’ve been created in the closed-source world first.

(Even web browsers are an example of this. Today’s Firefox would be instantly recognizable to, and usable by, someone familiar only with Netscape Navigator 3. Most of the changes have been iterative and under the hood. The biggest changes in the browser experience in recent years have been Flash and XMLHttpRequest, both borne of closed-source development, but it was the web’s open standards that made Flash and AJAX possible.)

For those of us who haven’t taken sides in what Adam astutely calls a “senseless techno-philosophical holy war” but instead recognize that both forms of development have their own advantages, it is important that each side have the flexibility to do what it does best. I wonder, then, what Lanier think of the “viral” GPL, which no doubt inhibits or slows some software revolutions that could be built atop software that is open source but covered by the GPL but are not because of the redistribution requirement. On the other hand, there is great value to iterative development from requiring that most changes to source code be distributed freely. Projects like FreeBSD (free source code, but no requirement that changes be shared), however, suggest that requiring changes to be made public may not always be necessary in iterative development.

In a dispassionate discussion (not debate), it might be a good resolution to conclude that sharing is a community norm and radical innovation an economic, market good–few business probably expect to make bundles off products that are only slightly better than everyone else’s. It’s an empirical question as to whether that would favor more flexible licenses, like the BSD and MIT licenses (which permit code to be used in closed-source development), but I think it’s likely.

Sadly, there is too often nothing but passion and absolutist rhetoric in the debates over (of all things!) source code licenses.

That’s why Lanier’s column is so welcome. With his background, credentials, and (yes) the dreadlocks, Lanier can command the attention of all but the most devout GPL adherents and perhaps goad them to think about their religion and reflexive hostility to close-source development. It’s not that open source doesn’t have its benefits–it does, and they’re big ones–but just the more modest point that it isn’t best for certain and important kinds of development. Why that is so controversial in some circles is beyond me.

  • http://enigmafoundry.wordpress.com/ enigma_foundry

    As Lanier explains, closed source is the font of nearly all paradigmatic innovation—the great revolutionary leaps.

    If this were true than why were the first web browsers open source? The browser nexus (originally called worldwideweb) developed by Tim Berners-Lee is open source. (Although not GPL)

    I’d also point to the filesystem ZFS, as state-of the art file system.

    Or how about blackbox, or Celestia?

    One thing I note in Lanier’s article is anecdote–he does not supply any metrics by which we could verify his hunches.

    Of course some points he makes are obviously true–that for example closed and open exist iteratively, and some projects are better being closed for a certain time.

    But he hasn’t demonstrated that there is anything open source can’t do.

  • http://enigmafoundry.wordpress.com eee_eff

    As Lanier explains, closed source is the font of nearly all paradigmatic innovation—the great revolutionary leaps.

    If this were true than why were the first web browsers open source? The browser nexus (originally called worldwideweb) developed by Tim Berners-Lee is open source. (Although not GPL)

    I’d also point to the filesystem ZFS, as state-of the art file system.

    Or how about blackbox, or Celestia?

    One thing I note in Lanier’s article is anecdote–he does not supply any metrics by which we could verify his hunches.

    Of course some points he makes are obviously true–that for example closed and open exist iteratively, and some projects are better being closed for a certain time.

    But he hasn’t demonstrated that there is anything open source can’t do.

  • http://bennett.com/blog Richard Bennett

    As luck would have it, I’ve been checking-out the EeePC for my archaeologist wife. It’s essentially a conventional laptop with two exceptions: very small size & weight, and a flash drive instead of a hard drive. It comes with Linux simply because it’s cheaper than Windows, and many EeePC users on message boards explain that the first thing they did when they got it home was install Windows. Given that Firefox, Thunderbird, and Open Office provide the necessary functionality, I don’t plan on replacing Linux with Windows when we get ours, but that will depend on the primary user’s preferences. The EeePC still has some issues with the tiny 7 inch screen size, which could be made wider, battery life, and limited storage (4 Gigs shipping now, 8 gigs promised, and it really should have more like 16.)

    It’s hard to see any particular connection between this machine and the Open Source religion one way or another.

    On Lanier’s actual point, I’d have to agree with him that Open Source is like offshore manufacturing, good at making things cheaper, not good at making things better. I’ve yet to see a single example of an innovative new technology coming from either offshore manufacturing, prison labor, or Open Source. Firefox is nice, but it’s just a browser, and we’ve had browsers for nearly twenty years now. Samba is great, but it’s just an implementation of a 25-year-old proprietary protocol, and the list goes on.

    Genuine advances in technology tend to the the work of the lone genius and an army of drones, and in Open Source you only have the drones. Drones don’t do much that’s revolutionary.

    The Lisp Machine narrative is interesting to me because I once worked for the company that did the most with it, Texas Instruments. TI’s Explorer group ultimately created the MicroExplorer, an add-on board for the Mac II that put Lisp in the hands of the great unwashed. It failed in the marketplace because Lisp is a PITA, not for want of effort.

    Programmers familiar with Stallman’s Emacs, the last surviving example of Lisp, are well aware that Stallman is a kook of the first order. His flagship product is just now getting support for anti-aliased fonts, something that Windows had fifteen years ago. What a joke.

  • http://bennett.com/blog Richard Bennett

    As luck would have it, I’ve been checking-out the EeePC for my archaeologist wife. It’s essentially a conventional laptop with two exceptions: very small size & weight, and a flash drive instead of a hard drive. It comes with Linux simply because it’s cheaper than Windows, and many EeePC users on message boards explain that the first thing they did when they got it home was install Windows. Given that Firefox, Thunderbird, and Open Office provide the necessary functionality, I don’t plan on replacing Linux with Windows when we get ours, but that will depend on the primary user’s preferences. The EeePC still has some issues with the tiny 7 inch screen size, which could be made wider, battery life, and limited storage (4 Gigs shipping now, 8 gigs promised, and it really should have more like 16.)

    It’s hard to see any particular connection between this machine and the Open Source religion one way or another.

    On Lanier’s actual point, I’d have to agree with him that Open Source is like offshore manufacturing, good at making things cheaper, not good at making things better. I’ve yet to see a single example of an innovative new technology coming from either offshore manufacturing, prison labor, or Open Source. Firefox is nice, but it’s just a browser, and we’ve had browsers for nearly twenty years now. Samba is great, but it’s just an implementation of a 25-year-old proprietary protocol, and the list goes on.

    Genuine advances in technology tend to the the work of the lone genius and an army of drones, and in Open Source you only have the drones. Drones don’t do much that’s revolutionary.

    The Lisp Machine narrative is interesting to me because I once worked for the company that did the most with it, Texas Instruments. TI’s Explorer group ultimately created the MicroExplorer, an add-on board for the Mac II that put Lisp in the hands of the great unwashed. It failed in the marketplace because Lisp is a PITA, not for want of effort.

    Programmers familiar with Stallman’s Emacs, the last surviving example of Lisp, are well aware that Stallman is a kook of the first order. His flagship product is just now getting support for anti-aliased fonts, something that Windows had fifteen years ago. What a joke.

  • http://linuxworld.com/community/ Don Marti

    The GPL’s impact on proprietary software innovation is overrated.

    Every copy of Mac OS X includes Samba.

  • http://linuxworld.com/community/ Don Marti

    The GPL’s impact on proprietary software innovation is overrated.

    Every copy of Mac OS X includes Samba.

  • http://metapundit.net/sections/blog metapundit

    >but just the more modest point that it isn’t best for certain and important kinds of development. Why that is so controversial in some circles is beyond me.

    Maybe some of the confusion is in not grokking the different outlooks of Open Source and Free Software people.

    Now arguably all software professionals should prefer Open Source software for a variety of reasons (although the argument that programmers should oppose Open Source software since it drives the cost of software and hence our own profitability towards zero is interesting if futile). Open Source proponents (technically exemplified by Linus Torvalds, philosophically by Eric Raymond) further argue that a certain style of development (the bazaar) is inherently better at producing certain kinds of software. Interestingly the arguments are primarily about bugs (given enough eyeballs all bugs are shallow) and security, not about innovation.

    Free Software proponents (Stallman) on the other hand believe philosophically that certain actions are moral and immoral. Preventing users from modifying their software is immoral. It doesn’t matter whether the software resulting from open or closed processes has a higher quality – it’s about a moral argument (one I don’t place much credence in, fwiw, but a moral one none-the-less). Saying to Free Software types “you shouldn’t oppose closed source development because you’ll hinder innovation” is to fail to understand their argument.

    I myself feel a bit conflicted about the argument. Let me make an analogy that will place you on the horns of the same dilemma: as a libertarian you probably feel that as a moral rule people should not be prevented from entering into any contract (with perhaps some Rothbardian exceptions around voluntary slavery). Undoubtedly you feel this is likely to lead to a “better” society as a whole but your faith in this proposition is not based upon obtaining maximally good results in every contingency (fascism might yield a more prosperous society given certain conditions but this would not commend it to you).

    Given this libertarian position, DMCA backed DRM seems abundantly defensible. One party enters into a contract with another party not to circumvent the digital restrictions on bits of software and media as a condition of obtaining them. If the first party does not like these conditions then certainly it is free not to make the purchase and enter into the contract. Et voila! We all have Liberty.

    And yet. Read Stallman’s dystopian tale The Right To Read which strikes me as a plausible preview of an increasingly digital world in which DRM applies to everything (no really, go read it! Or if you must, you could read the DiveIntoMark Cliff Notes application to today.) Through a combination of technology and freedom of contract, freely exercised by everyone, we may yet get a totalitarian society!

    Now if you’re morally committed to libertarianism it seems to me that this possibly future is no argument against freedom of contract – any more than innovation in closed source shops is an argument against Software Libre. Of course the possible implications, the ends of our ideologies ought to make us think about our moral commitment to a set of means. I feel uneasy about uncompromising purity of moral vision which is immune to such qualms (libertarianism no less than Stallman’s software-has-rights-too theories).

  • http://metapundit.net/sections/blog metapundit

    >but just the more modest point that it isn’t best for certain and important kinds of development. Why that is so controversial in some circles is beyond me.

    Maybe some of the confusion is in not grokking the different outlooks of Open Source and Free Software people.

    Now arguably all software professionals should prefer Open Source software for a variety of reasons (although the argument that programmers should oppose Open Source software since it drives the cost of software and hence our own profitability towards zero is interesting if futile). Open Source proponents (technically exemplified by Linus Torvalds, philosophically by Eric Raymond) further argue that a certain style of development (the bazaar) is inherently better at producing certain kinds of software. Interestingly the arguments are primarily about bugs (given enough eyeballs all bugs are shallow) and security, not about innovation.

    Free Software proponents (Stallman) on the other hand believe philosophically that certain actions are moral and immoral. Preventing users from modifying their software is immoral. It doesn’t matter whether the software resulting from open or closed processes has a higher quality – it’s about a moral argument (one I don’t place much credence in, fwiw, but a moral one none-the-less). Saying to Free Software types “you shouldn’t oppose closed source development because you’ll hinder innovation” is to fail to understand their argument.

    I myself feel a bit conflicted about the argument. Let me make an analogy that will place you on the horns of the same dilemma: as a libertarian you probably feel that as a moral rule people should not be prevented from entering into any contract (with perhaps some Rothbardian exceptions around voluntary slavery). Undoubtedly you feel this is likely to lead to a “better” society as a whole but your faith in this proposition is not based upon obtaining maximally good results in every contingency (fascism might yield a more prosperous society given certain conditions but this would not commend it to you).

    Given this libertarian position, DMCA backed DRM seems abundantly defensible. One party enters into a contract with another party not to circumvent the digital restrictions on bits of software and media as a condition of obtaining them. If the first party does not like these conditions then certainly it is free not to make the purchase and enter into the contract. Et voila! We all have Liberty.

    And yet. Read Stallman’s dystopian tale The Right To Read which strikes me as a plausible preview of an increasingly digital world in which DRM applies to everything (no really, go read it! Or if you must, you could read the DiveIntoMark Cliff Notes application to today.) Through a combination of technology and freedom of contract, freely exercised by everyone, we may yet get a totalitarian society!

    Now if you’re morally committed to libertarianism it seems to me that this possibly future is no argument against freedom of contract – any more than innovation in closed source shops is an argument against Software Libre. Of course the possible implications, the ends of our ideologies ought to make us think about our moral commitment to a set of means. I feel uneasy about uncompromising purity of moral vision which is immune to such qualms (libertarianism no less than Stallman’s software-has-rights-too theories).

  • http://tim.geekheim.de Tim Pritlove

    You wrote: “Apple’s OS X is another example, built on the same core (well, nearly) as the open-source FreeBSD OS. ” Well, this is not true, not even a bit. Mac OS X ist built around Mach and an “BSD” Layer, that is UNIX compatible but anything but FreeBSD.

    What is true is that parts of the FreeBSD userland is in use but this does not count as “core” as it is not part of the “core” (and not used at all by Mac applications).

  • http://tim.geekheim.de Tim Pritlove

    You wrote: “Apple’s OS X is another example, built on the same core (well, nearly) as the open-source FreeBSD OS. ” Well, this is not true, not even a bit. Mac OS X ist built around Mach and an “BSD” Layer, that is UNIX compatible but anything but FreeBSD.

    What is true is that parts of the FreeBSD userland is in use but this does not count as “core” as it is not part of the “core” (and not used at all by Mac applications).

  • http://enigmafoundry.wordpress.com/ enigma_foundry

    Richard Bennett said above:

    Genuine advances in technology tend to the the work of the lone genius and an army of drones, and in Open Source you only have the drones. Drones don’t do much that’s revolutionary.

    Richard, then the following folks are drones?

    Linus Torvalds, Richard Stallman, Eric Raymond, Andrew Tridgell, Sir Timothy John Berners-Lee

  • http://enigmafoundry.wordpress.com eee_eff

    Richard Bennett said above:

    Genuine advances in technology tend to the the work of the lone genius and an army of drones, and in Open Source you only have the drones. Drones don’t do much that’s revolutionary.

    Richard, then the following folks are drones?

    Linus Torvalds, Richard Stallman, Eric Raymond, Andrew Tridgell, Sir Timothy John Berners-Lee

  • http://bennett.com/blog Richard Bennett

    Your two biggest names, enigma, follow my model: TBL created his little web and his browser on his own, and Linus created his little OS on his own. Afterwards, they gave their stuff away, but the creative act itself was more or less solitary.

    But neither of these count as “radical innovation” in my book. TBL essentially delivered a discount version of Xanadu, and Linux is a nothing more than a Unix clone. So where’s the paradigm-shifting breakthrough?

    Stallman has overseen the re-writing of some code-munging tools, compilers and utilities, nothing new there. His editor is, as I’ve already said, a joke.

    These examples simply prove my point (echoing Lanier) that Open Source makes things cheaper but it doesn’t drive radical innovation.

  • http://bennett.com/blog Richard Bennett

    Your two biggest names, enigma, follow my model: TBL created his little web and his browser on his own, and Linus created his little OS on his own. Afterwards, they gave their stuff away, but the creative act itself was more or less solitary.

    But neither of these count as “radical innovation” in my book. TBL essentially delivered a discount version of Xanadu, and Linux is a nothing more than a Unix clone. So where’s the paradigm-shifting breakthrough?

    Stallman has overseen the re-writing of some code-munging tools, compilers and utilities, nothing new there. His editor is, as I’ve already said, a joke.

    These examples simply prove my point (echoing Lanier) that Open Source makes things cheaper but it doesn’t drive radical innovation.

  • http://metapundit.net/sections/blog metapundit

    Richard said:
    >Stallman has overseen the re-writing of some code-munging tools, compilers and utilities, nothing new there. His editor is, as I’ve already said, a joke.

    Richard it’s ok if you don’t get emacs. Some people like IDE’s, want to use a mouse, etc. There are tons of programmers who swear by vi and emacs however because they have entirely different requirements from your average Eclipse or Visual Studio user. Saying emacs took a long time to use anti-aliased fonts is like pointing out the Visual Studio can’t run in console mode: illuminating but not actually damaging from the point of view of people who use it.

    I tend to agree that Stallman is an extremist as an ethicist but serious programmers tend not to underrate his very real hacker credentials (and tend not to underrate Lisp which is certainly not limited to or even best expressed by Emacs Lisp).

  • http://metapundit.net/sections/blog metapundit

    Richard said:
    >Stallman has overseen the re-writing of some code-munging tools, compilers and utilities, nothing new there. His editor is, as I’ve already said, a joke.

    Richard it’s ok if you don’t get emacs. Some people like IDE’s, want to use a mouse, etc. There are tons of programmers who swear by vi and emacs however because they have entirely different requirements from your average Eclipse or Visual Studio user. Saying emacs took a long time to use anti-aliased fonts is like pointing out the Visual Studio can’t run in console mode: illuminating but not actually damaging from the point of view of people who use it.

    I tend to agree that Stallman is an extremist as an ethicist but serious programmers tend not to underrate his very real hacker credentials (and tend not to underrate Lisp which is certainly not limited to or even best expressed by Emacs Lisp).

Previous post:

Next post: