Arguing with the Inevitable

by Tim Lee on November 7, 2006 · Comments

This is a Red Hat marketing video, but I think it’s pretty cool:

I’ve written about this in the past: in the long run, open systems tend to triumph over closed ones. I don’t think that Red Hat will necessarily be the open system that conquers the OS market. Personally, I’m rooting for the increasingly-open Mac OS X. But sooner or later, the inefficiencies of creating large software products using a centralized, Soviet development model will render that model unsustainable.

Comments Posted in: Open Source, Open Standards & Peer Production

  • Tim, please explain this: "open systems tend to triumph over closed ones."

    Focus on detailing your conception of "open." You've read Professor Chesbrough right.

    Also, explain this: "the inefficiencies of creating large software products using a centralized, Soviet development model will render that model unsustainable."

    Inefficient? Did you consider the market. Many consider the market as the ultimate efficiency, and it says the most for IP protected technologies. You have an entirely different idea about efficiency, based on abstract principles. The difference between the market and your reasoning, of course, is production. Efficiency needs to factor in what is productive.
  • Noel, by "open" platforms, I just mean platforms that give their users and third parties greater freedom with respect to those systems. So platforms based on open standards are more open that platforms based on proprietary standards. Patent-encumbered platforms are less open than unencumbered platforms. Open-source products are more open than closed-source products.

    As for efficiency, I don't think that's a hard concept to understand. Microsoft spends an order of magnitude more money developing proprietary products like Windows, SQL Server, IE, and IIS than their open source competitors spend developing Linux, MySQL, FireFox, or Apache.
  • I hate it when Apple's propaganda convinces people that they are 'increasingly open'. Apple hates openness; perhaps even more so than Microsoft does. They encrypt their binaries so you can't study them; their iPod/iTunes combo is the most restricted hardware/software combo in decades; you can't choose what hardware to run OSX on; they wrote most of a new office suite without using open standards for documents; they chose BSD over Linux for a kernel because they could have more control; I could go on and on. They're certainly using more open technology, but that's not the same as being open.

    Fascism where the trains run on time is still fascism- this blog of all folks should know that.
  • Agreed Luis, Apple isn't playing as nicely as folks think.

    Tim, you differentiate between "open" and proprietary standards; what do you think of proprietary standards where the company actively licenses out necessary protocols for others to make itneroperable products, or puts those protocols under a BSD like license.

    ***As for efficiency, I don't think that's a hard concept to understand. Microsoft spends an order of magnitude more money developing proprietary products like Windows, SQL Server, IE, and IIS than their open source competitors spend developing Linux, MySQL, FireFox, or Apache.***

    And Microsoft recoups it in revenue right. Now put all the companies that make your FOSS products together. How do they stack up.

    Oh, and please never mention IIS to me again. That program is as reliable as a wooden tennis racket.
  • Luis,

    That's a fair point. Part of the reason I think they're "increasingly" open is that they started from a low base, openness-wise. And look: their web browser is an open source project. They've integrated open source projects like Samba and CUPS into the OS. I don't know how those projects feel about that, but at a minimum it helps to raise the profile of those products. From what I've read, they've wholeheartedly adopted GCC.

    Yes, they've built a lot of proprietary applications on top of that mostly-open foundation. And no, they're not an ideal open source citizen. As you point out, they quickly revert to building closed platforms when they think they can get away with it, as with the iPod.

    But maybe that's the point: even a guy as reflexively anti-openness as Steve Jobs is being forced by economic necessity in the direction of increasing engagement and integration with the open source community. Although that's clearly not as good as having Apple as a wholehearted part of the open source community like, say, Mark Shuttleworth, I think it's still a step in the right direction.
  • Tim, Like you I'm a fan of openness, but I wouldn't say universal openness is inevitable. I wouldn't compare closed systems of production to Soviet planning, but instead Coase's theory of the firm, with Apple being one of the best examples of a successful (closed) firm.

  • their web browser is an open source project


    Because otherwise they would have had to start from scratch. Had there been a BSD-licensed browser they could have forked, like their kernel, they would have. (If they'd wanted to actually work with a community, instead of dominate one, they would have worked with Mozilla instead of KHTML.)


    From what I've read, they've wholeheartedly adopted GCC.


    Only after the FSF nearly took them to court over their fork of it. (Well, it was NeXT who they nearly took to court, but same difference.)


    But maybe that's the point: even a guy as reflexively anti-openness as Steve Jobs is being forced by economic necessity in the direction of increasing engagement and integration with the open source community.


    Now that is a fair point :) I'm troubled, though, when open is chosen because of the pseudo-network-effects of 'free' and the GPL- it smells coercive, which is not the best way to build a robust and healthy ecosystem.

  • Jim Harper
    Noel, you've confused efficiency with profitability. The point is not to maximize corporate profits, but consumer welfare.
  • Tim, again, how do you address "open innovation," as espoused by Professor Chesbrough. It describes the increasingly vertically-decentralized aspect of the software industry, where firms have formalized inputs/outputs that they leverage by IP licensing and FOSS practices.

    You are different from Chesbrough. You look at the code level to describe "open," while Chesbrough looks at the firm, industrial and economic level. Bascially, you're trying to explain what Chesbrough does but you do so by generalizing from the most microscopic level. Hence, your conception of open is quite removed from his.
  • Doug Lay
    Luis:

    I could understand why the GPL might seem coercive as the market share of certain OSS projects approaches dominance, but what is a "pseudo-network-effect?" Are genuine network effects and market coercion mutually exclusive? Far from it, I think.
  • Jim Harper, let me take a different angle then.

    When you talk about effiency, you have a goal in mind. Right? If you want to optimize your means to that goal, say a software program, or R&D; process, you try to get to a point of economies of scale. Efficiency is measured by the rate of increasing returns. Tim's angle is that effiency should be measured in pure dollars, based on comparing FOSS and proprietary technologies (which is why I thought my point about revenue was relevant). He even mentions specific products that compete between these two camps. But just because those products perform similar functions, does not mean that they are comparable for measuring increasing returns. Consider that technology is rarely better in the market just b/c of technological superiority; how does it interact with other technologies, how much are third party developers attracted to make interoperable products. Microsoft's products are more integrated between their own offerings (when they want this to be the case..:) and with third party products that they partner with. Second, and this point relates to the first, what technology do consumers prefer; I'm not talking about the average consumer who wants to buy technology for its purpose, not to tinker with it, and here I'll argue that consumer wellfare is increased when they can buy products that suit their simple needs. The market share of proprietary technologies suggest that consumers are pretty happy with what Tim calls "Soviet" style business.

    These are just a couple examples of how you can't measure efficiency between FOSS and proprietary technologies by pure R&D; dollars along. Jim Harper if you have other things in mind to talk about consumer welfare, I'll address them.
  • Jerry,

    I was indulging in a bit of hyperbole. Let me see if I can flesh out my argument a bit.

    One way to think about the Soviet economy was as a really big firm. Obviously, it didn't work so well, because above a certain size, the firm form of organization has serious diseconomies of scale.

    Over time, people discover more and better ways to organize the production of software in a decentralized manner. At the same time, software products like operating systems and web servers get more and more complicated. Eventually, the diseconomies of scale inherent in the firm form of organization will make that form impractical for producing large software products.

    This isn't to say that there will be no commercial firms. Far from it. But commercial firms will tend to be smaller, and they will tend to build their products atop open foundations, rather than trying to build their own proprietary technologies from scratch. TCP/IP is the poster boy for this strategy. In the 1980s, lots of companies designed proprietary online services built on proprietary protocols. In the 1990s, those all got swept away by the superior efficiency of the Internet. We still have proprietary firms like Google, but they don't try to create their own proprietary network--they take the network as they find it and build tools that work within the larger system.

    By the same token, Apple is a proprietary firm, but Mac OS X is a thin layer of proprietary software atop a foundation of open source software. Apple largely built Mac OS X by taking a lot of off-the-shelf open source software developed by other people and combining them. That's in contrast to Microsoft, which generally tries to build its products itself from the ground up. I don't think it's a coincidence that Apple will be releasing its fifth major revision of Mac OS X (or at least fourth, if you want to call 10.1 a bugfix) since its release in 2001 at about the same time as Microsoft releases its first major revision since Windows XP. They've gotten third parties to do a lot of the work for them.
  • Tim, as long as you won't address my points on monopoly, efficiency, and openness, fine:):):):)

    But tie your argument to the actual industry. How is what you say about efficiency relevant to whats going on with companies rather than some chalkboard.

    Also, every fiscal year, corporations go through annual review period. Often, especially recently, R&D; groups had their budgets downsized, or diminished in annual growth, so they the company can focus on its prioritized innovating activity. So, are you saying that you know more about efficiency or how to plan the innovation process than all the execs and business managers at IP firms. Is this right. And how are firms who are consistent with your views doing.

    Tim, you miss several things. Corporations don't just work on existing products. Further, there's more to innovation than just the product. Companies spend a lot of time and money on integration, licensing, joint development and planning products for 5-10 years from now. DO you consider these things inefficient.
  • Jim Harper
    @Jerry: Having discussed this with Tim a lot lately, I think I get what he's saying. He sees the firm as an inefficient producer of many software products (calling it "Soviet" to add spice). This is consistent with Coase's argument that the firm arises to meet an organizational need the market can't (due to transaction costs).

    The Internet, various software, and Web sites lower transaction costs enough that the market can produce many information products, rendering the firm a less efficient way of producing these goods. (It's not a market that trades in money, but various chits like human capital, reputation, self-satisfaction, etc.)

    Tim sees software as more amenable to spontaneous creation (like a spoken language), than top-down design (like an automobile). His prediction of this as inevitable is mighty bold, but if you give him a long enough horizon, he's probably right.

    @Noel: Yes, efficiency can only be measured with a given goal in mind. If we're agreed that the goal is consumer welfare, then I read Tim's argument to be "I prefer companies, software, platforms, etc. that provide the same benefits at lower cost, in terms of both investment/inputs and price to consumers." So far as I could understand it, your counter was to cite Microsoft's revenue. ("And Microsoft recoups it in revenue right." [sic])

    Microsoft's revenues are not a measure of consumer welfare. Indeed, Microsoft's high and persistent profitability signals likely inefficiency and a lack of competition (again, goal: consumer welfare).

    Read less Chesbrough. Read more Bastiat.

  • Tim and Jim, Where do you two get off convincing me like that?!
  • Jim Harper writes: "This is consistent with Coase's argument that the firm arises to meet an organizational need the market can't."

    I would qualify "markets" as "perfectly competitive" markets. Tim's argument touches on the point of difference between Coase and Ken Arrow.

    Jim Harper writes "Microsoft's high and persistent profitability signals likely inefficiency and a lack of competition."

    How is MIcrosoft inefficient (it makes many more products than any other single firm, or software movement), and how is there lack of competition. A broader question is how is MIcrosoft harming consumers by its profitability.

    To merely cite profitability as harm to consumer welfare suggests a very proactive support for antitrust enforcement. Is this what you mean.

    I will point out though that economists find high turnover rate of incumbents, a healthy onslaught of new entrants and countless new technological products daily in the software industry. If Microsoft is harming ocnsumers, other players in the field are offsetting its effects. So I don't see Tim justifying any kind of change in the industry by calling big firms inefficient.

  • Are genuine network effects and market coercion mutually exclusive?


    No, of course not; I meant merely that 'free' has sticky, anti-competitive effects that are similar to the anti-competitive effects of broad market penetration (i.e., real network effects).

  • Jim Harper
    No, Noel, that is not what I meant. By observing the existence of inefficiency and deficient competition, I did not suggest "harm" to consumers in any meaningful sense of that term. These deficiencies just reduce their welfare by making them pay more than they otherwise could. The subject here is efficient ways to produce information goods. I wrote nothing about antitrust and thought nothing about it as I wrote. Do stick to my words, because your personal gloss seems to miss my meaning.

    And in case you missed it (". . . I don't see Tim justifying any kind of change in the industry by calling big firms inefficient"), the purpose of Tim's post was to make a powerful (if perhaps overstated) case for massive changes in what he sees as an inefficient software industry - though not through antitrust, I don't think.

  • ***Tim sees software as more amenable to spontaneous creation (like a spoken language), than top-down design (like an automobile). His prediction of this as inevitable is mighty bold, but if you give him a long enough horizon, he's probably right.***

    Jim Harper, where do you gather this. Tim sees software amenable to many things, except for the market. The prediction actually completely ignores the market (what will consumers buy), existing TCO analysis (FOSS is cheaper to produce, but consumers may not pay less for it), and technological progress (does FOSS make good mass market products).

    This whole post begs so many questions; Tim has explained 10% of them as hyperboles.
blog comments powered by Disqus

Previous post:

Next post: