Towards an economics of free software

Posted 9 Jun 2000 at 10:39 UTC by advogato Share This

A number of essays and papers purporting to address The Economics of free software recently. None have been particularly satisfying to this cat. While trying to get to sleep one recent night, some of the thoughts that have been percolating on this subject seemed to gel. Insomniac writings or insightful analysis? Read and decide.

Two of these showed up at Kuro5hin and Slashdot. Someone pointed me to the works of Alan McAdams, whose interest actually predates the whole "open source" hype wave. Do a web search, and you'll see quite a few more of these.

Most of these essays seem to address a very narrow question: why do people work on free software projects? After all, you don't see people manufacturing widgets and giving them away free. Why should software be any different?

To me, the answer is clear. We aren't manufacturing widgets. We are creating art. We do free software because it doesn't get in the way of our art anywhere nearly as much as working in a proprietary framework does. Any person who confuses this with manufacturing widgets is definitely going to come to the wrong answer. Economists just seem more likely than your average person to make this mistake.

Software is not just any old art. It's not just pretty pictures for people to look at and enjoy. It's the creation of tools for people to create, understand, and disseminate information. It's the creation of a digital environment where many of us feel we live, or at least spend much of our time. It's defining the geography and texture of life online. Most importantly, it's collaborative art. A computer artifact built by a single person is no longer particularly interesting. The artifacts that are lovingly built and maintained by vital, thriving communities, now those are interesting.

Software is not such a radical departure from other things in the past. Science is another perfect example of a form of art that is far more interesting in collaboration than as a collection of individual efforts. Isaac Newton articulated the power of collaborative science in his famous quote, "if I have seen farther than others, it is because I was standing on the shoulders of giants."

In this context, we can begin to analyze why the proprietary software market, based on a combination of capitalist, free market economics and artificially constructed "intellectual property" rights has done such a poor job, and why the free software movement has been able to kick its ass in a number of dramatic, important ways, even consuming only a meager fraction of resources.

Traditional free market economics is based on the theory of value and price. Someone produces a commodity at some cost. This commodity has some value to some other person. "The market" determines a price for the commodity, closely linked to the value to the buyer. The difference between the production cost and the sale price is profit. High profits encourage more production. In a large network of interacting producers and consumers, resources tend to get allocated fairly "efficiently," essentially that the sum of value to consumers is maximized. Anyone with a formal training in economics, please feel free to explain where this serious oversimplification misses the mark.

Of course, this model is based on a set of assumptions which are never met in real life. Applying this model to software fails completely. The problem is that value to the consumer is hardly related to the specific software delivered. A great deal of the value, probably nearly all of it these days, is in the network effect. I could write a custom word processor for you that is much better than, say, Microsoft Word. If you were writing a novel, such a thing might in fact be very interesting. But that's not how we use word processors these days - it's primarily to create documents to send to other people, and to view the ones sent to you. By this metric of value, my custom word processor is nearly worthless. You'd send your novel to your publisher, and they wouldn't be able to read it.

Thus, much of the real value of software lies not in the ones and zeros of the code running on your machine, but in the protocols, interfaces, and standards that comprise the network. If these are good standards, then stuff works well together, and there is lots of value. If they're bad standards, then stuff breaks, and there is not quite as much value.

And now we can start to see why proprietary software sucks so much. If you really want to make money in proprietary software, it's important to make sure that your competitor's product doesn't work very well. If you get good enough at the game, it's even sufficient for success. Making your own product work well is not so important. Of course, for an isolated piece of software, you can't negatively affect a competitive product. Ah, but gain control of some important standard, and suddenly you have a lot of leverage. In today's software world, the money is in the proprietary ownership of standards, not in the actual code.

Thus, the total value to consumers, so beautifully optimized by an invisible hand in traditional economics, is horribly distorted by a free market trading in brightly colored boxes full of "intellectual property," with an actual meaning of proprietary ownership of standards.

We can also begin to see why free software sucks less. As a free software developer, I create a tiny little bit of value, whatever I can do in my spare time or whatever, and with little in the way of resources. By the standards of proprietary software, it's not that much. However, all the value so contributed adds up, at least potentially. I'm not poisoning the pool for the other players. If I work on an operating system kernel, I have no incentive to make sure it doesn't work on the hardware of some manufacturer who hasn't paid me the licensing fees I demand.

Economists have noted that prosperity can be a natural (and happy) side effect of freedom. Indeed, this effect explains, at least in part, the relative prosperity of free countries compared with totalitarian regimes. The proprietary software world imposes lots of restrictions on programmers, including NDA's, non-compete agreements, and the lack of rights to proprietary technology. While this model provides monetary riches for the producers of software, programmers who choose the free software model are often able to create more value for consumers, in spite of the far more limited resources available. As in traditional economics, one result of the freedom is prosperity, in this case in the form of high-quality software for users.

I believe this framework also lets us analyze the public good of intellectual property laws. For an isolated, standalone work of art, copyright makes sense. It facilitates the artist getting paid for the work, and discourages rip-off artists from copying it and therefore destroying the price. The market funds the artist, and actually encourages production to more or less match the desires of people for art (ignoring, of course, the standard highbrow/lowbrow debates). In this sense, then, I believe copyright serves the public good.

Software is of course not an individual work of art. It exists primarily (or exclusively) as part of a broad network. Copyright and other forms of "intellectual property" can be used to deny competitors access to the proprietary standards, thereby forcing incompatibilities and poisining the pool. As I have argued, this process reduces the sum total value for consumers, and thus does not serve the public good. Intellectual property does funnel a lot of money into software development. However, it tends to funnel it to the companies that play the "proprietary advantage" game best, rather than the best artists. Thus, I believe it is far from an optimum allocation of resources.

In my humble opinion, these are the questions that economists should be studying, not the question of why artists create art. The implications of a thorough study could be interesting, to say the least.


2/3 Nails, posted 9 Jun 2000 at 15:33 UTC by mcelrath » (Apprentice)

It looks like you've hit 2 of 3 nails on the head. Just to summarize, free software "works" because:

  1. Authors of free software have an incentive to create and follow standards, where proprietary software does not.
  2. The "network effect" that is responsible for code improvement, interoperability, and even disemination of the software.

The last one is that software simply is not a commodity. A commodity by definition must be scarce. Software can hardly be scarce when the cost of distributing it is zero (or nearly so). People try to make software scarce with license agreements, copy protection, and so on, but the fundamental fact remains that I can copy software until I'm blue in the face and it won't cost me a penny.

I think the cost of production is maybe the major factor that the free software has popped up. You say "why don't you see people manufacturing widgets and giving them away for free?" And the simple answer is that it costs money. Widgets are a commodity. There have been various efforts to move open source ideas into other fields, including law, books, and microprocessor design. I think these will ultimately suceed or fail based on how much it costs to produce the widget in the end. books will probably succeed. I'm not sure about law. Microprocessor design probably will not. It's just to expensive to make microprocessors. So some people come up with a collaborative design? No manufacturer would make it. Once you have one you can't modify it and see the changes. Because they do cost money the manufacturers will try to maximize their profits by making proprietary standards, generally getting in the way of any collaborative, "open" effort. The low cost of production for software is a major factor.

The Short Answer, posted 9 Jun 2000 at 20:34 UTC by ncm » (Master)

The economic basis for participation is explained in a few short paragraphs here:

http://linuxtoday.com/news_story.php3?ltsn=1999-08-19-004-10-NW

Everything else builds on that economic fundament. We don't need to invoke Art, even if in fact Art is also involved.

On Invoking Art, posted 9 Jun 2000 at 22:28 UTC by lee » (Apprentice)

On invoking art: I don't see many artists giving their work away. Some do, most don't. Even those who work "day jobs" and pursue their art on their "free" time, typically charge for it. I'll cite the extensive network for art shows and craft fairs as evidence.

Personally, I suspect the success of OS has less to do with economics and more to do with politics. OS is the democratization of software development. Democracies seem to work better than most political systems, including those manufactured inside capitalist corporations.

Software as science, posted 9 Jun 2000 at 23:01 UTC by schoen » (Master)

Software is not such a radical departure from other things in the past. Science is another perfect example of a form of art that is far more interesting in collaboration than as a collection of individual efforts. Isaac Newton articulated the power of collaborative science in his famous quote, "if I have seen farther than others, it is because I was standing on the shoulders of giants."

This analogy is the best way I've found to explain to other people how free software makes sense to me and proprietary software doesn't. I've heard the claim that, once, before software-as-commodity was invented, software development was semi-free and was treated almost like scientific research, because nobody had yet thought of another way to think of it.

I grew up reading very enthusiastic accounts of the work of very idealistic scientists, who mostly believed that they were working on a shared enterprise which by right belonged to all of humanity. The cool thing is that there have actually been a lot of scientists who believed that, and who lived that way.

Those people deserve a lot of respect. But a lot of them also had fun and made a decent living, which is more than many people accomplish, never mind the "benefactor of humanity, explorer of the secrets of the universe" bit.

Nowadays -- but actually starting at least a hundred years ago -- a great deal of science is being done for profit. Fair enough. Unfortunately, the idealistic sense that the knowledge thus obtained belongs to all humanity seems to be fading. Patents (to pick on an old punching bag once again) have really exploded, and a larger proportion of discoveries are either remaining unpublished (trade secrets), or, on the other hand, published only after a patent application has been filed.

I can't say that this is hindering the overall progress of science, because that seems to be doing just fine, thanks. I do think it's hindering the scientific and intellectual culture, when scientists have to keep secrets or draw up patent rights agreements and NDAs before discussing their discoveries.

The public is well-served in one sense by the rapid for-profit development of proprietary drugs -- those drugs get discovered! -- but in another sense more and more of the scientific culture is stripped of the sharing and open collaboration which used to be thought to characterize it.

So, I can't say anything about the practice of science. I've never been a practicing natural scientist. All of my impressions are of cultural things, from reading books and magazines about the work of scientists. Sure, there have been uncollaborative, secretive, and proprietary-minded scientists for ages. Were they thought of as the norm and the paradigm of what a scientist's work is supposed to be like?

A more specific note, posted 9 Jun 2000 at 23:11 UTC by schoen » (Master)

The UC Berkeley Computer Science department was originally founded in the college of Letters and Science by ... mathematicians! This is why you can get a Bachelor of Arts degree in CS there (which is what I was once planning to do); the program was founded by some people who came from the Mathematics department. When I heard that, I thought it was awesome. (I'm very partial to mathematicians. And bookdealers.)

So, once upon a time, in the dark mists of computer pre-history, lots of people thought all this stuff was an exciting new branch of mathematics.

Considering that, it's so easy for me to see why Richard Stallman got upset even at a seemingly innocuous company like Symbolics, considering the contrast between the attitudes of the (idealized) academic and commercial worlds. Who could think of selling mathematics or keeping mathematics as a trade secret in order to profit from it?

Rick Moen said that the current resurgence of free software was a new opportunity, a fresh start, and a new golden era for computing enthusiasts. Isn't it?

I am paid for Free Software development, posted 9 Jun 2000 at 23:20 UTC by Uruk » (Apprentice)

Why people write free software is clear to me. I'm not claiming everybody does it for the same reasons, but I've found my motivations are common to some other people's, and it's not what ESR has outlined in the past.

You were right to say that programming is art. I feel that it is too. When I program, I create something of value, and then I give it away without charging money for it.

See, I think that people who are into other types of art, such as painting, sculpting, etc. would also give away their art for free if it weren't for the difference between software and pottery - you have to buy clay for pottery, but developing software costs nothing other than the computer you already have and some line time to download emacs/gcc/autoconf/linux in general. The production cost on software is nothing other than time.

But still, I'm getting paid for my time, because when I write free software it's FUN. It's interesting. It's stimulating, and it builds my skills, which is something I take pride in. So in that way, I'm being paid, because writing free software is often more entertaining than going to a movie, reading a book, or other diversions that you would normally pay money for.

So basically when I write software, I have zero cost, and I have to give up a lot of time to do it. But since I'm paid for that time, I don't mind doing it. There are also the fringe benefits, such as having people email you and tell you that your program kicks ass and that it's filled a gap in linux. You don't have to write sendmail apache, or any other extremely widely used piece of software to get this effect. In fact, I'm betting that the people who write the small niche apps that appeal to only a few people get the most "fan mail" since they're providing something that the OS wouldn't otherwise have, and people are thankful for that. "Fan mail" that most developers get if they stick to a program long enough is a form of payment for the coding, since I'd be lying if I said it wasn't an ego booster.

Just my $0.02

Some points.., posted 10 Jun 2000 at 02:17 UTC by mettw » (Observer)

  1. I'd say that free software authors are more like master craftsmen who take a lot of care in the quality of what they produce. But ofcourse even craftsmen demand to be paid (and highly) for what they produce.
  2. The idea that free software has been developed with little resources is simply not true. Most free software projects start life being funded in one way or another by the state through universities - either by students or academics. Linux was kik-started by Finnish taxpayers, GNOME in large part by Mexican tax payers and so on. It takes a lot of time to get a software project up and running and funding has to be found to support the programmers.
  3. The idea that software is like a science holds water for the cutting edge parts of the field, and state funding would be appropriate here, but the vast majority of software is bought for purely commercial reasons by businesses. I can't see how one could justify state funding for such software. Capitalism has a legitimate role to play here because the state shouldn't take on the risks associtaed with developing that much software, and if the free market won't fund that much free software development then proprietary software is a better solution to requiring taxpayers to fund software development.
  4. You've confused proprietary software with proprietary standards. Capitalism often produces results that we don't want (eg, the Microsoft monopoly) but this doesn't mean that we should throw away the capitalist system, just that state intervention is sometimes required. On the proprietary standards issue the state can successfully force companies to use open standards in their proprietary software by simply refusing to buy any software that doesn't use these standards. Governments, and especially the US government, have enourmous buying power and can pretty much make companies do whatever they want through it. The fix here is to convince ligislatures that it is a problem, and that it is one that they can easilly, and without expense, fix.

Followup, posted 10 Jun 2000 at 06:33 UTC by raph » (Master)

Thanks to everyone who commented. The essay was written at 3am, and is rather sketchy in some of its arguments. I hope to do an expanded version at some time.

In the meantime, I want to expand or clarify a few points.

1. By "art" I mean something rather broad. It would have to be so to include all of science. But hey, if SVG can include all of DOM and EcmaScript, I should be able to do the same :)

There are a couple of specific features of art that I feel are important. First, money is not the main reason most artists do art. On the other hand, money is the main reason people produce commodities. By using the word "art" I want to emphasize that software is not a commodity.

Second, the difference between good and bad software is extremely complex and difficult to quantify, as is also true of art.

2. I stand by my assertion that free software is developed with a small amount of resources compared with proprietary software. Yes, the free software lifestyle is subsidized pretty heavily by grad student grants. But compared to the resources poured into proprietary software, it's still very little.

3. I did try to emphasize the link between proprietary software and proprietary standards. But standards play such a central role in software that I think it would be difficult to build proprietary software that did not enforce a proprietary standard if you tried. If you broaden the concept of standard to include scripting interfaces, user interfaces, and configuration mechanisms, it's pretty close to impossible.

4. I very explicitly discussed standards in terms of "good" vs. "bad" rather than, say, "open" vs. "closed". Open standards are not always better. In fact, the question of what makes a standard good or bad is, like art, a tricky one. My argument is that proprietary companies have a strong incentive to create bad standards in the sense of making it more expensive and difficult for competitors to implement well. I didn't really say this explicitly, but free software developers have an incentive to support good standards because it's the quickest and easiest way to get things working together.

That said, creating a good standard is really tough work, something that most people (whether in free software, industry, or academia) don't have the stomach for.

5. I am not arguing for state funding of free software here. The point of my essay was to establish a framework for considering the economic implications of different models of software development.

6. I am similarly not arguing against capitalism here. I am, however, strongly arguing against the notion that intellectual property is a God-given, natural extension to the concept of real property, and that an economic system that works well (in the sense of maximizing total value) for real property is by extension going to work well in an intellectual property setting. I further argue that this is not merely ideology, but that the success of free software constitutes and empirical counterexample.

It should not be too surprising that intellectual property works by different rules as real property. In the case of real property, if I "take" it from you, you don't have it any more. In the case of intellectual property, you do. Another way of saying the same thing is that the incremental cost of production is zero.

7. I find it interesting that people still insist on discussing the question of "why participate in free software" even though it it not a central question addressed by this essay. Probably, it's a topic that hasn't really been explained well. Possibly a future Advogato's number should (re-)ask the question more explicitly.

8. I am not saying that artists (or craftsmen, or artisans) should work for free. My real argument is quite a bit more subtle: for software, the distortions imposed by a free market system working with the current artificial notions of "intellectual property" are a more powerful negative influence than the financial incentives for production are a positive one. For other forms of art that are not driven so powerfully by the network effect, the balance may well tip the other way.

9. I'm enough of an idealist to believe that it may be possible to devise a hybrid system combining the best features of both proprietary software (financial incentive for creating value) and free software (disincentive to support proprietary standards). Of course, it's way beyond the scope of this essay to propose any such thing, but maybe in a future one soon, hm?


It was difficult in writing this essay to walk the line between saying things that were obvious to the choir, and skipping so fast through arguments that they would not be convincing to, say, a traditional economist. Judging from the feedback received so far, I still don't have a good sense how I did. I hope the discussion continues.

Software and Art, posted 10 Jun 2000 at 07:13 UTC by yakk » (Master)

I think I kind of understand why painters `have to paint', composers `have to compose' and writers `have to write'. As a programmer I feel that same thrill when I'm coding. That adrenaline when you get that parser to work, or you take that algorithm from O(n^3) to O(n log n) - theres (almost) nothing like it. I go to work all day - writing code, and then I come home jump into /usr/local/cvs/gnome/ and start hacking - what else would I do? There has been times when I've considered quiting full time employment - just so that I've got more time to hack.

Of course this addiction I've got doesn't exactly help my studies, relationships and often work, but I can't help it - I'm hooked. I'm hooked on the thrill of creation - exploring ideas that may never have been explored before - of doing things that people think are impossible.

Followup to followup, posted 10 Jun 2000 at 08:34 UTC by mettw » (Observer)

I probably should have taken a bit longer in my first reply.. The big assumption in my first reply was that capitalism won't fund free software development. Companies like Cygnus are making a good living from supplying consulting services but they didn't have to fork out the money to develop initial versions of their most important tools.

I'm just very suspicious of whether we'll see any more companies like Digital Creations or AbiSoft in the future, since they don't have a monopoly over consulting services to get a good return on the initial development costs. I suspect that an Aladin style `proprietary for the current version' system or something similar will come to dominate in the future.

Hey you!, posted 10 Jun 2000 at 18:54 UTC by aaronl » (Master)

You hackers... pirating our software! You're trading it like a commodity instead of an art!

(seem familiar? ;-))

Let's get real, posted 12 Jun 2000 at 10:39 UTC by vipw » (Apprentice)

Can anyone truly say that he's not just in it for the chicks?

Software is different, posted 13 Jun 2000 at 02:33 UTC by nelsonrn » (Master)

Software is different than anything else we have created in the past. It is art, and it is a capital good at the same time. Software is an economic good. Software that is not written (e.g. a page layout program) is extremely scarce until it is written. -russ p.s. I'd say more, but I only have this little textbox to write in.

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!

X
Share this page