Comments on Homesteading the Noosphere

The following is a letter I sent to Eric Raymond right after reading his paper Homesteading the Noosphere, an exploration of the property and ownership customs of the open-source or free software culture and a sequel to the earlier and widely acclaimed The Cathedral and the Bazaar. Both papers are interesting reading.

(A brief warning, however: while these two bits of Raymond's writing added some useful vocabulary and analysis around open source, the man himself is a self-aggrandizing narcissist and a deeply unpleasant person who has, among other things, spread conspiracy theories about fabricated plots to fake sexual assault claims, denied that HIV is a disease, and claimed that gay people are more likely to be pedophiles. Some of his technical contributions are worthwhile, but his opinions share a lot in common with the dregs of Internet comment threads. If you decide to engage with him, be aware of the risk to your blood pressure.)

Eric Raymond opens "Homesteading the Noosphere" with the following apparent contradiction: While nearly all free software licenses state clearly that anyone can hack on the source code of an open-source project, the actual culture sets fairly strict guidelines on the manner in which this should be done. Guidelines that are not at all reflected in the source licenses. He goes on to analyze this in terms of conscious ideology and unconscious adaptation.

While I agree with nearly all of the remainder of the paper, my initial reaction to this apparent contradiction is different than Eric's.

I had an immediately different thought about the initial contradiction than the one that you present here. I guess I've never thought about it as a contradiction. Instead, my conception of the open source licenses, which I agree don't bear much resemblence in some fundamental ways to the way the community actually functions, is that they serve as dikes.

There is a surrounding exchange culture that has attempted to integrate into itself a system for handling intellectual property. This has become more important and more complicated over the years as more property of note and value has become intellectual rather than physical. As hackers, we've made an often very conscious choice to not participate in the surrounding exchange economy, and to not use this system of property for at least a substantial portion of what we write.

The difficulty that we encounter in doing this is that the exchange culture both surrounds us and is very well-entrenched. We all have to participate in it for other aspects of our life, and have to pay homage to the command culture that it uses to keep aspects of itself under control. In the event of a dispute of some sort within our gift culture, it would be very tempting for someone to poison our culture by turning to the surrounding exchange culture to obtain power to force their views on the other side of the dispute. Certainly such an act would be in violation of the rules of our culture, but it's possible that someone could get away with it and either poison our culture as a result or pull something out of our culture into the surrounding exchange culture.

In many ways, one can think of hacker culture as being in a situation somewhat analogous to the Netherlands and can think of the surrounding exchange culture as the surrounding ocean. Hacker culture has a lot of power, but it's small and based on an at least somewhat vulnerable foundation (post-scarcity economics). It also doesn't have its own system of force, whereas the surrounding exchange culture is fairly well-armed with a legal system and legal force to back up the dictates of that system. As such, we're in constant danger of being swept away by the ocean.

So in a move very similar to that of the Dutch, we've built dikes.

Open source licenses are not entirely creatures of hacker culture, hence the obvious contradiction between what they say and what we do. They're a defensive line drawn between our culture and the exchange culture. Basically what they say is that the legal system of the surrounding culture is not going to be permitted to intrude into ours and that we are willing to use artifacts of the exchange culture (namely their notion of intellectual property) to enforce this.

RMS made that sort of statement when he originally developed the GPL, and it's that sort of logic that lies behind the term "copyleft": the idea that we're going to take a tool spawned from a culture we don't particularly care for (at least in the context of software projects) and use it against that culture. The BSD license is less zealous but does basically the same thing for similar reasons; Berkeley didn't want to have to get involved in any legal hassles produced by the exchange economy.

So I would argue that licenses are not actually the conscious laws of our culture; rather, they're agreements we've made to guarantee our freedom. They're statements that while we don't think people should fork projects unless absolutely necessary, while we don't think people should be removing credit in any way, and while we don't think people should be trying to take over other people's projects, we refuse to use the legal system to force any of these actions. We prefer our own problem resolution mechanisms, and therefore keep the legal system at arms length with licenses that allow people to do almost anything.

(As an aside, the entire philosophical dispute over licenses basically reduces to whether one should use the license to prevent the movement of products of our culture into the exchange culture or not. It's not surprising that this disagreement is so heated and so core to our culture given that on one side one is keeping back a culture that is anathema to ours in many ways, and on the other side one is limiting a gift and thereby undermining the fundamental support structure of the culture.)

I think you'll find the notions that licenses are there to ensure freedom in emergencies, that licenses grant more freedom than should actually be used in the common case, and that there are other rules not in the licenses to be very widespread and widely stated beliefs in the hacker community. I don't think this is really at all unconscious.

Another interesting observation about licenses is that they reflect the fundamental distrust within the hacker community of problem resolution via appeal to written document. Nearly all of our documentation is guidelines, recommendations, distilled wisdom, and the like; very little of it comes down in the form of Rules, and statements of desire that are phrased like Rules tend to be met with scorn. This is so fundamental as to have resulted in the most Rule-like documents of our culture being called "Requests for Comments." Open source licenses are Rules being used to dispense with all other Rules; they're the one exception we're willing to make to our dislike of written authority because their action is to negate the effectiveness of nearly all other written authority.

Instead of law, hacker culture bases its problem resolution procedures on open communication, debate, and public opinion; the winner is the person who succeeds in raising their reputation among the onlookers in the community high enough that they have sufficient capital to earn victory over the losers. Usually, in the end, by simply producing better code and thereby proving that one is right on technical merit. Occasionally by producing better argument. Sometimes by simply understanding the culture better. Implicit in this is the belief that any principle which at one point would have been considered to have force of law may later prove not to be valid.

This tendency is of particular fascination to me. I've wondered for a long time if whether, in on-line communities, we may be seeing the first example of a governmental system that scorns written law and is capable of scaling.

A much later addendum: in the years since I originally wrote this, it's become obvious that my speculations in the last two paragraphs were rather naive, and the answer to the question posed in the final paragraph is clearly "no." Rather, on-line communities started with strong social controls via peer pressure, which broke down as they scaled following a typical pattern described well by Schneier in Liars and Outliers. As the scope and size of on-line communities have gotten larger, we've seen the same increased need for written rules and more formal, institutional enforcement mechanisms as seen by any other community.

Most of the apparently utopian aspects of early hacker culture were simply artifacts of the well-known fact that governance in small communities is usually simple and nearly automatic, since natural human social enforcement mechanisms work well. As hacker culture has scaled and become a larger part of the surrounding society, these early mechanisms have, predictably, broken down and been replaced with more formal governance processes.