Everything we learned from Fred Brooks and the N-Squared Law predicted that the results should have been completely awful, horrible, unusable, flakey, full of bugs. And yet when I tried it out, I immediately discerned that what I was looking at was, in fact, more featureful, more stable, superior in every way I could measure to every commercial operating system I had ever used. And I found this completely astonishing. After I started running Linux in fall of 1993, I spent three years trying to figure out what was going on. And when I figured out more or less what was going on, I decided OK, now instead of just vaporing about theory, I’m going to run an experimental test now.

There was a project that fell into my hands at that time and I decided to run this project as a deliberate test of this theory, replicating the behaviors that I thought were important, and the experiment came out positive. That piece of software completely took over its ecological niche, and at that point I figured I was onto something and I wrote the paper.

SIMS: What was that software?

"The thing about the Internet is you can’t coerce people over a T-1 line, so power relationships don’t work, and there isn’t really any scarcity economics."

RAYMOND: It was a utility called Fetchmail, which is used for remote mail retrieval, talking to pop and IMAP [Internet Messaging Access Protocol] servers. And ... it will retrieve mail from an arbitrary collection of remote mail servers and merge that mail into your normal Unix mail stream, which means that connections to these servers look as though they’re conventional SMTP [Simple Mail Transfer Protocol] push delivery, even though what’s really going on is this polling thing with pop and IMAP, but the user never has to see that anymore.

SIMS: When you say you tried to replicate Linux’s development methodology, you unleashed it on the Internet?
RAYMOND: The three key things that I did were, one, I sought the largest possible number of users and co-developers. Two, I released early and released often. And the third thing that I did was I gave people ego reward. When they did good things I told them that they had done good things, and I said that they had done good things in public.

The thing about the Internet is you can’t coerce people over a T-1 line, so power relationships don’t work, and there isn’t really any scarcity economics. There’s scarcity economics outside the Internet that affects what people do, but on the Internet itself there isn’t any scarcity economics. So the only game left to play is pure craftsmanship and reputation among peers. So if you can offer people the chance to do good work and be seen doing good work by their peers, that’s a really powerful motivator.

But in order to make that work, you have to be willing to reflect credit outwards and say nice things about people when they do good things. So that was point three, and to repeat points one and two, seek as many developers as possible, because the basic insight of the paper was the more eyeballs you get, the shallower the bugs get. And release early and often because that means that people get instant gratification. They get to see their work going back in, so they instantly feel better about it, and bugs get found and fixed faster.

SIMS: So the scarcity that you looked for was the scarcity of attention and reward?
RAYMOND: That’s exactly correct.

C O N T I N U E D . . . 2 of 2
SIMS: You brought up another point, too, which is that part of what you’re doing to proprietary companies is keeping them honest. Do you want to elaborate on that?

"We keep [proprietary companies] honest in the simplest possible way, by providing an alternative, which, unless they bust their butts really hard, is going to be superior."

RAYMOND: We keep them honest in the simplest possible way, by providing an alternative, which, unless they bust their butts really hard, is going to be superior. And given the number of brains we can put on any given problem, it may turn out to be superior even if they bust their butts really hard. So that puts enormous pressure on them to be humble, it drives their prices down, and it puts pressure on them to be closer to the customer, because if they don’t compete really, really effectively we’re going to eat them for lunch.

SIMS: You’re thinking of competing on quality or features, or both?
RAYMOND: All of the above.

SIMS: What’s the new paper that you’ve written that follows up on some of your thinking?
RAYMOND: Well, I just finished writing Homesteading the Noosphere, and that’s a paper that anthropologically looks at the open source culture, the hacker culture as I like to call it. I’m a hacker and proud -- and no, that does not mean computer criminal, and if you think it does -- go do your homework.

What I do in this paper is look at some of the social and psychological mechanisms in the hacker culture that underlie the development of open source. I expand on this analysis of, if you don’t have scarcity economics and you don’t have power relationships, what’s left? Well, what’s left is the reputation game. So this paper is an analysis of the reputation game and how it’s reflected in the customs that hackers have about who gets to modify source, who gets to redistribute source, and who’s considered the official maintainer of something and why. It turns out that there’s a lot of rich and interesting structure there that reveals some surprising things.

SIMS: Some of your research is probably based on the research project you did following your discovery about Linux, I would imagine.
RAYMOND: Part of the reason the Linux community is interesting is because I have been in open source communities before Linux, but Linux shows us an example of the open source dynamic in action that is both larger and, in a sense, purer than I’ve ever seen before in 20 years of developing software. It’s a really ideal laboratory to see how these mechanisms operate in practice.