ALL PHOTOS © GARY WAGNER

As one of the creators of Berkeley Unix, Bill Joy knows a thing or two about developing and marketing a free operating system. Sun Microsystems' chief scientist has survived the Unix wars and has watched both his company and its chief competitor, Microsoft, grow from tiny start-ups to industry giants. Though he has had a major hand in the development of such important Unix technologies as NFS (Sun's Network File System), the Berkeley Unix TCP/IP stack, and the vi text editor, Joy's current obsession is trying to build a thriving development community around Sun's Jini distributed computing technology and its not-quite-Open Source software licensing model. Joy recently accepted Linux Magazine's invitation to dinner, where he gave Publisher Adam Goodman, Executive Editor Robert McMillan, and Associate Editor Eugene Kim the lowdown on what Sun thinks of Linux and Open Source.

Linux Magazine: One of the reasons we wanted to talk to you was that you have a long history with and a broad perspective on Unix and free software. What do you think of Linux? A lot of people talk about it as more than just an operating system.

Bill Joy: It's actually less. It's just a kernel if you want to be technical about it. It's politically incorrect to conflate Linux with the applications. At least one person will get upset. So to be quite precise, it's just the kernel of the OS. When we did Berkeley Unix, we were doing the operating system and all of the applications.

In a lot of ways, the Berkeley Systems Distribution (BSD) was on the road to being free with source available and many of the things that Linux is. But it got hung up in this legal fight between the University of California and Unix Systems Labs.

Those are the accidents of history. Now with Linux, we have this new version of Unix written with similar kinds of values that BSD had. One of the great strengths of Unix is that it's been rewritten and reimplemented several times. Applications with similar names and similar functions are widely understood, which allows this healthy kind of invention and reinvention to occur.

LM: So if it weren't for the lawyers, we'd be called FreeBSD Magazine?

BJ: If BSD had been free, there would have been no reason to rewrite it. The new thing that happened with Linux was cultural. The Internet is now coupling people together in ways that probably couldn't have happened before. How else would the developers have found each other?

I did my work in the era of the magnetic tape. We sent Unix in source form to thousands of people; they sent us a few hundred dollars, because I had to pay for the postage and for the printing of the manuals, and that was our network. It was a postal-age speed thing. It was not very convenient.

LM:Were licensing issues as important back then as they seem to be now?

BJ: No. I knew I needed a license for BSD because at some point Berkeley was going to discover it. So I just took a license from the University of Toronto and modified it a little bit and started using that. I figured if I sent people a tape, and there was nothing for them to sign, they wouldn't take it seriously.

When you give things away for free, often people think that's what it's worth: Nothing. So charging them a small amount and giving them a license to sign actually created a perception of value. I'm not saying the tape didn't have value, but an awful lot of stuff comes across your desk that you just throw away.

LM: So, what did your license actually say?

BJ: I don't remember. It was a one-page thing. I didn't have any lawyers look at it and I'm not a lawyer. I just made it up as I went along.

What happened was that at some point we were getting to be big enough that we were sending out hundreds of these [Unix tapes] a year and charging hundreds of dollars for them. A quarter of a million dollars in revenue is a great deal of money for a graduate student. Scott McNealy likes to say: "To ask permission is to seek denial." And we were operating with that philosophy.

But there were huge amounts of money involved and we were becoming pretty visible. So eventually we decided to send AT&T a letter asking them: "Is this okay what we're doing?" And 18 months later they sent a letter back: "We take no position." We won't answer your question. So that's what it was like to deal with a regulated monopoly of lawyers. That same sort of legal structure is what caused [AT&T] to license the transistor for nothing.

So we couldn't actually get an answer from them and it was only years later that this whole fracas erupted around who owns the code. It turned out their code was as tainted with Berkeley stuff as ours was with theirs, so they eventually came to a truce. That's what I've heard second hand or just drinking wine with people. So there's a very tortured and funny history to all this code.

LM: Have you ever contemplated what it would have been like if you'd released your code under the GNU Public License (GPL) or something similar?

BJ: I don't see what the advantage to it is. The important thing is that people have the source code. I actually think it's fine that people can take BSD and make improvements to it and reap software profit.

I don't think, given where we were and what we were trying to do, that the license made that much of a difference. At Berkeley, we had the model that software is the result of your research. The university tradition is that when you do research, you publish. So not giving people the source code for software meant that you weren't publishing your research. A fundamental model of BSD was: Software is a result of our research. We'll publish it and other people will use it if they choose. If someone commercializes it, I don't particularly care, because if you publish research in a university, people can commercialize it. That's just the way it is.

The important thing in my mind is that people share stuff. We've done something at Sun -- Community Source Licensing -- which is another spin on this. But the fundamental principle in my mind is that people get to see the results of other people's work in a way that they can stand on shoulders rather than on toes. The details can vary; there can be many approaches and they work in different contexts.

I think the GPL is fine. I just don't necessarily agree that it will achieve everything that Richard Stallman thinks it will. I'm not as religious about this as other people are.