(9/20/2000)
Andy Hertzfeld was a member of the original Apple Macintosh team back in 1981, at which time he designed and implemented much of the original Macintosh system software. In that light, it's not surprising that he's still on the cusp of GUI development today as cofounder of Eazel. Eazel's first project, Nautilus, is one of the first open source development projects to aim for commercial success; its innovative file and system management model is poised to open up the world of Linux to a much wider audience. At this year's LinuxWorld Conference in San Jose, I sat down with Hertzfeld to discuss open source, the future of Linux, and Eazel.
Linux Center: What do you think about the size of this year's show compared to last year's?
Andy Hertzfeld: I think it's showing that Linux is growing up, making a transition into the commercial realm, and maturing, I guess, is the best word.
LC: There's an enormous amount of exhibits this year compared to last year. It's exciting.
AH: Yeah, and it's also painful to some people. That was one of the points I was trying to make in my talk (at the O'Reilly Open Source Software Convention), in that I was talking about the early history of the personal computer revolution, and how the commercialization that occurred at that time was painful. And I meant to try to show people that it has happened before, and we've come out on the other side in good shape. The personal computer industry didn't get wrecked by the commercial forces, it became fundamental to everyday life. I see open source making that same transition; it will become the bedrock of the commercial software industry.
LC: You really believe that?
AH: Yes, yes I do. I don't believe that every application will be free software. But I do believe the infrastructure will be free: the common libraries, the things that are the intersection of the applications, because that's where there's the most benefit, that's where it makes the most sense. One experiment you can do to think about that is to try to imagine the Internet with a proprietary architecture. I can't! When I try to imagine it, it just doesn't work. What you end up getting are seven different Internets or ten different Internets. And so as things evolve, I think, in ten years or so, it'll be hard to imagine the computer industry in that way.
LC: How do you see Eazel helping to move that along?
AH: Eazel is just one company among many, many companies making big contributions. The space that we bit off for ourselves is usability, partially because we see that as the next hurdle for the open source movement. Robustness is there, the functionality is getting there, but the usability was where it was falling down, not for every application, but for mainstream desktop-type users. And because we have experience at that, we thought, well, that's what we're trying to do, is make it usable enough, remove that hurdle for adoption.
LC: What qualities would bring Nautilus to the corporate desktop?
AH: Two things: One, we're trying to put the functionality of the machine at the users' fingertips, give them mastery over their machine, make it so their intuition about the software is correct--it works the way they think it does--so they can be more productive. That's very key. The guy sitting at the computer--if we can make it easier for him to get his work accomplished by removing those technical hurdles of trying to figure out why and how to do this or that, that'll help. Also, we're trying to address system management problems. That's very important for corporations, although maybe not our key market at first, only because they have paid people to address that. Whereas the smaller businesses really can't afford to have the guy with the ponytail come by when something goes wrong. We definitely would like to make our stuff great for the enterprise as well, but it's not our very first priority.
LC: In your keynote at the O'Reilly Open Source convention, you said that in proprietary development, it's your position in the organization that controls the decision-making, whereas in the open source model, it's the code that really matters. What's the likelihood of open source development succumbing to the proprietary model?
AH: It's much less likely to happen by virtue of the code being open. Because the code is open, it can be scrutinized by anyone who comes by. There's no barrier, from a third-grader who wants to look at it to Bill Gates himself. Therefore, the arguments are at a different level. When the code is secret, necessarily the technical arguments don't come to the fore as much, because it's not there to discuss... Because the code is open it naturally focuses on the technical merit of the code, and that can't be so when the code is secret.
LC: You also said that the GPL (GNU General Public License) plays a key role in platform fragmentation.
AH: Absolutely. The best way of looking at it is going back to the Unix wars of the '80s, where the commercial Unixes fragmented. Why did they fragment? Each one was trying to get a proprietary advantage, so they each added features that were not available to the other Unixes, in order to get customers to go over to their Unix. And eventually, repeat that for a year or two, and you Balkanize--the Tower of Babel is another way of saying it. The apps can't run across them anymore, they can't take advantage of any of these proprietary features. And the whole reason to add them is for the apps to take advantage of them. The reason the GPL acts as a unifying force is that it requires any added features be made available to all competitors... If you know your competitors can adopt any of these wrinkles you put in, you're less motivated to put them in for the purposes of differentiation.
LC: How would you describe the essential differences between open source and free software?
AH: The open source movement would say the reason to open the code is it makes better software. And Richard Stallman (founder of the free software movement) would say the reason to free software is because we have an ethical obligation to share with our neighbors. So it has a different philosophical underpinning. And the pragmatics also are a bit different at different levels. It's almost always true that anything that meets Richard Stallman's definition of free software would also meet the open source definition, but not vice versa. There's plenty of stuff that meets the open source definition that would not be considered free software.