Interview with Eazel VP Bud Tribble
Eazel burst on the scene earlier this year. With a large amount of media coverage, such as this (registration-required) New York Times article, its ties back to the developers who produced the Macintosh user interface and its intent to help make Linux a competitor on the desktop made for strong headlines. Eazel's introduction to the Linux community was amazingly amicable. They had made a strong effort to talk to people within the Linux community first and were welcomed by the GNOME community. " Eazel's participation is a truly exciting development for GNOME; they've been hacking on GNOME for quite some time, but the company hasn't been publicly announced before this past week or so. Their Nautilus project promises to be _the_ central feature of the GNOME 2.0 desktop."
LWN: What have the last six months been like?
Bud: We've been hard at work on Nautilus, the file manager for GNOME. We're aiming to have it go out with GNOME 1.4 as a replacement for the current GNOME file manager, gmc. Nautilus uses a lot of code from GNOME, including Bonobo, etc. All the views in Nautilus are using that component architecture. A lot of work we've done has been contributing to Bonobo and other associated GNOME projects that we needed for Nautilus. For example, we've contributed to gvfs, the virtual file system for GNOME. Our overriding goal, of course, was to provide an approachable file manager for inexperienced users that is also valuable for more experienced users.
The other issue is that we've been growing our company, getting funding, all of those sorts of things.
LWN: How well is Bonobo working for you?
Bud: So far, it is working wonderfully. We've been able to do everything we want to do with it for Nautilus. We've been able to have other people contribute Nautilus components back to us as well. For example, here at the show, Sun went ahead and Bonobo-ized the individual parts of Star Office, so you now have Nautilus plug-ins that allow you to view a spreadsheet within Nautilus, etc.
LWN: What about the Gnumeric spreadsheet?
Bud: I don't know the Bonobo status of Gnumeric but, knowing the people working on it, I presume it will be Bonobo-ized as well.
LWN: Does the integration of StarOffice have an impact on other parts of the GNOME community who were building competing Open Source projects?
Bud: I don't have any particular insight into that.
LWN: What about the complexity of CORBA/ORBit?
Bud: It is used kind of like a device driver within Bonobo. The average developer doesn't see it; the Bonobo API provides an approachable access to the engine. We can build high performance code because ORBit is a very lightweight implementation of CORBA.
LWN: What other lessons has your development group been learning over the past six months?
Bud: We've been learning the power of doing Open Source development. Our team is composed of about one third ex-Apple developers, one third ex-Netscape developers and one third hackers from the GNOME community. There has been a lot of learning for the people not from the Open Source community about the benefits of this development model. They've learned how to take advantage of it and leverage it.
Bud: Andy Hertzfeld is one of our key developers, from the Apple background. He was checking in some code to the GNOME CVS one evening, I forget exactly what, but within 5-10 minutes he had email from two people. One note was from London, pointing out a bug in his code. The other note pointing out a better algorithm to accomplish the same task and even provided the code to allow him to use that algorithm. That is very different from traditional development, where you might be lucky to have one other person look at the code. It has given us much faster development and much better.
We also noticed the impact of knowing that others will see what you write. That means you are publishing the code, you are the author, so you do the extra effort to make it good. That's an important part of it. In addition, the whole energy of the GNOME community, which is extremely high, has been great.
LWN: What other members of the GNOME community have you interacted with?
Bud: A lot of the people at Red Hat, plus others in the community, particularly those that we then hired.
LWN: Is there a competition between you and Helix to hire GNOME hackers?
Bud: A little bit, but we're on different coasts, so that splits things some. Now, if there is a GNOME hacker in England, we might both be looking to hire them.
LWN: Any specific concerns you have for the next six months?
Bud: GNOME, in general, is going through a maturing process. This is going to be a common type of maturation, where an Open Source project starts out as purely volunteer and then finds corporations are interested in participating as well. For example, Sun's recent announcements indicate that Sun plans to participate heavily in GNOME in the future.
LWN: 50 engineers dedicated by Sun to GNOME development?
Bud: (Shrug). This presents a challenge. How do you deal with these corporate interests in a way that is positive instead of negative. Part of the thinking behind the creation of the GNOME Foundation, is to provide an arena where corporations can go and interact with the GNOME community. It isn't unique to GNOME. Apache is a good model. As Apache matured, IBM decided to take a strong interest, so Apache formed an organization with a board of directors, etc., to provide a mechanism for corporations to interact. It will be an important time for the GNOME project.
LWN: What if the GNOME Foundation recommends something that the core developers feel is a bad idea?
Bud: That's why the GNOME Foundation Board of Directors are elected from the GNOME community. They aren't from the corporate entities; they are developers elected for that position.
LWN: How many people do you expect to see on the GNOME Foundation Board of Directors?
Bud: I'm interested in seeing a well-understood process, with a reasonable number of directors (10 or so, not 50, not 2). The process needs to be very open and inclusive, in terms of bringing the whole GNOME community into that process.
LWN: Then the board would be elected from the GNOME developers who happen to work for a company that happens to be in the GNOME foundation?
Bud: It will probably require that you be a contributor to GNOME, whether or not a company in the GNOME Foundation is actually paying your salary. It could be someone paid by a company or an organization that isn't in the GNOME foundation, a university professor or student, for example.
LWN: What is the development schedule for GNOME 1.4?
Bud: It is set by the GNOME Steering Committee. It has been discussed to be roughly the end of October or beginning of November. That isn't carved in stone, but that's the intent.
LWN: What are you most excited about with Nautilus?
Bud: The fact that we're finally at the point where people can try it out. Our background is building software for consumers. The most important part of which to do a good job is getting good feedback from consumers. There is a level of indirection there. It is easier for tool developers to build tools for themselves than for a much broader audience. You need to bring that broader audience into the community to get that feedback. There may be very subtle things that you can only find by getting feedback from the user.
In some ways, this is a new thing for the Open Source community. Currently, if a need is seen, someone with that need must write the code. Now we have to build a connection between people with the needs and the people with the code so that non-coders can get their needs met.
Here is one of the things that was kind of a revelation for me twenty years ago when we built the Macintosh. We set up test sessions and videotaped people who were trying out the software for the first time. Originally, we had a screen that brought up a confirmation window with two options, "Cancel" or "Do It". One person kept hitting "Cancel" and then cursed and tried again. We turned up the sound and found he was saying "Dolt or cancel? I'm not a dolt!". The potential misunderstanding is not one that would have occurred to the developer; we had to get the feedback from a non-developer using the software.
LWN: Is the Open Source process enough to bring you that feedback or do you need to do those videotaped trials?
Bud: I think it is a bit of a missing piece in the process. You need a place where you would have a website with videotaped user trials and a to-do list that you could grab to fix issues brought up there. Someone would have to do that, build the website and do the user trials, but there is no reason it couldn't be brought into the Open Source process. Andy has some interest in Eazel doing that once we get to be a bigger company.
LWN: Can you think of any questions you wished I'd asked?
Bud: Not really. The thing we didn't touch on is, so how do we make money?
LWN: Are you making money yet?
Bud: (Smile). No.
LWN: We want to report on that once you've had a chance to try out your ideas and see if they work.
Bud: No one has exactly discovered what is going to work and what isn't. It will be an exciting time.
LWN: What kind of parallels or differences do you see between the Helix coders and the Eazel coders.
Bud: The Helix group is very homogeneous; they've all been hacking on GNOME for a very long time. Eazel has this mixture of cultures and background that I mentioned before (1/3 GNOME hackers, 1/3 Netscape and 1/3 Apple). That causes the two companies to have different internal cultures.
LWN: How homogeneous were the original Mac developers?
Bud: We were pretty heterogeneous. Andy came out of Berkeley Unix, I, myself, was in medical school, Susan Kare was a graphic arts major, Bill Akinson was a drop-out from a biochemistry program; we had all sorts of different backgrounds. Joanna Hoffman specialized in studying ancient forms of writing, cuneiform. Steve Jobs who was Steve Jobs. Susan is now consulting with us half-time. Joanna comes by occasionally and is friendly; she is still good friends with Andy.
LWN: Bud, thank you very much for your time.
For information on Eazel's plans for building revenue, which were mentioned but not discussed in detail, check this March Salon article, which did one of the better jobs of articulating their revenue plans. "While the desktop will be free software, with the source code available to all comers, Eazel has a two-pronged plan that includes not only providing a friendly interface for mainstream computer users, but also making a business out of facilitating easy software installation and automatic updates of the rapidly evolving operating system and the applications that run on it. Its business model is to offer these services on a subscription basis, says Christensen, general manager of online services."
Additional Eazel LinksOriginal Media coverage:
Eklektix, Inc. all rights
Linux ® is a registered trademark of Linus Torvalds