De Icaza, a flamboyant 28-year-old programmer from Mexico, is founder of the Gnome project to outfit Linux with a polished graphical user interface. He's also chief technology officer of Ximian, a company that is trying to capitalize on Gnome by linking that user interface to e-commerce services on the Internet.
The success of Gnome, which top Linux seller Red Hat backed heavily, has attracted heavy hitters such as Sun Microsystems and Hewlett-Packard. Both computing giants plan eventually to adopt Gnome as their standard user interface.
Open-source efforts such as Linux and Gnome have a political dimension as well, as de Icaza believes the software will make computing more feasible for poorer nations. De Icaza's position at the crossroads of technology and politics merited an invitation to speak at the World Economic Forum, a gathering of top politicians and executives in Switzerland last month.
But de Icaza faces some challenges as well. Despite some lowered barriers between Gnome and the competing interface KDE, the rift persists. De Icaza acknowledged that the split was "a very bad situation."
CNET News.com's Stephen Shankland spoke with de Icaza on these and other issues in a recent interview.
Why did you change your company's name from Helix Code to Ximian?
It was basically impossible to have any product name called "Helix something." We would be infringing on a lot of trademarks. It wasn't worth the pain. We came up with lots of different names, some I liked, some other people liked. The problem is that we have a very diversified company, given that most people are from the Gnome Project and live in different countries. Nobody likes the same names because in their country that means something obscene or because it means something stupid. So coming up with a name is very hard, and it gets trickier every day.
How did you decide to make a company out of your work in the Gnome
I had a background with working with Linux for a long time. I was involved in the Linux Sparc port and the Linux SGI port. I did the Linux RAID stuff. But back then I remember Applixware was a big thing because it was the first office suite for Linux. They were making these applications available for Linux on Intel on Red Hat, which of course I couldn't use on Linux on Sparc or Linux on the SGI. If you wanted to run some kind of an application, you're basically tied to an Intel system, and you're basically tied to the major distributions. With free software you were not tied with that--you could actually take the software and move it into any platform. So when the KDE Project was launched, I talked to Red Hat and (Free Software Foundation founder) Richard Stallman and a bunch of other people, and I told them, "Guys, we should make sure that we ship KDE because KDE is this big thing for Linux." Of course I didn't actually notice the license at that point, but they said we have licensing problems here, which means that this software is not free.
And that's where the Gnome Project came from?
That's why we started the Gnome Project. On one hand, it was the fact that we needed to have a completely free system. On the other hand, I had just been at Microsoft doing an interview for the Internet Explorer team and for the first time heard about component systems. Initially Gnome stood for a component architecture--it was a Gnu Network Object Model Environment, which was a version of (Microsoft's) COM for Linux. Eventually Gnome became the desktop project, and a year and a half later, we actually produced Bonobo, which was a component architecture. And of course Gnome has grown into having an office suite and now working together with Sun, integrating open office and Bonobo and all that stuff. So it's definitely a lot larger project than it used to be.
What do you think about the split--having the two desktop user
I'm not particularly excited about it. If anything, I'd like everybody to switch over to Gnome and start developing Gnome--if anything because I think that our component architecture is sweet, that we're addressing a major problem in the Linux platform, which is lack of a component architecture. On the other hand, I guess people like the fact that two desktops are competing.
As long as the user interface can only do simple things--copy and
paste--then it's not a big problem. But once you start to add deeper
features, like component models, then it gets more complicated having
the two user interfaces.
Or printing architectures or the font system. It's a very bad situation. Of course we'll try to work with them, but there's a lot of pride involved in this thing.
The difference between KDE and Gnome goes beyond just making graphic applications. It's a matter of creating a foundation, which you can build powerful applications.
What do you think is the biggest challenge right now for Ximian, for
Gnome and for Linux in general?
What's frustrating for Ximian (is that) we don't want to make another Linux distribution. I think that's just stupid. We need to work with other distributions. That's why we support God knows how many distributions. Ximian is very easy to install on any Linux distribution. We have paid a lot of attention to the details, but this assumes that you already have a Linux system in place.
Our core competency is the desktop and making nice applications and focus on usability. So I think that Linux needs to focus on ease of installation, getting more partnerships with (computer manufacturers) so they can ship pre-installed machines, which basically would help a lot for the adoption of Linux.
I still think that configuring a Linux system shouldn't be hard. I think that all (computer manufacturers) have done a very lousy job at doing configuration tools for Linux. So we launched a project inside the company called the setup tools, which is targeted to (work on different versions of Linux or Unix)--it doesn't matter if you're running Solaris or HP-UX or FreeBSD or Linux or Red Hat or Debian--you always get the same user interface for managing your system.
How has the arrival of the Unix people such as Sun and HP changed
Well, I can tell you they're both nice people, and it's interesting that they have both realized that Gnome had a technical advantage over CDE, right? It's interesting to see these two big companies devoting resources to join the Gnome foundation and actually work with us. (One person at Sun) has his team working on different pieces of Gnome. They're doing things that nobody else wanted to do, like RAID documentation. They're contributing to Bonobo. And yesterday, I was talking to him about getting them to work also on doing some SOAP stuff.
Right now (Ximian is) working on a new thing which is called Soup, which is a SOAP implementation. So you can interoperate it with Microsoft.Net services if you want to. You can be a Microsoft.Net server or a Microsoft.Net client--it doesn't really matter.
What is the Ximian business model?
There are a number of things that we're pursuing. We're providing support to a number of companies. One of them is Hewlett-Packard, and we did have a contract before with Turbolinux. But the long-term plans are along the lines of providing services for the Linux platform and integrating services into the desktop.
So Ximian's software is a channel for people who want to have
services that show up on the desktop?
That's right. There are a lot of interesting services on the network that could be integrated into your desktop that would provide the information right away instead of having you know that there's www.whatever.com that provides a service. And that's exactly the vision of Microsoft.Net. That's why making SOAP services is so important because now it's very easy to create services that run on both Linux and Windows that expose the internals of an engine to the world. So now you can expose the internals of Amazon, of the entire suite of things that Amazon exposes in a programmatic way so that you can actually programmatically search for prices in a multiple set of sites.
Have you made technical changes to Gnome to accommodate more operating
Gnome already worked on Solaris and HP-UX before we did anything. But any person (even those without high technical skills) should be able to install Gnome on top of Solaris or HP-UX or anything like that. And the setup tools for instance are designed to support HP-UX and Solaris. And it's good to have somebody now to talk to directly and say, "Hey, I can't figure this out!" And it's part of their job to answer those questions, so that's good. Actually I guess I'm in love with the whole situation!