Linux advocate Bruce Perens doesn't seem to travel without his soapbox. This week, the open source pioneer and
developer will be stumping for the Linux desktop and his UserLinux initiative at LinuxWorld Conference and Expo. During his State of Open Source address, Perens will tell attendees about his mission to create a common-branded enterprise Linux desktop and server based on the Debian Linux distribution. We caught up with Perens to learn more about his views on UserLinux and the ascendance of the enterprise Linux desktop.
You've predicted that Linux desktops will capture about 30% of the enterprise desktop operating system marketplace by 2006. How did you come up with that number?
Perens Actually, I was looking at Windows licensing. Microsoft really hasn't been able to solve its licensing issues. We had a pretty big push from vendors looking at Linux because of that. Then, we lost some ground because of the licensing changes on Red Hat, but that's going to work itself out, I think.
Secondly, the software you use with Linux on the desktop now works. The Mozilla browser is getting to the point of stability. I just looked at the change log for most of Mozilla 1.6, and the changes are a lot about fixes and not really about big changes in functionality. StarOffice and OpenOffice are in good shape. For these reasons, I think that 2004 is going to see a shift from what pilot Linux deployments to (production) deployment of Linux desktops.
Some IT pros have told me that their primary hopes on the Linux desktop rest with Sun's Java Desktop System and Novell's Ximian and SuSE projects. Do you agree?
Perens I think that people tend to place their hopes on companies, and we have to understand that these are open-source projects. You do not need either Sun or Novell and Ximian to have the GNOME desktop. For example, it is an important point that Sun's Java desktop is not a Java desktop; it is GNOME. Sure, one of the products in the package is a Java Virtual Machine, but it is GNOME that they are selling. That is a pretty big deal.
So, do you see large vendors' interest in the enterprise Linux desktop as a mixed blessing?
Perens I would like to see more small players, rather than fewer. When I look at large players in Linux and open source, my biggest fear is that we will start to perceive Linux as coming from one or two companies rather than where it really comes from. I fear that smaller Linux and open source companies will be swept over by large companies. I mean, there is always an IBM looking around to buy one or two innovative companies. And we don't know if that big company will be working issues important to open source, such as the software patent issue.
So, what's the value in and what's the problem with having more small players?
Perens I think it is important to keep the market well distributed, to have a number of viable small players. Politically, they are in a better place to sponsor Linux and open source than larger companies. Traditionally, it's been the smaller company that is more threatened by the same things that threaten Linux and open source software, things such as patenting and other forms of restrictive legislation. Large companies always have a way around that.
The problem is that we have had small players, but we have never had structures for them to be viable.
In building acceptance of Linux, isn't the backing of large vendors a plus?
Perens I certainly find it a plus to have IBM, etc., involved because their involvement makes people consider Linux more real. On the other hand, you can also have distortion of the facts about open source.
How do facts get distorted?
Perens This happened with one company, whom I won't name because the people there saw that it was a mistake and corrected it. So, that company said publicly that they were going to bring quality engineering to Linux and open source. I called and said: 'You cannot speak that way about a partner.' They hadn't realized that the open source community was a partner, that they had to treat the community like any other business partner.
There is a good deal of naivety among the larger companies and, I would imagine, their customers. Hopefully, as they use this stuff more they will get some understanding of where it really comes from.
Now that a good many corporate IT shops have successfully deployed Linux in their data centers, they tell me that they're much more interested in using open source software. Did you start the UserLinux initiative to give them an simple alternative to big vendors' flavor of Linux?
Perens Yes. I'd rather see these guys working with the Apache team, the GNOME team, and the Linux kernel developers than I would see them working with a big vendor as their engineering partner. The opportunity is there for every company to use open source and for every company that can afford to, to have their own programmers involved directly in the open source project.
Do you think that some corporate executives and IT shops consider large vendors to be more stable than the open source community?
Perens When I look at the question of instability, I always turn back to what happens in the commercial world. Vendors suffer from the vicissitudes of the market. That is no more visible with any company now than it is with Sun Microsystems. Sun is in trouble, and we don't know what Sun's strategy will be several years from now. We don't know if Sun will be acquired or not.
Contrast this to an open source team. We know that many open source projects when initiated have failed. We also know that once an open source project is going, and once people are interested in it, is almost impossible to kill it off. The only thing you can really do is replace it with something better.
So I feel that open source is actually more stable that your typical commercial developer or vendor, and that that fact is not significantly acknowledged by the business.
Perhaps that impression that open source is unstable comes from the volume of debate about almost everything in the community. For example, there was well-publicized dissension when you announced that UserLinux would use the GNOME GUI instead of KDE.
Perens That was very interesting. In dealing with this problem, I had to go through about a month of e-mail protests from the KDE group. But now that is all over, and now our project is going forward. I thought that was really a small price and the benefit is that what we have is something that neither Red Hat nor SuSE has: a coherent desktop strategy. So no, I don't think open debate is a sign of instability, and the dialog with KDE and was not a very bad thing, either.
I suppose the concept of open debate about technology decisions is unheard of for companies that have only worked with commercial vendors.
Perens:If I were an IBM customer, and I had serious misgivings about IBM's strategy, I might take that up behind closed doors with someone there. But I would probably never get to the front office of IBM unless I were a really big customer. With Linux and open source, you almost always have access to a principal. I don't know of people who asked Linus Torvalds about technical issues and were not able to get to him. He is very good about answering the technical questions in his mailbox. I just don't see that happening with your typical CEO.
FOR MORE INFORMATION:Click here for SearchEnterpriseLinux.com's coverage of LinuxWorld
FEEDBACK: How high of a priority is Linux on the desktop to your enterprise?
Send your feedback to the SearchEnterpriseLinux.com news team.