No matter what strides the new generation of open-source companies make, they all owe a big debt to Linus Torvalds. In 1991 the Finnish programmer started Linux as a project at the University of Helsinki. Fourteen years later the reverberations are still being felt. Thanks to support from giant companies such as Dell Computer and IBM, Linux is now commonplace on big corporate servers -- posting 11 consecutive quarters of growth, according to market researcher IDC.
Hardware companies are selling more than $1 billion in servers to run Linux every quarter, while sales of servers running proprietary software continue to fall. And now, slowly but surely, Linux is making inroads on the desktop as well. According to IBM, 10 million desktops ran Linux in 2004 -- a 40% jump from a year ago.
COMFORTABLE DISTANCE. That progress has been an important foot in the door for all open-source companies. Marc Fleury, chief executive of open-source middleware company JBoss, describes the Linux operating system pioneered by Torvalds as the older brother who fought the tough battles and was able to get the curfew extended and the keys to the car, so that life was a lot easier for the rest of the open-source world.
But Torvalds doesn't relish the role of older sibling. He's reclusive, working from home just outside Portland, Ore., where he lives with his wife and three daughters. And he generally shies away from commenting on other businesses in the open-source or proprietary world. He works for Open Source Development Labs, a Beaverton (Ore.) nonprofit organization that administers and evangelizes the use of Linux in businesses.
In an e-mail exchange with BusinessWeek Online editors, Torvalds discusses his thoughts on where open source is heading and the challenges the Linux community faces. Edited excerpts of the exchange follow:
What do you think of the new generation of open-source applications and services companies popping up today? What does this means for the future of open-source software?
I don't think this is anything new. That has always happened in the open-source community, namely that the area of development ends up expanding. If you look at open-source development, it used to be relegated to only very core and very technical areas.
As those core tools end up being more and more mature, the development effort starts building "higher abstraction"-level tools on top of the old ones. [That] is not to say that the core developers necessarily move on, but what you see is that new blood tends to concentrate on the things that the old projects didn't do, and thus the horizons for open source keep on widening.
The applications and services [companies] are just a sign that the core competencies of open source have grown up enough that these things make sense. It certainly wasn't something you could do five years ago; the infrastructure just wasn't there.
What needs to be done to make Linux an even more popular choice for mobile applications and consumer-electronics devices?
Well, one of the things happening is just the devices themselves growing up, and that will obviously continue. A big issue keeping Linux away from [consumer-electronics] devices was simply that the devices generally weren't capable enough for it to make sense. We're only now starting to see that issue go away.
On the software side, the kernel is actually doing pretty well. The biggest problem I see on that front is just the psychological one: The [consumer-electronic] device manufacturers are coming to grips with the whole open-source community thing, but it will take them a while to be entirely comfortable with it. The same way it took the server-side companies a while to get used to the notion of open source a few years ago.
What about Linux on the desktop? Why hasn't it taken off?
Oh, it has absolutely taken off, but some people seem to think that "take off" means that suddenly everybody is running it. That's clearly not true. It's a very slow conversion. There are more people running it this year than there were last year, and it all looks and works a bit better all the time. But did everybody suddenly convert? No. It's one step at a time.
Some people seem to think that we got where we are overnight. Not so. I've been doing Linux for 15 years, and hey, if it takes another 5 or 10 for the desktop to be a big part of the market, that's what it will take.
As the owner of the Linux trademark, you have begun enforcing it by demanding licensing fees from companies that use it. Why are you doing this, and how do you respond to critics within the open-source community who call you a hypocrite?
I don't know why that thing became news. I've been the owner of the trademark for something like 10 years, and I have not "begun" enforcing it. It has been ongoing all the time. Has it been a big enforcement push? No. [In August, Torvalds did start cracking down on some companies in Australia.] That may be why it took some people 10 years to even notice. But the way trademarks work, you sadly have to enforce them to keep them valid.
Your employer, the Open Source Development Labs, recently proposed a patent commons project, where it would create a legal entity to administer open-source patents and a database to organize the claims, providing protection to open-source developers and companies who implement their software. Why is this important to the open-source community?
Patents in general are a big issue. There are worries about the situation at the U.S. Patent & Trademark Office among pretty much all tech companies. Absolutely everybody knows it's a seriously broken setup, although opinions differ on how to fix it.
The patent commons is just one thing to do. [We're] trying to at least make the situation a bit more bearable for open-source developers that under normal circumstances don't really have a lot of input on patents.
Companies that would be likely to sue open-source developers or users for patent infringement wouldn't likely contribute patents, so why bother?
Part of it is to let people know that yes, the kernel [and other projects] do actually use patented technologies, but they do so with the patents granted to them.
For example, there were all these reports about how the Linux kernel might be infringing on some random number of patents, without anybody to say: "Hey, it's O.K., we're fine with it."
So having a patent pool is part of saying: "OK, we know that project XYZ uses our patents, and that's perfectly fine." So they'd never have sued anyway, but the point is that by saying so explicitly, just the say-so will undo some of the fear, uncertainty, and doubt (FUD) spreading.
There's an effort under way to revise the General Public License, or GPL, that governs the use of Linux and many other open-source software packages. Why does the GPL need to be revised? What changes do you think are in order, and why?
I'm not even going to speculate until I see what the changes are. I'm not in the least interested in a second career as a lawyer.
Do you still worry about the SCO Group (SCOX )and its suit charging IBM with improperly taking Unix code and handing it over to Linux developers?
I don't think anybody does. What's a bit distressing is how long it has drawn out. That's a sad commentary on the whole situation in itself.
来自 “ ITPUB博客 ” ，链接：http://blog.itpub.net/10901326/viewspace-965934/，如需转载，请注明出处，否则将追究法律责任。