Interview with Linus Torvalds

Q01. When do you estimate the 2.5 development kernel source tree will open? What new features are planned to be included on kernel 2.6?
ANS: Linus Torvalds: I don't want to open a 2.5.x development tree until I'm happy with the pending issues for 2.4.x - it's taken longer than I hoped for, but it's getting there. Within the month..
The biggest issues for 2.5.x (and the eventual 2.6 or 3.0 release) will probably be NUMA and other big machine scalability, along with an overhaul of the disk IO layer. At the same time, most of what affects "normal" users is the continued driver development etc.
The other part of scalability is scaling down (that's the part that most "scalable" projects forget completely about), and there the interesting stuff is mostly about some of the QoS issues that some embedded users have. We'll see where that takes us.

Q02.What are the best parts of the linux kernel in your opinion, and what are the parts than need changing or even rewritting?
ANS: Linus Torvalds: I personally really like our filesystem layer, and in general the "core" code is in pretty good shape. The problem spots tend to be in outlying areas, especially driver code. Much of our SCSI layer really needs to be rewritten eventually, and that's one of the pushes for 2.5.x..

Q03. Are there any plans for a more "visual" way to add/remove drivers and reconfigure the kernel itself somewhat on the fly, which can be incredibly helpful for new users?
ANS: Linus Torvalds: I'm a big non-believer in manual driver and kernel configuration, be it visual or not. Most of the stuff happens automatically, and we're going to make that more and more common. Things like hot-plugging a device and the driver automatically getting loaded is how things are supposed to work, none of this "device manager" stuff.

Q04. Recently you stated that GUIs are important for Linux's market acceptance. Are you happy with the today's offerings of KDE, Gnome or WindowMaker as modern GUI systems?
ANS:Linus Torvalds: I really like KDE, especially the fact that it's more than just a window manager infrastructure, but that there are real applications developed under it too.
Which is not to say that there isn't more to be done - I'm personally happy with what we have now, but at the same time I'd be very very unhappy if it didn't continue to develop to become better..

Q05. What do you think of the FreeBSD 5 kernel and WindowsXP's new features from a clearly technical point of view?
ANS: Linus Torvalds: I don't actually follow other operating systems much. I don't compete - I just worry about making Linux better than itself, not others. And quite frankly, I don't see anythign very interesting on a technical level in either.

Q06. What is your opinion on Hailstorm, .Net and the rest of the technologies Microsoft is preparing to roll out in the years to come? Can these releases have an impact on Linux and if yes, in what way?
ANS:Linus Torvalds: See my answer about not caring what the competition does, but doing my own thing as well as I can..

Q07. Do you believe that the oh-so-many Linux distributions are a good thing for Linux's overall good and future, or a problem that creates forks and inconsistencies throughout the platform?
ANS: Linus Torvalds: Oh, choice is always hard. But we take it for granted in politics, and I take it for granted in Linux. Quite frankly, everybody has slightly different priorities, and working in lock-step simply isn't a good idea. Never has been, never will be.
When somebody who is different shows himself to be different in a _good_ way, that's how development happens.

Q08.What is your opinion on RMS insisting calling Linux as GNU/Linux?
ANS: Linus Torvalds: I don't mind what rms calls the system. I don't think his arguments for the naming are very valid, but hey, at the same time I really couldn't care less.

Q09. There was quite some discussion on the kernel mailing list some time ago about making Linux _truely_ preemptive and tear down the "big giant lock" around the kernel, which brings a number of good things, but for a price. What is your opinion on the issue?
ANS: Linus Torvalds: On the SMP side we've pretty much done it. For all intents and purposes there is no big kernel lock in any important area, and Linux these days scales pretty well, without getting into the nightmare scenario that some UNIXes got to where the locking granularity got so fine that it started impacting performance.
Some people have been playing with using the same locks on UP too, creating a fully preemptible kernel. A lot of people are playing around with the patches, and we'll see when/if I'll integrate them into the standard tree. It's not a high priority for me: they don't add performance (like the SMP scalability does), and if they improve latency noticeably I'd really rather look at why the latency is bad in the first place.
So right now as far as I'm concerned it's one of those "cool features" things, and it will need some prodding from the real world to show whether it is worth it.

Q10. How do you see the future of Linux for the next 5 or 10 years from an engineering but also a marketing eye?
ANS: Linus Torvalds: I don't use a marketing eye, I simply don't care. There are others who do, I'll let them worry about it.
From a technical standpoint, I believe the kernel will be "more of the same", and that all the _really_ interesting stuff will be going on in user space. That's not to say that there aren't problems to keep us occupied in the kernel too, I just don't think they make for all that interesting reading ;)

Q11. Let's think "big" for a moment. How do you see the general future of computing in the years to come? What kind of evolution is the next... revolution for software or hardware?
ANS: Linus Torvalds: I was never a "big thinker". One of my philosophies in Linux has always been to not worry about the future too much, but make sure that we make the best of what we have now - together with keeping our options open for the future and not digging us into a hole.
And I'm not a big believer in revolutions. What people call revolutions in technology were more of a shift in perception - from big machines to PC's (the _technology_ just evolved, fairly slowly at that), and from PC's to the internet. The next "revolution" is going to be the same thing - not about the technology itself being revolutionary, but a shift in how you look at it and how you use it.
What's that shift going to be? Who knows. Maybe it will have nothing directly to do with computers at all, just using computers to create new life-forms or whatever.. Where the _excitement_ is not the tool, but what you can do with it.