Oral History of Linus Torvalds
CHM Ref: X4147.2008 © 2008 Computer History Museum Page 22 of 41
remember if that was the original name, but there were like three or four different…
Slackware, SLS,
Yggdrasil.
Booch:
Better spell that one because it’s a curiou
s-
Torvalds: Y-G-G-
Booch: D-R-A-S-I-L.
Torvalds:
So it’s the name of the Tree of Life in Norse mythology I think.
Booch: Whoa.
Torvalds:
Which is why it’s spelled so oddly, but I had nothing to do with that. It was not my choice. I
think he chose a Norse
name from Norse mythology because I was from Finland, which I mean if you’re
in the U.S. it’s all the same. Never mind that Finnish mythology is completely different. So I think there
was some forethought there, but kind of not. Off the track. But that’s
kind of been a common theme and
one that I really enjoy how it has not been really planned. There has never been any central planning. I
certainly haven’t done any central planning, and a lot of these things just happened around Linux so the
only interaction that I literally had with Adam was he would send me patches that he wanted, I mean he
needed for his distribution. And that’s actually one of the things that I found to be most important. All the
commercial distributions, they had very different needs than the technical people, and it really added a
whole new balance to the whole thing where if you only technical people who make decisions, the end
result is usually not very good. If you only have marketing and non-technical people who do it, the end
result if really horrible, but you need both. And that was one of the big advantages from the whole
commercial approach was that it gave that kind of balance. I remember that I ended up rejecting a lot of
the patches, because I thought that they should be done differently. But it still gave-- I mean quite often I
ended up rewriting them and saying, okay, I see where you want to go, but I really don’t want to do it that
way. And I was very true early on that I kept a fairly strict control of the source code by just saying quite
often instead of applying somebody else’s patch, I would just say “Okay, that’s a good idea, but I’m going
to do it this way.” And it actually took a while before I learned to just trust people. Partly it was trusting
people in general, bu
t partly it was trusting then specific people so much that okay, I said “This guy knows
what he’s doing. I’ll just apply his patch and not worry about it.” Obviously that’s what I’m getting really
good at these days and that’s really my job.
Booch: Well,
the level of complexity has grown so you can’t go into those details.
Torvalds:
Right, but it’s also one of the things that I think some projects have a hard time sometimes
accepting. Where a lot of Open Source projects tend to have a core group that real
ly can’t let go, and
want to continue to control things at that level and it was something I did early on.
Booch: When did you start letting on? What year would that have felt like that you were letting go?
Torvalds:
It’s gone through many different phas
es. I mean it went to the point where I would trust some
people so much that I would just say okay. When they send me a patch, I worked with them now long
enough that I would say okay, this guy, he knows how I work. I know how he works. I’ll just apply it.
And
Oral History of Linus Torvalds
CHM Ref: X4147.2008 © 2008 Computer History Museum Page 23 of 41
that’s kind of the first stage, but then especially as we started upgrading on our tools and I ended up at
some point being really the limiting factor in applying patches from people. We ended up having this
much more distributed system where now I ju
st don’t apply patches from people I trust. I trust other
people to apply patches from people they trust so it’s this network of trust, and when they say “Hey, take
this whole set of patches,” and you say “Hey, I trust you.” And it’s been an evolution. In fact it’s been one
of the biggest things that’s been evolving is just how you work together with people across the globe,
while still, you have your balance. You have to take a new code, because if you don’t take a new code,
people just get really frustrat
ed and they just say there’s no point in me working at this, but at the same
time you do want to have-- I mean you need to have quality requirements, and balancing those two well is
really, that’s probably the biggest problem we’ve had, and I think we were
doing really well on it
considering how things-
development is going. But that’s been one of the important issues.
Booch: So back in the earliest days what machines were you now developing on, because you probably
moved on beyond that first piece.
Torvalds: It was still PCs but by then I mean we started expanding on support for different kinds of
hardware. Not just you didn’t have the right [ph?] code or the size of the hard disc anymore, and you
could use SCSI hard discs instead of the ST506 controllers and IDE and things like that. And the CPUs
were obviously also getting support for more modern CPUs from Intel.
Booch: How were you supporting yourself during this time frame?
Torvalds: I was at University. In Finland, University is free. Well, you have to pay for
extended health
care which is fifty bucks a year.
Booch: Or a few copies of Linux, how nice.
Torvalds: Yeah. And by the way, you also get these basically now, you get money from the state to
support you, and you get a basically no-fee loan too. Finland has a really good educational system. If you
want to be educated you can do so. You don’t have to have money. You don’t have to have an income.
Booch: And you were teaching a little bit too.
Torvalds: I ended up teaching also first just being like a TA and eventually actually doing some
introductory courses but early on I was just a plain student.
Booch: Which is how, by the way, you met your wife I understand in one of the courses you taught. Can
you relay that story?
Torvalds:
You’re not supp
osed to do that, I guess, but I married her so I made an honest woman out of
her. Where would I meet people otherwise in my basement on my computer? There were no dating sites
back then. So yeah, I met my wife at University.
Booch: There was also in this time frame that you and Andrew got into a bit of a tiff I understand over
microkernels versus-