Oral History of Linus Torvalds
CHM Ref: X4147.2008 © 2008 Computer History Museum Page 28 of 41
Torvalds:
No, I didn’t. Somebody else gave it the name Tux. It didn’t have a name as far as I was
concerned. The name just stuck. It was a penguin. Calling it Tux makes perfect sense. I’m not going to
argue against that.
Booch: So you were with Transmeta from when until when?
Torvalds:
I was with them from ‘97 to ‘04 or something like that, which was unusually long in the Bay
Area. People seemed to change jobs about twice a year or something. Well, maybe the downturn kind of
put a damper on that, but it certainly felt that way. But I’ve always been the kind of person who I’ll do
something and I’ll stick with it. By the end I was one of the senior people there just because most people
didn’t stick with one thing in the Bay Area.
Booch: You mentioned Oracle and their announcement to port over to Linux and IBM did something
interesting too around that time. Tell me about that.
Torvalds: The IBM announcement.
Booch: What was that announcement?
Torvalds: IBM had been fairly active in Open Source for a while, but then at some point, I
forget exactly
when this happened, they announced that during the next one or two years they were going to put a
billion dollars into Linux or something like this. I forget the exact number. The details escape me, but it
was like a big number.
Booch: Did that surprise you?
Torvalds: That did, yeah, I mean I think it took a lot of people by surprise, and I have to admit I was
pretty skeptical. I mean when it comes to companies making PR announcements it’s a lot easier to do the
talk than walk the walk, right?
Booch: Very true.
Torvalds:
And so I was okay, let’s see how that works out. It did not seem to be a match made in
heaven, let’s be quite honest. My personal view of IBM was this pretty staid company. I think most people
would agree that was kind of an odd bedfellow thing. But I have to say, I mean, that they really followed
up on it, and they put a lot of engineers into it, and they’ve been, I mean I was surprised, and I was
wrong. I’ll freely admit the fact that that really did work out really well.
Booch: Around this time too Linux was starting to make inroads in the embedded space. Did that
surprise you? That was not its original bent.
Torvalds: It was not, but I mean a lot of the things that-- the embedded space was funny. It really, some
of the really early things surprised me because I remember it was one of the first Linux Worlds or
something that some embedded people came up to me and told me how they had worked on this gas
pump that was running Linux because it had a screen on it with like web browser or a TV or whatever. I
Oral History of Linus Torvalds
CHM Ref: X4147.2008 © 2008 Computer History Museum Page 29 of 41
forget the exact details. And now you
can actually see those things, right? But this is probably late ‘90s
something, and the whole notion of running Linux on a gas pump, I was like okay, your guys are crazy,
but in a good way, right? And there was a fridge also.
Booch: A refrigerator running Linux?
Torvalds:
A refrigerator with yeah again LCD screens so that when you’re really bored instead of just
opening the door you can just stand there. Maybe in between, this is the U.S. In between opening the
door and snacking you can just stand there and drool in front of the fridge, so some of that took me by
surprise. At the same time a lot of the other embedded spaces where the actual use wasn’t quite as
lackey [ph?]. I mean, Linux actually made tons of sense because the whole not having to license fees and
things like that, and the flexibility. In that sense the embedded space and Linux make a lot of sense.
Booch: So, in the beginning you mentioned the issues of the influences of the technical and the
commercial kind of really impelling in certain ways. Did the move of the embedded space also push you
in certain ways with the Linux release?
Torvalds:
To some degree the embedded space has been very problematic. I think that’s cultural.
Embedded people are--
they’v
e had--
it’s been getting a lot better over the last couple of years but
especially early on, so what you would have is the embedded manufacturers, they always, I mean they
have some specific platform in mind. And then they port Linux to it. But they are, to them, Linux is a
necessary evil that they need in order to get their real work done, so they would usually do the fastest,
ugliest port that they could possibly do, and not in a way where I mean they make their changes
available, but they wouldn’t really be useful to anybody else. And then two or three years later, they’d do it
all over again, and start basically from scratch because two years later the Kernel they wanted to start
with was completely different from what they had done last time around. So the embedded space-- and I
think that’s what they were kind of used to doing. The embedded space has not generally been all that
great at buying into this whole incremental development and release often, release early kind of model.
Plus the hardware an
d the environment tend to be so specialized that it’s not that relevant to anybody
else. That’s changing partly because the embedded space is growing up.
Booch:
“Growing up” in what sense?
Torvalds:
In a pure hardware sense. It’s too expensive to make. In
the embedded space, it used to be
that you did everything custom. It’s just too expensive to do that. In the embedded space today it’s much
more reasonable to just take a general purpose CPU. They used to be too expensive or too power hungry
or something, but hardware has advanced enough that most of the time when you get an embedded
board, it’s going to be a big standard ARM core, and there’s going to be some specialized stuff around it,
but basically the embedded space has come to where the PC was ten years ago.
Booch:
If I’m not mistaken isn’t the Mars Pathfinder running Linux?
Torvalds: It may well be so. One of them was running BXware [ph?].
Booch:
I’ll have to dig that one up.