Oral History of Linus Torvalds
CHM Ref: X4147.2008 © 2008 Computer History Museum Page 14 of 41
theme continues. I mean, I'd already spent all my money on getting the computer. I thought 50 bucks
was outrageous. Right? So I had to make due without it. At a later date people actually ended up
sending me a copy, right. So now I do have a paper copy of it. And obviously POSIX actually made
some of the standards, most of them available for free on the Internet, but that was-- by then it was
already way too late. But what happened was, Artie Lencke [ph?] who was the person who told me hey,
there's no standard, also said, "But hey, if you get far enough I maintain…," he had this interest in
operating systems which was why he emailed me. And he maintained one of the FTP sites at the
University of Technology and said-- he ended up saying that, "When you're far enough along, I have a
place for you where you can upload this program." And so during summer of '91 I was basically working
towards something that I knew was going to be an operating system, and my target was basically being at
about the level of MINIX, except with a much better terminal emulation package. That was literally my
target at the time. And I had been talking to some people just because I'd been looking for information.
But I wasn't ready to release yet. And at the time my plan for the release name was to call it FREAX.
Linux was kind of always the code name, and let's face it, I thought it was a bit too egoistical to actually
call it Linux.
Booch: Did you first come up with that name, or did your colleague?
Torvalds: I forget. It might have been my own name that I kind of used. It may have been Otti, it might
have been Lars who was another person who was kind of aware of what I was doing who called it Linux.
But I didn't want to release it under that name. So I wanted to release it under FREAX because it was a
free Unix, but also actually written like a F-R-E-A-X because it was kind of a freak thing. It wasn't to be
taken serious. But I think in September or something like that I actually had my first version where the
definition of the first version was-- I could actually run something under it. It couldn't do a whole lot, but it
actually ran real programs, mainly a shell.
Booch: Do you remember your first email address back then?
Torvalds: My first email address was actually on the VMS system, and that was linus@helsinki.fi or
something like that. But then they realized they'd made a mistake and they couldn't actually use first
names, even unusual first names that nobody else had. So my email address at the time was Torvalds.
It had been changed over, and I kept it that way since, even afterwards when people say now you can
call yourself Linus, I'm now so used to be Torvalds at; it was torvalds@helsinki.fi. But I think the machine
I was mostly on was ‘clob.’ So if you'd seen my old emails they'd usually torvalds@clob.helsinki.fi. And I
think that's the one I used when I-- so now, actually the first announcement wasn't even public. It was
sent out to just the few people I'd talked to saying, "Hey, it's there." I wasn't that proud of the first version.
It ran a few programs, but it wasn't actually really usable for anybody, and it had some known issues. So
the first version was not really ever made publicly announced.
Booch: There was the August announcement in '91. Was that that announcement or was that a
subsequent one?
Torvalds: I think the August announcement was really a pre-announcement. That was the
announcement that I was working on. I hadn't really made a release yet, if I remember correctly. I mean,
this is going back some years. And I was telling people that hey, this is what I'm working on. I'm trying to
basically do a clone of MINIX, which was the target at the time. And I was asking for input. I was really
asking people-- at that point it was like, it actually did a lot of what-- I could see that it would do everything
Oral History of Linus Torvalds
CHM Ref: X4147.2008 © 2008 Computer History Museum Page 15 of 41
I wanted it to do at that point. I mean, this was very naïve
looking back, but I had really low expectations
of what the whole system was supposed to do. But I also wanted to continue my project because I used
to always do programming. And I wanted to have motivation to continue this project, and I could kind of
see that it was nearing its end. So the first announcement was actually asking for people to comment on
what they wanted to see from a MINIX clone.
Booch: What kind of reaction did you get from that announcement?
Torvalds: It wasn't a huge reaction at first. The only people who would be-- hadn't even seen it since it
was made to the MINIX newsgroups where people who were kind of interested in operating systems to
begin with were MINIX oriented, usually at the universities and still wanted to maybe see something else.
And most people probably just thought I was full if it. And so you can't expect to take somebody seriously
who says, "Hey, I'm writing my own OS." I mean, crazy people do that. And so I had a few people
contact me. And there was some thread-- in fact I think most of the reaction to that it was-- at the same
time I'd also made this silly hack to make my finger information change automatically. And I think that
may have been the same announcement where I told people in a PS,
"PS, try to finger me…," nobody
uses finger these days anymore. But that was a way of getting information. And I have this magic named
pipe thing that makes it give different things every time you did it to me, and I thought that was cool. And
apparently a lot more people thought that was cool in my operating system, because I have this pretty
strong memory of that one getting much more attention than the original announcement itself, you know.
So…
Booch: So things became reasonably stable in the September timeframe, and then October you
released the first official version, I understand.
Torvalds: Actually, it was a huge relief to actually get something working. And I think that's why doing
operating systems is so crazy, because there's this-- I mean for six-plus months it was completely
useless. I mean, it wasn't very useful even in October. But before that it was-- it did nothing at all. So for
six months I worked alone on this thing pretty much full time. And it was not usable for anybody. So
there's this huge initial step you have to take to get it into any shape at all. But then in September,
October what happened was some of these small programs started working. And when something starts
working suddenly it's not that hard anymore. Suddenly it's all incremental. I mean people didn't actually
start helping me until like December, January of '92 maybe. But even my own productivity just shot up
when I could just start compiling programs and saying, "Okay, that doesn't work." And the way I actually
approached that was when-- I would take random open-source programs off that net. And when they
didn't work, instead of porting the programs to Linux, I would see that as okay, this is a deficiency in
Linux. So I usually would try to configure it to look like SunOS, which was [what] I was most comfortable
with, because that was the university OS. Sometimes it's MINIX, and just say, "Hey, I will make this work.
I will take this unmodified source code from the Internet and I will make sure that I extend on my
capabilities so that it works, instead of changing it so that it works despite my lack of capabilities."
Booch:
So sort of a random selection of this was an interesting program and…
Torvalds: I mean, it wasn't completely random. Most of them were programs that I used under MINIX.
So it was some very basic-- I mean there's all the really basic Unix tools, the shells, the compiler itself
was actually a big step when I could compile the compiler itself under Linux. Actually the first step was