[[!tag linux ]]

I gave a talk about the early days of Linux at the jubilee symposium arranged by the University of Helsinki CS department. Below is an outline of what I meant to speak about, but the actual talk didn't follow it exactly. You can compare these to the video once it comes online.

  • Linus and I met at uni, the only 2 Swedish speaking new students that year, so we naturally migrated towards each other.
  • After a year away for military service, got back in touch, summer of
  • C & Unix course fall of 1990; Minix.
  • Linus didn't think atime updates in real time were plausible, but I showed him; funnily enough, atime updates have been an issue in Linux until fairly recently, since they slow things down (without being particularly useful)
  • Jan 5, 1991 bought his first PC (i386 + i387 + 4 MiB RAM and a small hard disk); he had a Sinclair QL before that.
  • Played Prince of Persia for a couple of months.
  • Then wanted to learn i386 assembly and multitasking.
  • A/B threading demo.
  • Terminal emulation, Usenet access from home.
  • Hard disk driver, mistaking hard disk for a modem.
  • More ambition, announced Linux to the world for the first time
  • first ever Linux installation.
  • Upload to ftp.funet.fi, directory name by Ari Lemmke.
  • Originally not free software, licence changed early 1992.
  • First mailing list was created and introduced me to a flood of email (managed with VAX/VMS MAIL and later mush on Unix).
  • I talked a lot with Linus about design at this time, but never really participated in the kernel work (partly because disagreeing with Linus is a high-stress thing).
  • However, I did write the first sprintf for the kernel, since Linus hadn't learnt about varargs functions in C; he then ruined it and added the comment "Wirzenius wrote this portably..." (add google hit count for wirzenius+fucked).
  • During 1992 Linux grew fast, and distros happened, and a lot of packaging and porting of software; porting was easier because Linus was happy to add/change things in the kernel to accomodate software
  • A lot of new users during 1992 as well.
  • End of 1992 I and a few others founded the Linux Documentation Project to help all the new users, some of who didn't come from a Unix background.
  • In fact, things progressed so fast in 1992 that Linus thought he'd release 1.0 very soon, resulting in a silly sequence of version numbers: 0.12, 0.95, 0.96, 0.96b, 0.96c, 0.96c++2.
  • X server ported to Linux; almost immediate prediction of the year of the Linux desktop never happening unless ALL the graphics cards were supported immediately.
  • Linus was of the opinion that you needed one process (not thread) per window in X; I taught him event driven programming.
  • Bug in network code, resulting in ban on uni network.
  • Pranks in the shared office room.
  • We released 1.0 in an event at the CS dept in March, 1994; this included some talks and a ritual compilation of the release version during the event.