Review of Unix

From: Nathan Dire (ndire_at_cs.washington.edu)
Date: Tue Jan 06 2004 - 23:28:12 PST

  • Next message: Ian King: "Review: Ritchie and Thompson, The UNIX Timesharing System"

    My first reaction to "The UNIX Time-Sharing System" is that the UNIX
    environment has changed very little in the last 30 years. I don't know
    whether this is due to the consistency and utility of the interface, a lack of
    innovation in subsequent systems, or merely the early success and prevalence
    of UNIX systems. In any case, it is certainly admirable that the interfaces
    have proved so durable in the face of so many technological advances in
    computing.

    On the other hand, as a result of advances hardware, one aspect of UNIX which
    certainly has changed significantly is the scale of the systems on which it
    runs. The paper mentions some limits which appear quite restrictive by
    today's standards: seven protection bits on files, 14 characters for file
    names, and a maximum file size of 1MB. This is of course a reflection of the
    "fairly severe size constraints of the system and its software", one of the
    major considerations which influenced the design. These limits don't appear
    to have been much trouble to extend; inodes simply have more levels of
    indirection to support longer file sizes, for example.

    Another major consideration in the design is that the authors wanted "to make
    it easy to write, test, and run programs." While this of course makes the
    system popular with the technical community, the average user often finds it
    difficult to master the UNIX environment. This remains an issue even for
    today's versions of UNIX.

    An interesting omission in the system is user-level file locking. The authors
    "take the view that locks are neither necessary nor sufficient in our
    environment to prevent interference between users of the same file." I would
    expect advisory locking on files to be quite useful, especially in the absence
    of any other form of IPC besides pipes.

    It's easy to point out what's been overlooked in this paper by looking at
    modern UNIX systems. In my mind, the remarkable consistency with modern UNIX
    environments is a testament to Thompson's original work.


  • Next message: Ian King: "Review: Ritchie and Thompson, The UNIX Timesharing System"

    This archive was generated by hypermail 2.1.6 : Tue Jan 06 2004 - 23:28:13 PST