McKusick, M.K., Joy, W.N., Leffler, S.J., and Fabry, R.S., A Fast File
             System For UNIX, ACM Transactions on Computer Systems, Vol. 2,
             No. 3 (August 1984), pp. 181-197.

2-line Summary:
This paper describes a reimplementation of the Unix file system to improve
performance.  Details a number of improvements, but doesn't justify a lot
of "magic" numbers.  


Performance Enhancements:

First trick: Increase block size
  pros:
    can transfer more data at once from disk.
  pros:
    can have more data contained within blocks directly referenced from
      inode
  cons:
    many files are smaller than block size (solution: use fragments of
      blocks)

Second trick: Keep reserve disk space
  pros:
    keeping reserve disk space allows disk allocation mgr to ensure that
      blocks belonging to the same file are placed together on disk
  cons:
    we're basically wasting 5-10% of the disk
  q: why 5-10% ?

Third trick: improve layout policy
  * use physical characteristics of disk to inform layout policy
  * cluster related information (inodes, etc).
  * use global and local allocation routines

Results:
  Performance improved greatly

Functional Enhancements
  * long file names
  * advisory locks
  * symbolic links
  * rename system call
  * quotas