From: Chuck Reeves (creeves_at_windows.microsoft.com)
Date: Tue Jan 06 2004 - 20:36:22 PST
Having focused primarily on OS/2 and Windows development in my career, I
found this limited review of UNIX to be an interesting read. Written in
1974, the introduction leverages details from a specific installation of
a PDP-11/45 at Bell Laboratories used for language, operating system and
computer network development. The text includes a detailed description
of the hierarchical file system including discussion of ordinary and
special files, directories and some limited access control. The ability
to mount removable file systems at different locations in the hierarchy
seemed quite intuitive and extensible. The description of the file I/O
primitives were quite familiar, paralleling my experience with other
systems very closely. Extending this model to I/O hardware devices seems
logical to me. The section on the implementation of the file system was
effective in helping me visualize the physical layout of the system
tables and their relationship to the bytes making up each file in the
system. However, a diagram would have been useful. The brief mention of
the problem encountered in enforcing quotas in the presence of links was
interesting. I would be inerested in hearing about different strategies
for handling this problem. The section describing the program execution
environment synchronization, and interprocess communication using, fork,
pipe, execute, wait and exit was pretty straight forward. The
description of how fork could use a label was new to me. I had always
heard that both processes would begin execution at the same location
requiring further conditional logic to determine the execution path. I
found the discussion of the shell simplistic with the exception of the
use of parentheses, which I had not encountered before.
This archive was generated by hypermail 2.1.6 : Tue Jan 06 2004 - 20:36:27 PST