The UNIX Time-Sharing System is not a very academic paper. It has the prose and
style of a user's guide or technical overview. The technical details of the
operating system were mentioned in shallow detail in order to motivate
discussion of the power that UNIX offers to users, developers, and
administrators. For example, the design choice of the shell as both a user
program and the primary interface: it offered a close interface to the system
calls offered by the operating system as well as very advanced scripting.

The aspects of the paper that I found the most interesting were those that
dealt with the *choices* that the creators made in designing the system.
Oneh choice is that the kernel will not dicate the contents of files. The very
idea that an OS could restrict the contents of a user's is suprising to someone
who has grown up in a post-UNIX world, but perhaps this shows that the choices
made with UNIX were either revolutionary or popular enough to be replicated
elsewhere. Another choice is that the use of hardware devices is abstracted
into the same level as the filesystem. This allows the programmer to use
everyday hardware devices (TTYs, printers, network sockets) as if they were
regular files. There is no need for extra APIs.

The design of the shell and the file system API given to the programmer and
user of UNIX allows for the pipe model that so many UNIX
utilities are built around. It also simplifies the contruction of the kernel[1].

In conclusion, a design tenant, if you will, that I take away from this paper
is that the OS writer should not dictate how a user must use the system,
enabling the user to be creative with her use of it. Next steps for reading
would be either a more technical paper about the internals of UNIX from the era
of it's creation, or a modern paper that expands on the UNIX-style of kernel.

Further interesting reading:
1 - "the UNIX kernel is nothing but an IO multiplexer"
    http://www.linuxjournal.com/article/2792