The main philosophical difference between the two multiprogramming
systems is their generality. The "The" system was built with one
purpose in mind, and that was to facilitate the running of batch
processes. The Nucleus concept allowed for a much more general
platform. The idea behind the Nucleus concept was to make it an
extensible system where the programmers could change its mode of
operation dynamically.

The "The" system had a fixed hierarchy of processes which provided the
necessary layers of abstraction on top of the hardware. The Nucleus
concept did not have a fixed hierarchy of processes -- instead users
of the system could dynamically create and destroy processes. The
Nucleus system's only function was to implement the creators' process
concept. This concept allows users of the system to add other programs
which controlled operator communication, program scheduling, and
resource allocation, giving it the ability to change to different
modes of operation without having to remake the entire system.

A more basic philosophical question with regards to these two systems
is - what should comprise an operating system's kernel?  Should
kernels be static, monolithic giants which do everything from
allocating resources to program scheduling (much as the "The" system)?
Or should they have a single, general function upon which everything
else is built as separate programs (like the Nucleus concept)?