From: Shearer, James E (james.e.shearer_at_boeing.com)
Date: Thu Jan 08 2004 - 09:51:33 PST
A review of "The Structure of 'THE' - Multiprogramming System"
The Structure of "THE" - Multiprogramming System (Dijkstra, 1967)
describes an operating system organized in a simple hierarchy of 4
layers below the application layers, introduced an information unit
abstraction ("segments"), and introduces the idea of abstracting all
activities into "sequential processes", an idea that would in later
systems manifest as "light -weight processes" or threads. It is not
clear to me that any of these were actually original ideas at the time
the paper was written, but it does describe them clearly which in itself
is useful.
However, the presentation of these ground-breaking ideas is greatly
clouded by the overall arrogance of the presentation. Two sentences in
particular elicited strong (and opposing) emotional reactions from me.
The first, which ends "... the intellectual level needed for system
design is in general grossly underestimated." is, in my experience, a
significant understatement. I could rant about shoddy systems
engineering for pages, but I will spare you. The second sentence that
caused a strong reaction ended with "... the resulting system is
guaranteed to be flawless." I think something similar was said about
the Titanic. Good design methodology rigorously followed by good
designers, and good implementation practices rigorously followed by good
developers can greatly decrease the incidence of flaws and make testing
easier and more accurate, but still I cannot imagine a system simple and
clean enough for me to claim it is "guaranteed to be flawless". The
arrogance of this statement alone made it difficult to take the rest of
the paper seriously.
James E. Shearer
This archive was generated by hypermail 2.1.6 : Thu Jan 08 2004 - 09:53:36 PST