Existence Precedes Essence - Meaning of the Stored-Program Concept - History of Computing: Learning from the Past - page 174

maxim of French philosopher, Jean-Paul Sartre. “Existence precedes essence.” In this
Sartre wished to indicate that the meaning of human life is not given but created by
individual human deliberation and action. [16] In a parallel I want to agree with Ifrah
that the mere existence of the necessary hardware does not confer the essence of the
computer, rather the operators' use of the machine determines its status.
I propose that while the traditional elements of the von Neumann architecture (a
large fast electronic memory separate from a processor, serial processing of instruc-
tions, a stored program and so on) were necessary for the modern computer they were
not sufficient. What was required beyond these hardware elements was a regulating
ideal that the machine should be as general purpose as possible and a realization that
the instructions themselves were subject to manipulation and generation by the ma-
chine. If endless plugging of wires on the ENIAC had simply been replaced by
endless hand coding of data and instruction numbers in later machines, then the im-
provement in power and flexibility of those machines would have been limited. Fi-
nally, a stored-program machine should have practical achievements that exemplify
these characteristics and aims, not mere theoretical hardware capabilities.
It is illustrative to note that the concept of subroutine libraries were quickly im-
plemented in the first stored-program computers developed. The concept was first
proposed by von Neumann and Goldstine working on the EDVAC project. The ED-
SAC, completed and running in 1949 at Cambridge, was the first machine to employ
such a library. The developers hoped to avoid error by using computer code known to
be reliable and save time in writing code. They disseminated this technique by pub-
lishing a programming manual in 1951 that detailed many of these subroutines. [7]
The Manchester Mark I, which became operational a little later in 1949, also made
use of a subroutine library. This library was an important resource for those who
bought the Ferranti Mark I computer based on the Manchester machine such as the
University of Toronto. [17] Note that these subroutine libraries, while only directly
applicable to identical machines, served as a means of distributing standard program-
ming techniques. Rather than each new computer having to be operated in a com-
pletely new way, techniques could be accumulated.
The reuse of code exemplifies both the standardization that is a hallmark of the
modern computer and the use of the machine to make programming easier. In this
case the operators take advantage of the fact that computer instructions were encoded
in machine readable and therefore machine reproducible form. This feature was
hardly exclusive to the modern computer, but is still an important element of its suc-
cess. The importance of digital copying is even more prominent today as more and
more media is distributed digitally in various ways.
Actual preparation of code by machine methods was slower in coming. One of the
first programs written at Cambridge was called “Initial Orders.” This program con-
verted programs written in programmer's notation of the machine code into the actual
binary code of the machine. [7] While a small step, such a scheme represented an
understanding that the computer could aid in construction of its own programs. How-
ever, operators would have to wait until 1952 before development of more complex
programming aids like programming languages began development. [3] Of course it
was only in 1952 that computers began to become more widely produced. Therefore
developments of these elements of the stored-program concept can be seen to occur
relatively quickly. Programming languages also allowed programs to be distributed