Information Technology Reference
In-Depth Information
prove more a nightmare and mathematical tables would fade from use. The point of
this paper is to illuminate the source of such predictions. Vernon had just finished
reviewing developments and practice in scientific computing up to 1939. In part his
prediction simply projected an extension of previous practice. Punched card machines
used switchboards and tables punched into machine readable forms. Therefore so
would the machines of the future. The developments during and immediately after
World War II changed, or added to, the nature of machine computation in fundamen-
tal ways.
Many attempts have been made to capture the essence of the modern computer.
The goals implicit in such definitions have been various, including attempts to charac-
terize a “computer age” distinct from what came before it and the less reputable task
of establishing priority for inventors. A requirement given by most analysts for a
modern computer is that it be electronic. Electronic speeds made computers capable
of thousands of computations in a second, far outstripping the speed of any human
effort. However, Vernon's description of the dream machine implies a machine with
just such prodigious speeds. So it is not speed alone that separates Vernon's dream
from later developments. In parallel most commentaries suggest that the speed of the
modern computer is only part of the story. Accounts point to the design elements that
make a modern computer a general purpose machine applicable to a wide range of
problems.
An example of these developments in machine computation is the ENIAC. The
ENIAC, Electronic Numerical Integrator and Computer, began operation in 1946 and
resembles Vernon's dream machine. The ENIAC's electronic tubes allowed computa-
tion at speeds a thousand times faster than previous machines and operators controlled
the ENIAC by rewiring switch boards. However, the physical rewiring of the ENIAC
proved onerous and a new system was implemented in 1948. This system controlled
the ENIAC's operations using a series of instructions encoded in a numerical format,
a program. The operators preferred the new system of set-up despite a significant
reduction in the speed of actual computation. [2]
Traditionally historians attribute the articulation of key design elements of the mod-
ern computer to the “Draft report on the EDVAC.” The report became synonymous
with John von Neumann, but was based on the work of a team at the Moore school,
responsible for constructing the ENIAC, lead by John Mauchly and Presper Eckert.
The EDVAC was to be the successor to the ENIAC and an improvement upon it. The
report details various broad design features of the new machine. The report was circu-
lated among many other researchers who incorporated key design features into their
own projects. Examples of seminal features of the EDVAC design were serial opera-
tion of the machine, separate mechanisms for calculation and storage, a hierarchical
arrangement of memories with a small amount of expensive fast storage and progres-
sively more cheap slow storage and the use of a stored-program for instructions. Often
the innovations of the Draft Report are called in summary “the stored-program con-
cept,” after a single feature. The stored-program has been summarized as the storage of
instructions and numeric data in the same format in the memory of an automatic calcu-
lator. [3] This feature makes it possible to perform a conditional branch by manipulat-
ing the instructions based on intermediate results. This in turn makes a machine Turing
complete, that is with the addition of unlimited storage the machine would be a
machine capable of carrying out any possible computation.
Search WWH ::




Custom Search