Information Technology Reference
In-Depth Information
of 1948. This machine sits right at the cusp of the era of the modern computer, but
receives limited attention from histories. The SSEC stored instructions and data in the
same format, could automatically manipulate instructions and used this feature to vary
operation in response to intermediate results. Therefore the SSEC embodied the nar-
row definition of a stored program computer. Yet this feature often went unmentioned
even in accounts that emphasized the narrow definition of the stored-program com-
puter as a, or the, key element of the modern computer. Later machines would be
cited as the first stored program computer. [7]
In light of the emphasis placed on the stored-program concept by the literature, I
found it incongruous that this feature of the SSEC would be so often ignored. This led
me to pay attention to those accounts that recognized the novel features of the SSEC,
but gave reasons to deny it the status of a modern computer. The accounts of the
SSEC by IBM historians such as Charles Bashe and Emerson Pugh suggested some
reasons. The SSEC's use of relays for the bulk of its “fast” memory meant that its
operating speeds were somewhat limited compared to later machines and even the
earlier ENIAC, despite its use of electronic arithmetic to perform computations. Per-
haps more importantly the SSEC had limited impact on later developments. IBM's
first commercial computer the IBM 701 was based on von Neumann's computer at
the Institute for Advanced Study. The only machine design to take direct inspiration
from the SSEC machine was the IBM 650. [8]
Computer scientist and historian Brian Randell offered explicit justification for his
placement of the SSEC. While noting that the SSEC was probably the first machine
capable of arithmetic manipulation of stored instructions, what I call the narrow
stored program concept, he denied it status as the first stored program computer. He
cited its small electronic storage size and claimed it operated more in the style of a
fixed program tape machine. In disqualifying the SSEC Randell in part makes refer-
ence to the need for electronic speed, since he allows that the Manchester SSEM test
machine was the first computer, although not a practical one, despite the fact that its
memory was only larger than the SSEC's with reference to the electronic storage and
smaller when relay storage was included. [9]
However, it is clear that for Randell the stored program means more than the nar-
row concept and electronic speed. In introducing the stored program concept he noted
its significance as allowing universal computation but added that its lasting signifi-
cance consisted in the concept of using the computer to help prepare its own pro-
grams. Randell saw in the stored-program concept the seed of things like compilers
and operating systems. Fixed program tape machines do not realize this sort of flexi-
bility and the SSEC would have had trouble holding its full program in its memory.
[9] Elsewhere Randell rejected the notion that merely storing instructions and data in
the same part of the machine is sufficient to define the stored program concept, refer-
ring to the stored program as a practical engineering realization of Turing's universal
calculator. [10] Randell's analysis argues that the stored-program was not merely
hardware, but a use and attitude towards the potential of that hardware.
Randell was not alone in viewing the stored program concept as richer than the
bare storage of instructions and data together. The late Australian historian of com-
puters, Allan G. Bromley, broke the origin of the “stored program concept” into 10
distinct stages of development. For Bromley the development began with the separate
efforts to create digital electronic arithmetic and storage circuits and ended with the
Search WWH ::
Custom Search