Existence Precedes Essence - Meaning of the Stored-Program Concept - History of Computing: Learning from the Past - page 175

relatively directly across platforms and so served to unify machines that were still
diverse in terms of hardware.
The importance of these elements can again be seen by comparison with the lim-
ited development of the SSEC. The SSEC made use of subroutines in its programs
often storing them as actual loops of paper tapes in an extensive number of tape
drives. However, engineer A. W. Brooke could recall no instance where subroutines
were re-used in different programs in the four years the SSEC was in use. [15] The
SSEC was never used to generate or check coding because of scarce machine time,
but the IBM 604, a punched card calculator, was used to perform simple arithmetic
checks and manipulations of the instructions. [18] Finally, just as the hardware of the
SSEC left few descendants, the programming techniques also apparently had limited
impact. The lack of standardized, or widely communicated, programming techniques
in the IBM SSEC show how it failed to achieve the full potential of the stored-
program. At the same time the lack of such success helps explain its rapid fall into
relative obscurity.
4 The Protean Computer?
Historian Michael Mahoney made a similar claim about the nature of the computer
when he suggested that the computer was a “protean” machine, its nature determined
by the tasks set and programs written for it. According to Mahoney it has no nature to
guide the way it is adopted as compared to other technologies. [19] Whether his posi-
tion is different from mine or not, I would make a difference in emphasis on certain
points. First Mahoney seemed to seat the protean nature of the machine in its mere
hardware, but, as I explained above, hardware is not sufficient to achieve the requisite
degree of generality. The computer finds wide application because people are ready to
use it in that way. Perhaps more importantly it seems computers become ever more
universal because that is one of the ambitions of their developers.
Components that would make a computer's features too specific were eschewed by
developers, ignored by users or viewed as troublesome, as in the problems associated
with optimal coding of magnetic drum memories. In drum memories information was
only available at one point in rotation, therefore the computer might have to wait for
an entire revolution to retrieve a piece of data or instruction. Users often sought to
optimize these retrievals by distributing instructions and data so that minimal time
was lost to waiting for the drum to rotate. [3] Such optimization was completely
hardware specific though. Later memory technologies became more truly random
access and generic.
Also, I wholeheartedly agree with Mahoney's emphasis on seeing how various us-
ers and communities adapt the computer to their use. However, this does not mean
that the computer is inert. In fact, if the computer were truly inert then adoption would
be a straight forward matter. The problem already mentioned of the “von Neumann
bottleneck” illustrates one example of the difficulty in reshaping the stored-program
computer. Adopting new calculating technology often leads to change in practices and
goals. The fading away of mathematical tables as a resource in scientific computation
provides a good example.