Information Technology Reference
In-Depth Information
to demonstrate constant progress toward machines that were smaller,
faster, and more reliable, their colleagues in software only seemed to
discover new and more perplexing challenges and diffi culties.
That computing itself was a curious amalgam of disciplinary tech-
niques and traditions drawn from mathematics and engineering was both
an asset and a liability. There is no question that nascent computer pro-
fessionals benefi ted immensely from their ability to make themselves
useful to a broad range of academic researchers. But having interdisci-
plinary appeal was not the same as owning your own discipline. Computer
center personnel had diffi culty shedding their image as service providers
rather than legitimate researchers. In a report to the ACM Curriculum
Committee in 1966, the noted computer scientist David Parnas warned
that computer science was “viewed by other disciplines as a rather easily
mastered tool.” “It is easy, in any fi eld, to confuse the work of a techni-
cian with the work of a professional,” suggested Parnas, “but this is
easier in computer science because a worker in another discipline will
consider himself an 'expert' after learning to use a computer to process
his data.”
41
The development of high-level programming languages exacerbated
this situation. For example, by 1958 the majority of users of IBM's line
of scientifi c computers were using FORTRAN to develop their software.
FORTRAN had been developed specifi cally for scientists, with its syntax
deliberately mirroring conventional arithmetic notation.
42
There was no
reason why a department of mathematics or physics could not offer a
FORTRAN programming course suffi cient for the needs of its faculty
and graduate students.
In fact, this is just what happened. Departments of mathematics,
engineering, and business were able to develop what they saw as perfectly
serviceable courses of instruction in computer programming. Anticipating
a debate that would soon develop in the commercial computing industry,
they considered disciplinary-specifi c training as being
more
relevant than
that provided by computer specialists. It was not computing per se that
was important or interesting; what mattered was the application of
computing to a particular problem domain, and who was better qualifi ed
to teach scientifi c programming than a specialist in that domain. A good
physicist could easily pick up enough programming to get by on, but
even the best programmers could never learn enough physics to become
truly useful. If the physicist's code was not quite as optimal as the pro-
fessional programmer's, it was always possible to buy a more powerful
computer.
