a purchase, either through lack of work or a shortage of financial
means, have recourse to commercial computing centres, which are
very willing to effect calculations for clients. The number of such
centres is steadily growing. In the United States of America alone
there are about a thousand of such service centres at the moment.
At best, one can submit ones calculation problems along with
numerical data to the computing centre by telex connection. One
can have the results returned in the same way. More conventionally,
data and results are transported by mail or by special service. This
procedure, by using either one's own computer, or a commercial
centre should in the long run, be considered particularly unprofi
table. Principally because any bureau would have to have every
possible programme at its disposal in order to comply with the
wishes of any customer, and these for many hundreds of available
computer types. In future, and as far as I am concerned this will
not be in the too distant future, this might lead to mergers of
computing centres and in the long run to the coming into existence
of a kind of "Public Computing Utility", which may be compared
with a public utility for supply of water, gas or electricity, or with
the telephone service.
Somewhere within a country, or in a group of countries, a "giant
computer" would operate at an incredibly high speed, and be
capable of processing all possible programmes. The client would,
by special telex connection, submit his problem to the Centre, where
a sub-computer (a large number of these are operating) receives
input data, which after arrangement, are passed on to the central
computer at high speed as soon as the machine becomes available.
The central computer would probably operate for a single client
for a limited period, say eight seconds, and would then proceed to
the problem of another customer. The result of the problem, which
might then be partially solved, is stored in a sub-computer and is
resumed only when all customers waiting at that time have had
their turn.
This sort of collective use will allow the operation of such a giant
computer at a very low cost per second, and we should not forget
that most geodetic and photogrammetric calculations in future
(10-20 years) will be effected within eight seconds. The beginning
of the development mentioned here is already evident in the USA.
It should be remarked that thanks to the development of steadily
more refined micro-circuits the described giant computer will not
necessarily be very bulky. Reflections on the future of electronic
computers are found in [41].
One can also see the development in computer science from the
rapidly growing number of computer makes and models that are
marketed. From the middle of 1963 to the middle of 1966 this
number increased from 186 to 341. Ten countries and approxi
mately sixty firms were involved in this at the end of the period
referred to. Half of these firms are established in the USA. Together
176
