Firstly a central processing unit for command-controlled processing of alphanumeric data by either mathematical or logical rules;
Secondly a write-read memory for variable data and a read-only memory for programmed commands;
Thirdly a bus system for sequential transmission of all these data and commands as bi-univocally indicated through binary addresses by pages and columns.
With these three parts, von Neumann machines articulated the fundamental structure of information technology as a functional interrelationship of hardware elements. No matter whether their environment supplies alphabetic or numerical data, that is, writing or media-generated values, the commands, data and addresses are all represented internally by binary numbers. The classic distinction between functions and arguments, operators and numerical values has become permeable. However it is precisely this breakdown of the alphabet which also permits operations to be applied to operations, and ramifications to be automated. Which is why computers in principle comprehend all other media and can subject their data to the mathematical procedures of signal processing.82
Data throughput and access time depend solely on physical parameters. Since 1948, when the transistor replaced the tubes/printed circuits of the second World War, and 1968 when integrated circuits replaced the single transistor, in each case reducing the space and time requirement by a factor of ten, real time analyses and real time syntheses of one-dimensional data flows (of speech or music for example) are no longer any problem.83 So the sound engineer can go home. However, for multi-dimensional signal processing in real time, such as is required for television pictures or computer animations, the von Neumann architecture becomes a bottleneck. For this reason large numbers of parallel computers are already in use, and biological and optical circuitry such as is required above all for the simulation of brain functions, is already under development. The day is not far off when signal processing will reach the physical limits of feasibility.84
This absolute limit is where the history of communication technologies will literally come to an end. Theoretically there remains only the question as to what logic this completion will have obeyed. From Freud 85 to McLuhan the classic answer to this was a generic subject – humanity which before of an indifferent or interferent natural world would have externalised first its motor and sensory interface, and finally its intelligence, in technical prosthetics. However if Shannon’s mathematisation of information rested on his “fundamental idea” of inferring, through a conceptual transfer, the “information efficiency of a jammed transmission” from its cryptoanalytical efficiency 86, interference will only be understandable as the interventions of a hostile intelligence, and the history of communication technologies as a series of strategic escalations. Without reference to the individual or to mankind, communication technologies will have overhauled each other until finally an artificial intelligence proceeds to the interception of possible intelligences in space.87