Arthur W. Burks

Preliminary discussion of the logical design of an electronic computing instrument (1946)

Inasmuch as the completed device will be a general-purpose computing machine it should contain certain main organs relating to arithmetic, memory- storage, control and connection with the human operator. It is intended that the machine be fully automatic in character, i.e. independent of the human operator after the computation starts. A fuller discussion of the implications of this remark will be given in Chapter 3 below.

First General-Purpose Electronic Computer

The conception, development, and design of the ENIAC are presented in the context of a causal history. Early influences, particularly the differential analyzer and the work of Atanasoff, are detailed. Architecture, electronic and logical designs, basic elements, main units, and problem setup are described, together with the historical contributions to each aspect. Finally, the place of the ENIAC in the history of computers is delineated, both quantitatively and qualitatively, through a comparison with machines from the mechanical, electromechanical, and electronic technologies. Reproduced here with permission of Arthur W. Burks and Alice R. Burks

The first electronic computer: the Atanasoff story

Tells of the design, construction, and subsequent controversy over the first special-purpose electronic computer

The Logic of Automata—Part II

We are concerned in this paper with the use of logical systems and techniques in the analysis of the structure and behavior of automata. In Section 2 we discuss automata in general. A new kind of automaton is introduced, the growing automaton, of which Turing machines and self-duplicating automata are special cases. Thereafter we limit the discussion to fixed, deterministic automata and define their basic features. We give methods of analyzing these automata in terms of their states. Four kinds of state tables--complete tables, admissibility trees, characterizing tables, and output tables--are used for this purpose. These methods provide a decision procedure for determining whether or not two automaton junctions behave the same. Finally, a class of well-formed automaton nets is defined, and it is shown how to pass from nets to state tables and vice versa. A coded normal form for nets is given. In Section 3* we show how the information contained in the state tables can be expressed in matrix form. The (i, j) element of a transition matrix gives those inputs which cause state S~ to produce state Sj. Various theorems are proved about these matrices and a corresponding normal form (the decoded normal form or matrix form) for nets is introduced. In Section 4* we first show how to decompose a net into one or more subnets which contain cycles but which are not themselves interconnected cyclically. We then discuss the relation of cycles in nets to the use of truth functions and quantifiers for describing nets. We conclude by relating nerve nets to other automaton nets.

Logic, biology and automata-some historical reflections

This is a historical and philosophical survey of the relation of logic (including automata theory and inductive logic) to the biological sciences, broadly conceived. Aristotle and his successors formalized portions of deductive discourse, but Leibniz was the first to suggest formalizing language as a whole. Since then many different formal languages have been constructed; they are fair models of certain aspects of language. Leibniz saw the computational possibilities of a formal language, which were later made explicit by Turing (1936–37) and Post (1936). In the 1880s Peirce suggested that Boolean algebra could be used to design relay computers and that evolutionary processes and inductive processes are analogous. The first well-developed applications of logic to biology were McCulloch & Pittss (1943) idealized neuron networks and von Neumanns self-reproducing automata. While these are interesting models, their fit to actual biological phenomena is rough. How may the fit of logic to biology be made closer? Various ways are suggested: more detailed applications, the development of biology in the direction of automata theory, and by using formalisms that combine deduction with induction. Evolution is a statistical or inductive process, but genetic strings play a deductive role. Formal languages are different from natural languages in rigor and precision, yet they give fair models of deduction, induction, and grammar. Other uses of language, such as the empirical, seem basically informal. Consider, however, a computer with appropriate input and output devices which interacts with its environment and communicates in a sophisticated language. It would understand empirical concepts and would verify empirical statements, and hence would model the empirical use of language. The logical design and initial state of any computer, and a fortiori of this computer, can be expressed as a recursive formula of a formal language. Hence the empirical aspect of language is also formalizable.

From ENIAC to the stored program computer : two revolutions in computers

Publisher Summary This chapter discusses two main revolutions in the computer field. The first was the employment of vacuum tubes to make a fast, reliable, powerful, general-purpose computer. This development began with John Atanasoffs slow, special-purpose electronic computer called ENIAC. The second revolution was the stored-program computer. The ENIAC necessarily used electromechanical equipment for input and output, and thereby tested the relative merits of the two technologies for computing. It is claimed that the arithmetic design of ENIAC was influenced mainly by two kinds of calculators: mechanical desk calculators, electrically powered and hand operated; and electromechanical card operated IBM machines. During 1946, the final design and construction stage of the development of the stored-program computer was beginning. It is suggested that von Neumann was the first to see and exploit the fact that when orders or instructions are stored in a high-speed read–write electronic memory, they can be manipulated arithmetically and modified by the machine. His variable address EDVAC code was the basis of the modern computer software revolution.

The invention of the universal electronic computer: how the electronic computer revolution began

This is the story of the causal sequence of the three programmable digital electronic computers that launched the Electronic Computer Revolution: the ENIAC (Electronic Numerical Integrator and Computer); the EDVAC (Electronic Discrete Variable Computer); and the Von Neumann, or IAS (Institute for Advanced Study), Computer. All were designed and built from 1943 to 1951. The chief designers were Presper Eckert, John Mauchly, John Von Neumann, Arthur Burks, and Herman Goldstine.The interacting roles of truth-functional and memory logic with digital electronics are explained, together with the relation of these electronic computers to the theoretical calculating systems of Kurt Godel and Alan Turing.

Models of deterministic systems

The definition of “model of a system” in terms of a homomorphism of the states of the system is evaluated and an alternative definition in terms of sequence generators is proposed. Sequence generators are finite graphs whose points represent complete states of a system. Sequence generators include finite automata and other information processing systems as special cases. It is shown how to define models in terms of a projection operator which applies to any sequence generator which has an output projection and yields a new sequence generator. A model produced by the projection operator is embedded in the system it models. The notion of embedding is discussed informally and some questions raised about the relations of deterministic, indeterministic, and probabilistic models and systems.

The Logic of Evolution, and the Reduction of Holistic-Coherent Systems to Hierarchical-Feedback Systems

The distinction between holistic-coherent systems and hierarchical-feedback systems is best explained by means of examples. We will use examples of both natural systems and of man-made systems, such as philosophical systems and computing systems.

Computers, Control, and Intentionality

Publisher Summary This chapter discusses the nature of intentional information processing and control in humans and robots. Mans highest control capacity is his ability to formulate goals explicitly and work toward them intelligently. Intentional goal seeking employs several underlying computer capacities: sensing, reasoning use of a knowledge base, and action. Insofar as a robot is instructed to respond in definite ways to definite stimuli, the concept of desire is not really needed: the robot just reacts to each stimulus with its appropriate response. Bona fide desires enter the picture only when the value structure is complex and the environment restrictive, that is, when there are competing goals and difficulties in achieving goals. Language plays an essential role in intentional goal seeking; in the future, robots may be used to study various issues about the nature of language. In nature, a genetic program typically performs its construction task by starting operation within a small organism—the fertilized egg.