Dawn of a computing future
The story of the birth of this new science emphasizes and honors the mathematicians and engineers behind the physics.
Turing's Cathedral
The Origins of the Digital Universe
By George Dyson
Pantheon. 432 pp. $29.95
nolead ends nolead begins
Reviewed by Richard DiDio
Deciding which came first, the ChickenPad or the EggApp, is guaranteed to freeze up any operating system. The more interesting question is: How did our world of hardware and software magically appear out of the ether?
In Turing's Cathedral: The Origins of the Digital Universe, George Dyson describes how a remarkable group of innovators rode the dangerous swells of post-World War II politics and real weapons of mass destruction, creating a new science of computing and the machines to carry out its program. Their almost unbelievable accomplishments unleashed a new age of science, and a new future for our society.
The beginnings of electronic computing have not been revealed in this detail before. With access to long-forgotten materials, and interviews with surviving eyewitnesses, Dyson's history is a fascinating combination of the technical and human stories behind the computing breakthroughs of the 1940s and '50s.
Like many technological innovations, the alpha-point of modern computing is located at the nexus of war and weapons. The role of physicists at the beginning of the atomic age is well documented. It is, therefore, gratifying that Dyson, who will discuss his book Tuesday night at the Free Library of Philadelphia, honors the activities of the mathematicians and engineers behind the calculations that accompanied the physics.
No one played a more important role than Hungarian American John von Neumann, who fled Europe in the 1930s, ultimately joining the faculty of the Institute for Advanced Study (IAS) in Princeton. This independent center for theoretical research has been home to many of the world's most famous scientists, including Albert Einstein.
Von Neumann's mathematical brilliance may be unmatched in the 20th century. With rare skill at blending pure theory and applied mathematics, his forceful presence made him a roving consultant for academia, industry, and the military. It is no wonder that in 1946, the IAS, normally a place where pencil and paper were the sanctioned research tools, agreed to fund von Neumann's proposal to build a radically new type of computer. This machine would fully implement the theoretical ideas of British mathematician Alan Turing.
A decade earlier, Turing had developed his seminal theory of computing machines. He proved that a universal machine with paper-tape memory and appropriate tables of rules could be fashioned to perform any calculation or mimic simpler computing machines. What is remarkable in Turing's work is that the paper tape served as memory that contained both data and instructions for his theoretical machine, i.e., he imagined a stored-program computer. Von Neumann proposed to build a Universal Turing Machine at IAS.
This was a daunting proposal, considering that at the time there was no software, most memory was made up of punched cards, and the word bit had not yet been coined.
Turing's Cathedral is essentially the story of the Electronic Computing Project at the IAS. For five years, von Neumann and principal engineer Julian Bigelow led a group of mathematicians and engineers through political battles, scientific dead ends, academic rivalries, unreliable equipment, and extreme work conditions to produce MANIAC - the Mathematical Analyzer, Numerical Integrator, and Computer.
We are fortunate that they prevailed. The mathematics and engineering developed for the project led to the von Neumann architecture of control unit, central processor, input/output channels, and memory that is the template for all present-day computers.
In parallel, software was developed as algorithms and operating systems naturally evolved. Snippets of code from those days are now used as building-block memes and data in all software programs.
Thus the chicken-egg question has no answer because software development demands hardware just as much as hardware development demands software.
Von Neumann's computer dreams and the race to develop thermonuclear weapons provided a positive feedback loop of need and ingenuity. Not surprising, MANIAC's "birth" is marked by its first trial run in 1951, when it ran nonstop for 60 days computing the dynamics of an H-bomb blast. But a stored-program computer that could be configured to do anything soon spawned science that could not have been invented before MANIAC. By 1953, it was used to simulate weather and climate, biological evolution, artificial life, and neural processing. (All this with only 5 kB of memory - the approximate file size of this review!)
Dyson pays ample homage to those scientists and machines that inspired MANIAC, with special credit to John Presper Eckert and John Mauchly's ENIAC at the University of Pennsylvania. He also presents needed histories of the new sciences engendered by stored-program computing, and the implications of our new world of cloud computing, in which software and data exist independent of hardware.
At its core, though, Turing's Cathedral is Dyson's tribute to von Neumann, whose drive to construct Turing's sacred space, which had previously existed only as pure thought, brought our digital universe into being.
Science historian George Dyson, born in 1953, grew up near the IAS where his father, physicist Freeman Dyson, was a visiting scholar. His boy's-eye view of the dawn of the Universal Turing Machine, combined with his interest in how technology evolves and affects society, give him a unique perspective on this era.
Turing's Cathedral is an important work. It demonstrates that the power of human thought often precedes determination and creativity in the birth of world-changing technology.