[SIGCIS-Members] A Biography of the Pixel

[Brian Randell]

Hi:

I've now completed reading "A Biography of the Pixel", and preparing the following review, which I've just uploaded to Amazon:

Alvy Ray Smith's "A Biography of the Pixel" is simply wonderful. It's not just very readable, it's a real tour de force, in terms of the amount of fascinating historical and technical information it contains, and the brilliant way it explains everything, from Universal Turing Machines to the intricacies of the computer-animation techniques involved in the making of the movie Toy Story. The author, an eminent computer research scientist, was one of the two co-founders of Pixar, the pioneering and highly-successful American computer animation studio, now a subsidiary of Walt Disney Studios. He argues very convincingly that just three ideas - waves, computations, and pixels - underlie all the apparent complexity of Digital Light. This is his term for "the vast realm that includes any picture, for any purpose, made from pixels. It extends from parking meters to virtual reality, from dashboards to digital movies and television . . . anything mediated by pixels." (Pixels are profound abstract concepts - "samples of the visual world that are represented by little squares of color".) The book's starting points are the work of Joseph Fourier on waves - an analogue idea - and the (digital) concepts of computation and computability of Alan Turing.  I must admit that before I read this book I did not understand that: "Nothing is lost by going to digital. A discrete digital thing can faithfully represent a smooth analogue thing". This is a very subtle idea that Smith traces back to the work of the Russian scientist Vladimir Kotelnikov in the 1930s on the Sampling Theorem, which he explains and illustrates with remarkable clarity. As Smith puts it: "Each [of these three ideas] is intuitively simple, profound, and beautiful. These are the technological cornerstones of our modern world, and you don't need mathematics to understand them." The first three chapters present these foundational ideas and provide fascinating accounts of the people who made them possible. The following chapter is about the race to build the first actual electronic computer - a race (very narrowly) won by Freddie Williams and Tom Kilburn with their 'Baby' machine at Manchester University, which Smith shows was also the first computer to display a picture constructed from pixels, i.e. to "create Digital Light". In this and subsequent chapters Smith does an excellent job of placing all the developments that he describes in their historical context. He avoids the trap of providing simplistic linear narratives of history. Instead, "for the history of each technology, [he] designed a family flow chart [showing] who got what from whom (whether by hook or by crook) and the often-dense interplay of the cast of players". Each such chart acts as an introduction and explanatory guide to a chapter of detailed stories of the people involved and intuitive presentations of their ideas. By these means, Smith tells the story of Digital Light across a number of eras, each characterised by huge improvements in digital technology and computer capability. He takes this story up to the end of the 20th century, and the first digital movies, most notably Pixar's immensely successful Toy Story, the first entirely computer-animated feature film. It is a great story, and Smith tells it brilliantly.
Full dislosure: The author sought my advice on some aspects of the early history of digital computers, and kindly mentions my name in the Acknowledgements. However I did not see any of the rest of his book prior to its publication.


Expanding a little on the above: The account of the history of computers is very good, but is understandably focussed on issues of most relevance to Digital Light, and on fundamental computational concepts rather than technology, though the impact of technological progress and of Moore's Law are very well covered. Thus it does not provide information on all the early computers and computer pioneers, or cover the growth of the computer industry in general. ENIAC and Colossus are discussed, but not detailed, and neither John Mauchly nor Tommy Flowers are mentioned. However, its detailed well-illustrated explanations of the Universal Turing Machine, and of Sampling Theory, and later of various very sophisticated computer animation techniques, together with the biographical details given of the many scientists and engineers involved, are great!

Cheers

Brian Randell

—
 
School of Computing, Newcastle University, 1 Science Square, Newcastle upon Tyne, NE4 5TG
EMAIL = Brian.Randell at ncl.ac.uk   PHONE = +44 191 208 7923
URL = http://www.ncl.ac.uk/computing/people/profile/brianrandell.html