FROM THE BOOKSHELF

Looking ahead, Looking back, Looking closer

The changeover to the year 2000 is two-and-a-half years away. Many are concerned about this millennial shift: some astrologically, others computationally, very few scientifically. (I renewed a magazine subscription for three years recently and double zero showed up subsequently in the date field on the mailing label. I'll take this as a good sign, computationally, not astrologically speaking, that is.) In preparing for that chronological transition, it is useful to learn what the experts are saying about the future, to remind ourselves of the remarkable progress in the past, and to be more critical of evidence offered up in the present. Here are three new books to help us in those preparations.

Beyond Calculation: The Next Fifty Years of Computing
Peter J. Denning and Robert M. Metcalfe
Copernicus/Springer-Verlag, 1997
313 pages, http://www.acm.org/acm97

As part of the celebration in March of the 50th anniversary of the founding of the Association for Computing Machinery in 1947, this volume of 20 essays was assembled. Denning and Metcalfe provide introductions to the three somewhat weakly labeled sections that group the contributions by the 22 invited experts. The sections are: The Coming Revolution; Computers and Human Identity; Business and Innovation. Not surprisingly, one of the themes that runs through these sections is the convergence of computation and communication.

The first essay by Gordon Bell and James Gray provides the expected technological forecasts of the various performance parameters associated with hardware and software. (Be sure that you can count beyond a gigabit.) They say that we will go from LANs to WANs to BANsþfrom local area networks to wide area networks to body area networks. Thus, computers have gone from being remote and impersonal to being `up-close and personal.' Over the next 50 years, computers will go from being widely distributed to being intensely ubiquitous. Vinton Cerf speculates in his essay on what it will be like when they're everywhere. He sees greater incorporation of sensors into computing and gleefully proclaims: þAt last, our VCRs will be programmable via the Web.þ (OK, but at that point I suspect that VCRs will be obsolete.) The prudent programming language originator, Edsger Dijkstra, warns us not to be overwhelmed by the technical forecasts and instead urges us to seek simplicity amid the chaos and complexity. In a neat, dynamic extension of the forest/trees metaphor, Dijkstra says concentrate on the tide, not on the waves. Mark Wieser and John Seely Brown take him on by saying that within the next 50 years we will have the capability to move seamlessly back and forth between the tide and the waves. They say we are witnessing the `Coming Age of Calm Technology.'

Artificial intelligence takes it on the chin from several of the authors. The field seems to have generated initially a hubris not matched by subsequent achievements. At the same time, several other authors see benefits coming from a greater partnership between computers and experts working on the same problem. Accordingly, Terry Winograd argues for a focus on the design of this interaction. Computer science's biggest problem, says David Gerlernter, is how to handle too much data. Donald Norman's approach would be to build computers based on biological cells rather than on silicon chips. Both he and Gerlernter agree that we need to be more aware of the thought-style or cognitive-gait differences between humans and current computers.

The essays in the third section are less provocative than those in the two preceding sections, because they examine topics by now well-traversed: e.g., the trials and tribulations of IBM, the vulnerability of information systems, privacy, leadership in the cyber age, telecommuting and telepresence in general, and the marriage of new research and educational paradigms. As a more particular example, consider Abbe Mowshowitz's essay entitled "Virtual Feudalism." By this he means an emerging political economy based not on land ownership, but rather one based on "abstract forms of wealth that may be located anywhere and moved at will." Traditional `monarchs' of global activity will be challenged by a new `barony of virtual organizations.' This argument has, of course, been raised elsewhere by others, notably by Kenichi Ohmae in his discussion of the decline of the nation-state.

This interesting, but not overwhelming, volume includes biographies of the contributors, a glossary, and an index. Denning and Metcalfe provide an introduction to each section, but it would have been more helpful had they offered a concluding essay.

Looking ahead 50 years is a challenging assignment in any field, much less in computing. Looking back over 350 years of history about the launching of the quantitative era has to be equally daunting. But University of Texas history and geography professor Alfred Crosby was not deterred by the task.

The Measure of Reality: Quantification and Western Society 1250-1600
Alfred W. Crosby
Cambridge University Press, 1997
245 pages, http://www.cup.org

A slim tome with over 400 footnotes most would judge as a scholarly and probably dry monograph. Well, this one is far from dry. Consider this sample quote: þVisualization and quantification: together they snap the padlockþreality is fettered (at least tightly enough and for long enough to get some work out of it and possibly a law of nature or two).þ Crosby's main thesis is that when Europe dumped the þVenerable Model,þ which used numbers merely as symbols, for the þNew Model,þ which uses numbers additionally to count and measure, the West then leaped ahead of the rest of the world. In between the presentation of these two models, Crosby examines the evolution of the thought and practice through the Renaissance with regard to time, space, mathematics, music, painting, and bookkeeping. Along the way, the author reminds us that early clocks had no hands nor faces (the tolling of bells revealed the hour), that early mathematics had no signs for plus, minus, nor for equals (they appeared in the late 1400s), and that Arabic numerals were not invented by the Arabs (the Hindus did, perhaps borrowing them from the Chinese). Decimal notation didn't show up until the 17th century. As people felt increasingly the need to count, whether it was the passage of hours, the passage of sounds, or the passage of coins from one hand to another, quantification progressed. Companions to quantification were alphabetization and tables of contentsþoften overlooked aids to imposing order on an expanding knowledge base. But the real leap forward came with concepts and tools for visualization, argues Crosby.

We are familiar with the development of perspective in the Renaissance and its consequences for drawing and painting, but how many of us realize that the musical staff was, in effect, the first graphþtime on the x-axis, pitch on the y-axis? Once music could be written down, instead of being memorized, more voices, human and instrumental, could practice and participate together in its re-creation. Mapmakers studied the methods of the perspectivists, like Duerer and Alberti. Mercator's grid knowingly distorted the reality of the globe, but clarified the task of the navigator. The debit-credit grid of a ledger page became more comfortable for the growing number of merchants to record a sequence of transactions. Bookkeeping is praised (excessively?) by Crosby as a shaper "of more bright minds than any single innovation in philosophy or science." Pacioli, the propagator (but not inventor) of double-entry bookkeeping procedures, was a colleague of Leonardo da Vinciþthe master visualizer. There clearly was a cross-fertilization of ideas in this era. Some may find Crosby's generalizations a bit too sweeping and the historical loose ends insufficiently tied up. But he is an engaging story-teller of this crucial period in the development of the West and readers wanting a big-picture view will enjoy this book.

To continue with the theme of the importance of visualization to precise thinking, we examine the latest work of Edward Tufte, author, designer, and publisher of two previous works (The Visual Display of Quantitative Information (1983), and Envisioning Information (1990)) that have drawn wide acclaim.

Visual Explanations: Images and Quantities, Evidence and Narrative
Edward R. Tufte
Graphics Press, 1997
156 pages, http://www.cup.org

The third in a trilogy (and one hopes there are more volumes to come) of works on how to 'paint pictures' of high information content and thought-value. The first book in the series dealt with pictures of numbers, the second dealt with pictures of nouns (mainly maps), and the latest, with pictures of verbs. Building visual chains of reasoning without weak links is Tufte's goal in his latest work. "What are good strategies for the effective portrayal of explanations?" he asks. A statistics professor at Yale, Tufte has brought together an incredible diverse set of sources in providing good and bad examples to make his points.

There is an introductory chapter on images and quantities, wherein he laments the unthinking de-quantifying of data (e.g., scale omissions and changes). Chapter 2 deals with presenting evidence for decision making, which includes extended coverage of the graphical aspects of the post-mortem of the disastrous Challenger shuttle launch. Tufte observes here that the failure of pre-launch chart makers to display causally what pre-launch analysis had developed causally lead to the fatal go-decision. Had they scaled certain plots rather than just ordered the data, then the role of temperature in o-ring deterioration would have been more clearly revealed.

Another chapter deals with attempts to explicate magic tricks. One can learn a lot about visual explanation in trying to document sequences that the magician doesn't want you to see. Magicians, unlike good teachers, try to avoid clarifying repetition. In the chapter entitled "The Smallest Effective Difference," the author argues that using this strategy (e.g., for coloring iso-layers in contour plots) actually allows for more differences to be portrayed. The final three chapters address: parallelism in thought and picture, multiples in space and time, and interesting juxtapositions in visual narratives. The author derides the excessive "spatial imperialism" present in many web pages and computer-screen layouts wherein excessive icons and supposedly clever backgrounds reduce the space for meaningful content. Tufte demands that a good chart say the most with the least amount of ink. This book, like its two predecessors, is attractively and intelligently presented; all three deserve the reader's attention.