IF YOU thought an academic’s life consisted of marking a few exams, leisurely
lunches and long, long holidays in Tuscany, you should read The Visionary
Position by Fred Moody. Moody’s book chronicles the rise, but mostly fall,
of the virtual reality industry in the mid-1990s.

At the centre of the story is Thomas Furness, VR visionary and founder of the
Human Interface Technology Laboratory at the University of Washington, Seattle.
The book tells how Furness, childhood rocketeer and frustrated astronaut, became
embroiled in a bitter high-tech soap opera surrounding the fortunes of assorted
VR start-up companies, all competing for a position in the emerging industry. As
Moody says, Washington State had its own “digital gold rush”.

Furness had developed the head-mounted display for VR that allows images to
be beamed directly into the user’s eyes. Full of ideas but strapped for
cash—a condition perhaps typical of all university research
laboratories—Furness made an understandable mistake. He entered into a
disastrous partnership with a company.

The company failed to exploit the technology or fund research, alienating
Furness’s key scientists. The scientists left the lab to join other companies,
but instantly found themselves in much the same situation. They in turn became
caught up in an entrepreneurial whirl of venture capital, IP0s (public
floatations of a company on the stock market), and deal-making for which they
were unprepared. Furness, faced with seeing his life-long dream of cheap and
freely available VR technology fade into nothingness, became disillusioned and
eventually ill through overwork. He ended up in hospital.

Moody, a science writer whose book displays little sympathy for the company
suits, chronicles all of it with ugly precision: the lust for stock options, the
boardroom battles between scientists and marketeers, the random hirings and
firings. He concludes that the inventor will invariably lose out to the venture
capitalist, who has an eye for the deal. As one of the protagonists in the story
puts it sadly, “they always have to shoot the inventor.”

Information Ecologies by Bonnie Nardi and Vicki O’Day, MIT Press,
ÂŁ19.50, ISBN 0262140667

Lasting fame is a rare beast. Out of the billions born since history began,
only a handful of the great, the good and the celebrated are still alive to us.
What propels some into Who’s Who and others into obscurity?

Joseph Leidy isn’t one of history’s darlings. Few, in fact, will have heard
of him. Yet according to biographer Leonard Warren, this 19th-century
Philadelphian was “the last man who knew everything”.

The further we delve into the facts, the odder it seems that this
extraordinary polymath has sunk from sight. In his concise, lucid biography
Warren credits Leidy with being, among other things, the founder of American
vertebrate palaeontology; an international authority on anthropology,
entomology, botany, zoology, mineralogy, anatomy; the greatest authority on
microscopes in his country, one of the first to pioneer experimental cancer
research methods and an expert on gems consulted by the British Museum.

So why was Leidy edited out of history? Warren suggests several reasons, but
they all boil down to this: Leidy was not connected to the right networks. In
his lifetime, American scientific endeavour was all but ignored by Europeans.
Worse, Leidy’s approach—detailed naturalistic observation—was
“old-fashioned” compared with the experimental clinical investigation taking
over. It wasn’t timidity that kept him from the cutting-edge theoretical debates
of his time, but, sadly, his working-class roots: he felt ill at ease with the
wealthy gentleman-scientists of Philadelphia, argues Warren. Yet while the man
disappeared, his knowledge wasn’t lost. It is as if Leidy’s mass of data was a
remote, fertile island, and the papers he wrote seeds that drifted to distant
shores, where they generated disciplines whose roots have since been
forgotten.

A fanciful metaphor? Not at all: in Information Ecologies, Bonnie
Nardi and Vicki O’Day argue that the most productive way to look at people,
organisations and technologies is indeed as “information ecologies”. As in
natural ecologies its elements have complex interrelationships and exhibit
diversity, evolve and co-evolve, and have their own keystone species.

Nardi and O’Day would argue that metaphors don’t just describe, but actually
shape, our relationship with technology. If we view technology as a tool we see
ourselves as controlling it. If we treat it as a “text”—as do critical
theorists like Bruno Latour—we see technology as embodying aspects of our
culture which can be “read” in different ways at different times. If we view
technology as a system, say Nardi and O’Day, we see ourselves as caught up
inside it.

Their prescriptions for evolving a thriving information ecology—”work
from core values”, “pay attention” and “ask strategic questions”—may be
redolent of management manuals. But they are surely right to say that the
ecological view has value in encouraging us to play an active part in our
information ecologies rather than simply passively using whatever is provided
for us—especially on the Internet.

The latter, of course, replaces “knowing everything” with “knowing how to
connect”. Nardi and O’Day argue that the openness of the Internet is threatened
by powerful commercial interests. And openness seems vital for thriving
information ecologies. But when they call for people to “nurture and defend
local ecologies”, Nardi and O’Day sound like internetworked environmentalists
railing against the tide of “crass commercialism”. Commerce and the creation of
wealth are vital in motivating all human activity, including the Internet. Are
Nardi and O’Day reluctant, one wonders, to stray from their own comfortable
information ecology? And if so, might their fate be similar to Leidy’s?

ALARM bells should ring in your head when you read the next sentence.
“Familiarity with more advanced mathematical concepts (such as vector algebra
and differential calculus) might occasionally help.”It appears in a book on
mathematics that claims to be “written with a large public in mind”.

Like all enthusiasts, mathematician Arturo Sangalli tends to assume that
everyone has the same capacity to engage with the often incomprehensible
language in which such enthusiasms are expressed. To be fair, however, Sangalli
gets as far as page 25 of The Importance of Being Fuzzy without the
appearance of a nonlinear differential equation, 24 pages more than I had
expected.

His topic is fuzzy logic, a way of representing knowledge which has become
increasingly important in making computer and other artificial systems function
more like humans. As its name suggests, fuzzy logic uses the imprecise
possibilities of language because they capture the loosely-defined categories
and generalised strategies that humans need to “learn”.

If you are prepared to search them out, the book does have gems. Sangelli
succeeds in giving the reader an idea of the possibilities, for example, when he
describes the commercial applications of fuzzy logic. You learn about the cruise
control system of the 1994 Mazda Sentia, with fuzzy controllers to prevent
upward gear-shifting when travelling uphill; that the 1988 Volkswagen Beetle
uses fuzzy controllers to create an adaptive automatic transmission; and that
there were more than 200 domestic appliances using fuzziness of one kind or
another by the end of 1990. As to the fundamentals of fuzzy logic, however, I
remain, well, fuzzy. Sangalli manages to convince the reader unschooled in
mathematics and computing that it is All Very Hard Indeed and should be Left To
The Experts.

THIS is the best book on computers I have ever read. And I write that as
someone who has struggled through interminable computer science textbooks in the
search for something that students might find even remotely engaging,
interesting or at least small enough not to need a small backpack to carry.

After all that, Daniel Hillis’s The Pattern on the Stone is a breath
of fresh air. Hillis knows what he is talking about: he began his career at MIT.
He became co-founder and chief scientist of Thinking Machines (makers of the
Connection Machine, one of the first and speediest parallel processing
computers) and most recently vice-president and Disney Fellow at Walt Disney
Imagineering.

But for me Hillis’s most impressive achievement has to be that, as a small
child, he built a mechanical man out of motors and lights, without seriously
injuring himself. His lifelong interest in computers started with the
realisation that what was missing from his robot was the ability to think or
compute.

Recalling the standpoint of a small child captivated by possibilities,
complexities and challenges of building a fully-functioning robot, Hillis has
written “the book I wish I had read when I first started learning about the
field of computing”.

In 150 or so short pages, Hillis takes us on a lightning tour of the
fundamentals of computing, beginning with Boolean logic, finite-state machines
and programming languages, through functional abstraction and recursion, to
Turing machines, quantum computing, public-key encryption, parallel computing,
neural networks and the simulated evolution of intelligence. Nowhere does Hillis
lose sight of the fact that what is important is not the detail of these issues,
but the story that flows through them and the rationality of thought that
connects them.

His other achievement is not to pretend that he is setting down the “truth”
about his field. Scattered throughout are Hillis’s reflections, opinions and
statements of personal interest. When discussing parallel architectures, for
example, you’re getting the story straight from the source. Hillis arrived at
MIT’s artificial intelligence laboratory as an undergraduate in 1974 when
the field was in a stage of explosive growth.

“The first programs that could follow simple instructions in plain English
were being developed, and a computer that understood human speech seemed just
around the corner,” he recalls. Several years later, Hillis rejoined the lab as
a graduate student at a time when the intractability of some AI problems was
becoming apparent. As he explains: “Although lots of new principles and powerful
methods had been invented, applying them to larger, more complicated problems
didn’t seem to work.” He was one of the first advocates of the parallel computer
architecture that solved a few of those problems in the late 1970s.

Parallel computers, argues Hillis, are a powerful solution to the inherent
limitations of standard, serial computer architecture, where one thing is done
at a time like a production line, with each worker performing a specific task.
Parallel computers, by contrast, do many things at once, like a team of people
working on one task together, and so have the potential to be very fast indeed.
But the idea of using computers in parallel—linking many
processors—seemed beset by problems. “I spent a large part of my early
career arguing with people who believed it was impractical, or even impossible,
to build and program general-purpose massively parallel computers,” Hillis
recalls.

One of those people was Gene Amdahl, who in the 1960s proposed the law that
says, essentially, that there will always be a part of any computer operation
which is sequential: that is, it can only be done one step at a time. Even if 90
per cent of a task is possible using a parallel approach, the final 10 per cent
will form a bottleneck. Hillis’s insight was that the sequential part of a task
was much smaller than 10 per cent. By dividing the task between the various
processors in a parallel computer you could avoid the bottleneck altogether.
Hillis says his approach is based on the way the human brain combines many
neurons working in parallel. “Because I knew that the brain was able to get fast
performance from slow components, I also knew that Amdahl’s law does not also
˛ą±č±č±ô˛â”.

Hillis also makes no bones about the fact that he is firmly of the belief
that human intelligence is largely a matter of the computational properties of
our 10¹¹ self-organising neurons and their 10¹⁴ connections, even if we don’t
yet really understand the way the brain works in any more than a sketchy
outline. He responds to the critics of the computational view by pointing out
that “Most of us do not appreciate being likened to machines. This is
understandable: we ought to be insulted to be likened to stupid machines, such
as toasters and automobiles, or even today’s computers.” As he says, “The simple
caricature of thought within today’s computers” has much to teach us.

Hillis also does a good job of rehabilitating some of the more extreme
pronouncements of AI’s pioneers, who have predicted the arrival of the
artificially intelligent computer for a long time. Hillis simply suggests that
such an intelligence may be evolved, and that is why such a long time is
required. But he does not merely assert this. He takes the trouble to identify
the mechanism most likely to achieve the required result: evolutionary
development of a variety of intelligences in successively richer simulated
environments, beginning with computers we would recognise and ending with
something intelligent, evolved but not born and beyond our imagination.

The latter will prove a Pandora’s box, Hillis argues. If we assist the
evolution of artificial intelligence, we will usher in problems of selfhood, of
authority and rights. The arrival of an artificially intelligent computer
heralds the arrival of a whole tangle of moral issues to confront. What, he
asks, will we do if, for example, we are faced with the dilemma of pulling the
plug on an artificially intelligent being?

And there’s plenty more to come, much of it beyond our imaginings. The
Internet, for example, is an ocean of e-mail at the moment. What, he wonders,
will happen when the Net connects physical devices such as the computers in
telephone networks to home appliances. For Hillis, the direct communication
between them will provide a rich environment, a source of “emergent behaviour
going beyond any that has been explicitly programmed into the system”.

It’s this mixture of anecdote, insight and challenging ideas that marks this
book out from the herd. The Pattern on the Stone will prove excellent
reading for many an audience, from undergraduates to readers who just want to
know what all the fuss is about. I plan to donate my copy to my 76-year-old
father along with his new Apple Macintosh and first Net account. I hope that he
sends Daniel Hillis an e-mail and tells him he enjoyed the book.

TECHNO-SNOOPING has come a long way since the good old days of steaming open
envelopes and applying a wineglass to the wall. It now includes CCTV
surveillance cameras (300 000 on the streets of Britain alone), nannycams, sound
pickup networks to monitor the streets for gunfire, enhanced computer vision
technologies, and flying robot cameras. What will such technologies do to the
delicate balance between an individual’s right to privacy and the desire of
society to monitor its citizens’ activities? Will our every action be visible to
an all-controlling state—and to anyone who wishes to sell us shampoo or
life insurance?

Trying to curb the reach of surveillance technologies by law is not the
correct response, according to David Brin in The Transparent Society.
The right course, he says, is to embrace the concept of transparency, and allow
surveillance technologies to proliferate. If we can all observe each other, then
we can all be held to account for our behaviour, he argues. In Brin’s
transparent society, it won’t be the police alone who are responsible for
monitoring the outputs of video cameras and identifying wrongdoers. Every
citizen will be able to dial up any camera and monitor everyone
else—including the police. Transparency brings communal
accountability.

The problem with this argument is that it is hard to see how the transparency
movement will ever get started. Overwhelmingly, surveillance technologies are
owned by the state or rich corporations. And when it comes down to it, most
people actually don’t care to think too much about what their fellow
citizens—except those few with serious criminal intent—are up
to.

For these reasons and more, Brin’s call for a transparent society is likely
to fall on deaf ears. There are serious privacy issues that he could have
explored more fully: British proposals for the registration of sexual offenders,
the Clipper Chip proposals for regulating what American children see on TV and
various legislative attempts to ensure freedom of information.

Perhaps this is not surprising as his book comes from the same stable as
cybercritiques such as Clifford Stoll’s Silicon Snake Oil (Review, 13
January 1996, p 43) and The War of Desire and Technology by Rosanne
Stone. All share with Brin a lack of reasoned and balanced argument and patchy
coverage of key issues.

These facts and a lot more can be found in The Future of Software, edited by Derek Leebaert, researcher in the management of technology at Georgetown University, who brings together contributions from 14 scientists and strategists. In an earlier hook, The Future of Computing and Communications (MIT Press, 1991), Arthur C. Clarke’s preface gave the various chapters the air of a rarefied science fiction collection (Clarke, of course, envisioned communications satellites – and wrote a story describing how the Soviet Union used them to beam pornography into the US). In contrast, The Future of Software is long on facts but a little short on adventurous prediction. Apart from the those above, there are detailed histories and discussions of, among other things, the development of time-sharing systems, user interfaces, client/server computing, software standards, and business process re-engineering.

As well as the history of software, the papers range from the future of software from the perspectives of standards development, through computer-supported collaboration, to end-user programming, knowledge representation and knowledge work.

Leebaert suggests that there are two key themes rupning through this collection. First, the increasing ability of software to collaborate. Contributors examine the ways in which future software will be “without borders” (as exemplified in open systems, plug-and-play, ATM technology, workflow systems and multimedia publishing, among others). Secondly, the way in which it is all too easy to take things for granted when discussing computing technology. Authors predict the “fall of software’s aristocracy” – computer scientists and trained programmers – in favour of an end user with intelligent visual programming tools, the development of intelligent agent-based software capable of exercising “discretion”, and the increasing use of groupware by the “knowledge organisation”.

All these developments are just around the corner (we have rudimentary intelligent agent systems, there are several groupware products and visual programming tools abound), but this book provides an interesting and at times lively restatement of the trends.

My favourite chapter is written by Louis and Morrow, and takes the form of a science fiction short story illustrating the future of workgroup computing. Characterising the period before 1997 (and the “LastEverComdexHurrah!”) as the end of the “middle ages of computerdom” during which people were victims of centralised communications systems and hierarchical organisations, “The Prairie School” illustrates the future of software in a way more engaging than many of the other contributions. The chapter is a series of vignettes showing how intelligent communications systems would operate in a different paradigm to current broadcast systems. Since their chapter is a story, and one doesn’t like to spoil the endings of stories, I suggest you read the book and find out if the cowboys and farmers become friends.

The difference between the atoms of Negroponte’s laptop and the bits of information it contained is the point of Negroponte’s book: how information, recorded and transmitted in digital form, is the valued commodity, and how people are increasingly enabled by communications technology to become members of the “digerati”, shipping bits of information around a seamless global medium.

Negroponte is director of the Media Laboratory at the Massachusetts Institute of Technology, Boston. Funded mainly by the private sector, his organisation has carried out some of the most fundamental work in computing and communications technologies – ranging from work on the Spatial Data Management System that laid the groundwork for the desktop concept and the icons of personal computers, to the real-time, full-colour video holograms which make 3D images appear to float in space. To its credit are some of the most important developments in user-interface design, including intelligent, speech-based and gestural interfaces.

This book describes many of these developments and more, and is partly based on Negroponte’s published columns (composed and delivered by e-mail, of course) which appear in Wired, the cult American magazine. Negroponte discusses developments that will change the face of information and communications in the next millennium: “bitcasting” via optical fibre cables and satellites, digital television, computer graphics, computer-generated music and computer-enhanced learning.

Negroponte predicts the increasing globalisation of work into a seamless, digital whole, with software constructed by moving work-in-progress around the digital medium to coincide with international time zones, where people are released from the limitation of geography as the basis for friendship, collaboration and “neighbourhood”, and where each new generation becomes more digital than the last.

It is telling that one of Negroponte’s fundamental themes is the future of television. The current industry preoccupation with developing “set-top boxes” that allow consumers to receive more television channels or movies on demand, is based on the premise that the set-top box is a kind of gatekeeper, allowing, among other things, for providers to charge for bits of information as they flow into the television. Rather, suggests Negroponte, the difference between the television and the computer is becoming less obvious as computers are already being used to receive TV programmes, telephone and news services. Negroponte points out that the problem with the television is that it is the dumbest appliance in your home – even the microwave contains more computer chips.

The impact of digital technology on television, he suggests, is not about watching the O. J. Simpson trial in greater resolution, or about receiving 1000 channels (999 of which by definition you are not watching), but about downloading bits which can be used later. Television will become a random-access medium that pulls bits from the bit stream – just like a computer attached to a network.

Negroponte devotes much space to discussing the mechanics of “bit allocation” in broadcasting – how the Federal Communications Commission and other regulatory bodies in the US are allocating parts of the broadcast spectrum to carry information. Such regulation is impossible in a digital world, he argues. Providers will eventually allocate bits to a particular medium, encoding the bit stream to tell your computer “here comes some television, here comes The Wall Street Journal, here comes radio”.

In the distant future, such allocation will be done not by the transmitter but by the receiver, the smart computer, which uses local intelligence to transform the bitstream into news, telephone calls, television or spreadsheets. A wider effect of this, he says, is that monolithic media empires are already dissolving, as the economic leverage of owning newspaper print plants is being replaced by cottage bit industries.

How soon is Negroponte’s vision likely to become a reality? If you compare the hype surrounding the superhighway with experience of actually surfing the Net, you might be doubtful. The Internet is currently little more than a collection of useless information that is difficult to access without expensive and cumbersome communications technology.

But the development of increasingly sophisticated, intelligent software will allow consumers to select and filter information more effectively. There can be no doubt that the digital landscape is changing quickly: I am writing this review in a Chicago hotel room and will e-mail it to my office via a CompuServe account, accessed on the hotel phone line, and across the Internet; over the last week I have sent and received over 200 e-mail messages from Nashville, Boston and Chicago, and will write another 50 on the flight home to England for transmission back at my office hangin’ out with the digerati.