Mar 1, 1998 12:00 PM

Ant Wisdom for the Web

By James Glave

| CRUCIAL TECH

| Ant Wisdom for the Web

Super Taster

Sound Technology

The Great Push Forward

Telecom Goes Qwest

Tiny Transmission

Character Recognition Sheds Its Neurons

Mugspotter

<h4>#### essor Paul Kantor's digital-information pheromones sniff out the good stuff on the Web. But keep your antennae up for intellectual fads and poisoned bait.</h4 Wpeople need a metaphor to describe Web navigation, they usually reach for a spider. Paul Kantor is partial to ants. "They've evolved these chemical systems for communicating information," says the Rutgers University professor of information sciences. "When you look at people dealing with any kind of information system, you realize that each person's decisions – those he or she makes in the course of getting to the right information – are essentially lost to the rest of the world."

Er digital-information pheromones, or DIPs, the concept at the core of a new Rutgers University project that aims to "antify" the Web by allowing people to leave pointers for those who might follow in their footsteps. But dealing with DIPs is no picnic. Kantor's answer is a network of Ant World Servers and AntApplets that will allow searchers to vote on how well a particular page they land on has satisfied their query.

Shat would an ant-enabled site look like? "Our current thinking is that you'd see a tiny ant icon next to a link," explains Kantor. Clicking on the insect would pop up a dialog box describing how useful the Web page behind the link had been to previous visitors.

Tidea dates back to 1987, when Kantor, then a distinguished visiting scholar at the Online Computer Library Center in Dublin, Ohio, sought a way for patrons to leave pointers from one book to another. As the Web grew, so did Kantor's project, until Darpa – realizing how critical information management is to national security – threw US$1 million into the undertaking. Kantor and colleagues Benjamin Melamed and Endre Boros expect their ants to break out of the lab and tunnel onto the Web by 2000.

Auple of nagging problems persist. Aside from the inevitable pheromone abuse by unscrupulous marketers, there exists the troublesome issue of trendmongering.

"ellectual fads are dangerous and wasteful of time," Kantor explains. "If you accumulate a well-worn path to a particular page, how do you get people to another page that has surpassed it in value?" Furthermore, Web searchers themselves are a flighty bunch. Asking them to rate their findings may be tough when online altruism – and attention – is in increasingly short supply. Admits Kantor, "A lot of our success will depend on not being seen as another flashy ad."

<ark Frauenfelder</em

<a hWisdom for the Web</a>S

r Taster

<Transmission</a>

<potter</a>

A hund times more sensitive than current devices, this biosensor detects compounds such as drugs or bacteria at concentrations as low as 9 quadrillionths of a gram per square millimeter. Created at The Scripps Research Institute and the University of California at San Diego, the sensor uses a chip of porous silicon – with an effective surface area of several square feet – to "taste" the biosample.

Beftwich</p

<ref=om for the Web</a> r

dchnology

rr Recognition Sheds Its Neurons</a>

rer</a>

s sodattle-sized device would make a swell hood ornament on Buck Rogers's rocket ship – but it's what is going on inside</the unat's really the stuff of science fiction. Resonating under the shiny shell are sound waves of an astonishing amplitude – more than 1,600 times higher than any made by humans. Put another way, Los Alamos National Laboratory's Gregory Swift says, "If you were able to somehow find yourself inside the small resonating cavity of this device, hearing loss would be the least of your worries. Your hair would catch fire." The

ngy is called resonant macrosonic synthesis (RMS), a revolutionary method of generating and harnessing superhigh-energy sound waves that's finding its first uses in home refrigerators and air conditioners. Unve

t December in San Diego at the 134th meeting of the Acoustical Society of America, RMS is the innovation of Tim Lucas, founder and CEO of MacroSonix in Richmond, Virginia. By e

iting with gas-filled chambers of different cone and bulb shapes, Lucas and his colleagues found it possible to eliminate the pesky shock waves that typically limit the energy levels of sound waves, yielding unprecedented pressure levels. Beyo

orefrigeration, RMS may find uses in process reactors, noncontaminating compressors, and pumps for commercial gases and ultrapure or hazardous fluids (technologies crucial to the semiconductor and pharmaceutical industries). On another front, RMS could be combined with pulse combustion to convert fuel to electric power.

ru Stout <p

oCRUC</strong> </

r"httor the Web</a> <a h"

tsh Forward <a h

"cognition Sheds Its Neurons</a>

"a> <p

nese Se Economic and Trade Commission has unleashed its recipe for jump-starting 166 state-of-the-art technologies during China's ninth Five-Year Plan (1996-2000). The document, which targets China's industrial sectors – including electronics, transportation, power generation, and telecommunications – is yet another quinquennial projection spewed forth by the Chinese government, a practice dating back to the early days of the People's Republic in 1953. But don't mistake the blueprint for just another bureaucratic wish list: it's China's high tech R&D hot sheet. The gove

nopes that the six Chinese firms slated to receive priority investments of 20 million yuan (US$2.4 million) each will rank among the world's 500 largest companies by 2010. The funding is awarded under one condition: the firms must promise to undergo technological development while steadfastly adhering to state regulations. Though k

wto government policy may seem an anachronistic gesture in China's increasingly market-driven economy, the state still calls the shots, according to Wei-chou Su, managing director of the US Information Technology Office in Beijing: "The state regularly designates key R&D projects where it provides the people, the money, and the facilities to boost development. This has been the tradition here, and it works quite well in the Chinese context." Literatu

esed by the Chinese State Economic and Trade Commission also touts the efficacy of R&D "in the Chinese context," claiming that 1,384 new products were developed last year and that sales volume on these products is expected to reach 180 billion yuan ($21 billion). Deng Xiaoping would be pleased: China's cat is not only still catching the mouse, it's being spoon-fed multivitamins to do the task even better than before. <h4>An upst

f#### uses the Internet and state-of-the-art fiber laid alongside railroad tracks to offer phone service at half the going rate. By Steve G. Steinberg</h4> Qwest C iions's December announcement of 7.5-cents-per-minute long distance phone service was the opening shot across the bow of the telecom behemoths. It wasn't so much that the Denver telco had undercut the competition by 50 percent, it was that it was using voice-over-IP (VOIP) technology to do so. Protesti

hthe technology just isn't ready yet, AT&T, Sprint, and even WorldCom have taken a cautionary position on the idea of unifying data and voice over a single network. Qwest, on the other hand, went out and did it. While the big guys were captive to their aging networks built for voice, Qwest took advantage of the fact that its network was designed for data. "What's

gis a revolution in telecom," says Joseph Nacchio, Qwest's CEO and a former AT&T exec. "It's going to be as dramatic as the shift from the telegraph to the telephone," echoes Nayal Shafei, Qwest VP and a graduate of the MIT Media Lab. "We aren't a telco, we're a multimedia carrier." Tired rh

iperhaps. But the fact that a 6-year-old start-up led by an unlikely team of old telephony hands, computer scientists, and construction experts now has a market cap of US$6 billion and its competition on the run is reason enough to listen. And once you learn what lies behind the 7.5-cent solution, it's hard not to believe that Qwest is right. The Qwes

obegins with Philip Anschutz. A billionaire who made his money from oil and railroads, Anschutz bought Southern Pacific Railroad for $1.8 billion in 1988 and sold it eight years later to Union Pacific for $5.4 billion. But the real coup was that he kept the rights-of-way that run parallel to Southern Pacific's tracks. These narrow strips of real estate became the basis for Qwest's network, providing a home for 13,000 miles of fiber-optic cable strung underground across the US. Buried a

sthe train tracks are now two conduits. The first contains state-of-the art cable, called nonzero dispersion-shifted fiber, that can carry far more data than the older fibers laid by companies like Sprint and AT&T during the 1980s. The second conduit lies empty, giving Qwest an open track to lay next-generation technology quickly. Combine Qwest's fiber with the latest in data transmission and you have a network far ahead of the competition. "Our network can't be duplicated by the other carriers," says Nacchio. "It would be like trying to refurbish a 10-year-old PC with a new processor and hard drive rather than buying a brand-new one. It just doesn't make economic sense." The fast

k, straight-out-of-New Jersey Nacchio should know. He was the head of AT&T's consumer business. Nacchio maintains that the telephone companies aren't stupid, they're just hamstrung by their shareholders. "The tel

rize that a revolution is occurring, but what are they going to do?" he asks. "If they say they are going to cut their margins by 50 percent to compete in this new competitive environment, their stock will drop 30 points." Qwest is

tto those old margin structures. The firm's Internet-savvy technology and massive network capacity allow it to achieve with 7.5-cent rates the same margins the other carriers get on tolls twice as high. For the most part, these savings come from lower equipment costs. Instead of using switches from Lucent or Nortel that cost tens of millions of dollars, Qwest uses Cisco routers priced at maybe a million. The savi

acome from the inherent advantages of packet switching over circuit switching. Instead of tying up an entire phone line's capacity no matter how much is actually being sent, IP sends packets only when there are packets to send. Most amazing of all, none of this savings requires a trade-off in voice quality. This is

pof the plan. Right now, no one is using VOIP, so bandwidth isn't a problem. Still, while Qwest will likely be able to meet the technical challenges, the company is untested when it comes to functions like customer support. So what'

xor Qwest? Shafei says that the company will offer data services like virtual private networks and concurrent engineering, where engineers collaborate over the network using high-bandwidth CAD images. Maybe ca

gest a multimedia carrier isn't as hackneyed as it sounds. Steve G. Steirg (<a href="http://org</a>) is a</Wiibuting_tor and a consultant for a New York investment firm. Portfolio managers he consults for may represent the companies mentioned. <p_

r*lder <p*

CIAL rong> <h

"p://he Web</a> <a href=p

sn <a href=

pition Sheds Its Neurons</a> <a h

pp>

lem wiors the size of a grain of sand is their commensurately puny power output. The solution? A micro gearbox – the one shown here was developed at Sandia National Laboratories – that increases the torque (and proportionally reduces the rpms) of a micromotor. This technology opens up a wide range of applications, from satellites to surgical instruments. B

J*/i> |*

ITECH>  

parcheb</a> <a href="htta

in Sheds Its Neurons <a href="htt

aHuma

a hardh time detecting their own handwriting – imagine trying to make software smart enough to comprehend the penmanship of every sloppy writer on the planet, and you see the challenge optical character-recognition software developers have had for the past 30 years. But the s

Molis company Silicon Biology believes that it has a far more accurate OCR program than its competitors, which rely on technology the firm considers fundamentally flawed. Dubbed Fermat, Silicon Biology's program uses a preclassification system based on a genetic algorithm akin to natural selection. In contrast, other OCR programs use a neural network based on the theories of the late Russian mathematician Andrey Kolmogorov. The neural model studies the shape and slope of handwriting in determining content, while Fermat assesses the approximately 20,000 ways a human could write a letter of the alphabet or a number. But does

eave other OCR programs beat? Yes, says Tony McKinley, a consultant with Pennsylvania-based Intelligent Imaging, who tested Fermat against 50 competitors. "It's not 100 percent accurate, but it outperformed other OCR systems by a factor of 50 percent or better." After a s

se to get the firm off the ground, Silicon Biology founder Eric Anderholm and his staff of 30 have begun to carve out a slice of the US$15 billion form-processing industry, attracting a handful of clients, HMOs and insurance companies among them. But data forms may not be the only area the company applies its expertise. CEO Doug Johnson says that the technology can also be applied to classifying spoken words, Asian-language characters, and white blood cells (a process now performed by the naked eye and a microscope). By Da

o*/i> | <st*

T</stp> | <at

aive.a> <a href="http://i

ieds Its Neurons</a> Mugspotter</

iiminals be

: Perstter may be watching. This new face-recognition software, based on biological vision, can spot a moving mug and know in seconds whether it's on the most-wanted short list. Even a mustache or sunglasses can't fool the system. "Our goal is

eious than face recognition," says Person Spotter's codeveloper, University of Southern California professor Hartmut Neven, who with his colleagues at USC and Germany's Ruhr University-Bochum founded the company Eyematic Interfaces to bring their system to market. Indeed, the p

sced enough to recognize individuals walking into a room and will soon be able to determine whether the visitors are grinning or scowling. The process works by first identifying a person's location in an image, then analyzing color and motion cues, and finally extracting the outlines of features at a fine scale, which it compares with patterns in the database. Person Spotte

trcial task is to control access to sensitive internal areas in offices at Germany's Deutsche Bank. Neven also visualizes airports using the system to combat terrorism by tracking passengers and their luggage.