The Internetchanging the way we

communicate

rom a sprawling web of computer

networks, the Internet has spread

throughout the United States and

abroad. With funding from NSF and

other agencies, the Net has now

become a fundamental resource in

the fields of science, engineering,

and education, as well as a dynamic

part of our daily lives.

F

To u nde r s t a nd t h e c hange s brought by the Internet, one has only to go

back a few decades, to a time when researchers networked in person, and collaboration meant

working with colleagues in offices down the hall. From the beginning, computer networks were

expected to expand the reach—and grasp—of researchers, providing more access to computer

resources and easier transfer of information. NSF made this expectation a reality. Beginning by

funding a network linking computer science departments, NSF moved on to the development

of a high-speed backbone, called NSFNET, to connect five NSF-supported supercomputers.

More recently NSF has funded a new backbone, and is playing a major role in evolving the

Internet for both scientists and the public. The Internet now connects millions of users in the

United States and other countries. It has become the underpinning for a vibrant commercial

enterprise, as well as a strong scientific tool, and NSF continues to fund research to promote

high-performance networking for scientific research and education.

The Internet — 7

A Constellation of Opportunities• A solar wind blasts across Earth’s magnetic

field, creating ripples of energy that jostle satel-lites and disrupt electrical systems. Satellitedata about the storm are downlinked throughGoddard Space Flight Center in Greenbelt,Maryland, passed on to a supercomputer cen-ter, and uploaded by NSF-funded physicists atthe University of Maryland and at DartmouthCollege in Hanover, New Hampshire. Using theInternet, researchers work from their ownoffices, jointly creating computer images ofthese events, which will lead to better space-weather forecasting systems.

• An NSF-funded anthropologist at Penn StateUniversity uses his Internet connection towade through oceans of information. Finally,he chooses the EMBL-Genbank in Bethesda,Maryland, and quickly searches through hugeamounts of data for the newly deciphered DNAsequence of a gene he’s studying. He finds it,highlights the information, downloads it, andlogs off.

• It’s time to adjust the space science equip-ment in Greenland. First, specialized radar is pointed at an auroral arc. Then an all-skycamera is turned on. The physicist controllingthe equipment is part of a worldwide team ofresearchers working on NSF’s Upper Atmos-pheric Research Collaboratory (UARC). Whenshe’s finished making the adjustments, thephysicist pushes back from her computer inher Ann Arbor, Michigan, office, knowing thejob is done.

• The Laminated Object Manufacturing (LOM)machine’s camera, or LOMcam, puts a new picture on the Web every forty-five seconds, andthe molecular biologist watches. From his office,he can already tell that the tele-manufacturingsystem is creating an accurate physical modelof the virus he is studying. The San DiegoSupercomputer Center’s LOMcam continues topost pictures, but the biologist stops watching,knowing that he will soon handle and examinethe physical rendering of the virus, and learnmore about it than his computer screen imagecould ever reveal.

This is the Internet at work in the lives of sci-entists around the globe. “The Internet hasn’t onlychanged how we do science, it permits entirely newavenues of research that could not have been con-templated just a few years ago,” says GeorgeStrawn, executive officer of NSF's Directorate forComputer and Information Science and Engineering(CISE) and former division director for networkingwithin CISE. “For example, the key to capitalizingon biologists’ success in decoding the humangenome is to use Internet-based data enginesthat can quickly manipulate even the most mas-sive data sets.”

A Public NetThe Internet, with its millions of connections world-wide, is indeed changing the way science is done.It is making research more collaborative, makingmore data available, and producing more results,faster. The Internet also offers new ways of dis-playing results, such as virtual reality systems thatcan be accessed from just about anywhere. The newaccess to both computer power and collaborativescientists allows researchers to answer questionsthey could only think about a few years ago.

Working from one of the NSF-supported

supercomputing centers, researcher

Greg Foss used the Internet to collect

weather data from many sites in

Oklahoma. He used the data to create

this three-dimensional model of a

thunderstorm.

8 — National Science Foundation

It is not just the scientists who are enthralled.While not yet as ubiquitous as the television or aspervasive as the telephone, in the last twentyyears, the Internet has climbed out of the obscu-rity of being a mere “researcher's tool” to therealm of a medium for the masses. In March 2000,an estimated 304 million people around the world(including nearly 137 million users in the UnitedStates and Canada) had access to the Internet, upfrom 3 million estimated users in 1994. U.S.households with access to the Internet increasedfrom 2 percent in 1994 to 26 percent in 1998,according to the National Science Board’s (NSB)Science and Engineering Indicators 2000. (Everytwo years, the NSB—NSF’s governing body—reports to the President on the status of scienceand engineering.)

In today’s world, people use the Internet tocommunicate. In fact, for many, email has replacedtelephone and fax. The popularity of email lies inits convenience. No more games of telephone tag,no more staying late to wait for a phone call. Emailallows for untethered connectivity.

The emergence of the World Wide Web hashelped the Internet become commonplace inoffices and homes. Consumers can shop for goodsvia the Web from virtually every retail sector, frombooks and CDs to cars and even houses. Banksand investment firms use the Web to offer theirclients instant account reports as well as mech-anisms for electronic financial interactions. In 1999,the U.S. Census Bureau began collecting informa-tion on e-commerce, which it defined as onlinesales by retail establishments. For the last threemonths of 1999, the bureau reported nearly

$5.2 billion in e-commerce sales (accounting for0.63 percent of total sales), and nearly $5.3 billionfor the first quarter of 2000. More and more,people are going “online to shop, learn aboutdifferent products and providers, search for jobs,manage finances, obtain health information andscan their hometown newspapers,” according to arecent Commerce Department report on the digitaleconomy. The surge of new Internet businesspromises to continue, with some experts estimating$1.3 trillion in e-commerce activity by 2003.

Among the Web’s advantages is the fact that itis open twenty-four hours a day. Where else canyou go at four in the morning to get a preview ofthe upcoming art exhibit at the New York Museumof Modern Art . . . fact sheets on how to preparefor an earthquake or hurricane . . . a study on theimportant dates of the Revolutionary War . . . orhands-on physics experiments for elementaryschool teachers?

With all of this information, the Internet has thepotential to be a democratizing force. The sameinformation is readily available to senior and juniorresearchers, to elementary school students, andto CEOs. Anyone who knows how and has theequipment can download the information, examineit, and put it to use.

While the original Internet developers may nothave envisioned all of its broad-reaching capabili-ties, they did see the Internet as a way of sharingresources, including people, equipment, andinformation. To that end, their work has succeededbeyond belief. The original backbone rates of 56 kbps (kilobytes per second) are now availablein many homes.

In March 1997, NSF launched itsPartnerships for Advanced Compu-tational Infrastructure (PACI) program.“This new program will enable theUnited States to stay at the leadingedge of computational science, producing the best science and engineering in all fields,” said PaulYoung, former head of NSF’s Directoratefor Computer and Information Scienceand Engineering.

The program consists of the NationalComputational Science Alliance (theAlliance), led by the University of Illinoisat Urbana-Champaign, and the NationalPartnership for Advanced ComputationalInfrastructure (NPACI) led by the SanDiego Supercomputer Center. Thepartnerships offer a breadth of visionbeyond even what NSF has hoped for.They will maintain the country’s leadin computational science, further theuse of computers in all disciplines ofresearch, and offer new educationalopportunities for people ranging fromkindergartners through Ph.D.s.

The Alliance’s vision is to createa distributed environment as a proto-type for a national information infrastructure that would enable thebest computational research in thecountry. It is organized into fourmajor groups:

• Application Technologies Teams thatdrive technology development;

• Enabling Technologies Teams that convert computer science researchinto usable tools and infrastructure.

• Regional Partners with advancedand mid-level computing resources

that help distribute the technologyto sites throughout the U.S.

• Education, Outreach, and TrainingTeams that will educate and pro-mote the use of the technology tovarious sectors of society.

In addition, the leading-edge site atthe University of Illinois at Urbana-Champaign supports a variety of high-end machines and architectures enablinghigh-end computation for scientistsand engineers across the country.

NPACI includes a national-scalemetacomputing environment withdiverse hardware and several high-endsites. It supports the computationalneeds of high-end scientists and engineers across the country via avariety of leading-edge machines andarchitectures at the University ofCalifornia at San Diego. It also fostersthe transfer of technologies and toolsdeveloped by applications and computerscientists for use by these high-endusers. Another major focus includesdata-intensive computing, digitallibraries, and large data set manipulationacross many disciplines in both engineering and the social sciences.

NSF recently announced an awardto the Pittsburgh SupercomputingCenter to build a system that willoperate at speeds well beyond a trillioncalculations per second. The TerascaleComputing System is expected tobegin operation in early 2001, whenPittsburgh will become PACI's latestleading-edge site. Through thesepartnerships, PACI looks to a strongfuture in computational science.

PACI: Computer Partnerships

10 — National Science Foundation

From the outset, NSF limited the amount of timeit would support CSNET. By 1986, the network wasto be self-supporting. This was a risky decision,because in 1981 the value of network services wasnot widely understood. The policy, which carriedforward into subsequent NSF networking efforts,required bidders to think about commercializationfrom the very start. When the 1986 deadlinearrived, more than 165 university, industrial, andgovernment computer research groups belongedto CSNET. Usage charges plus membership feesranged from $2,000 for small computer sciencedepartments to $30,000 for larger industrial mem-bers. With membership came customer support.

The Launch of NSFNETWhile CSNET was growing in the early 1980s, NSFbegan funding improvements in the academiccomputing infrastructure. Providing access to com-puters with increasing speed became essential forcertain kinds of research. NSF’s supercomputingprogram, launched in 1984, was designed to makehigh per formance computers accessible toresearchers around the country.

The first stage was to fund the purchase ofsupercomputer access at Purdue University, theUniversity of Minnesota, Boeing Computer Services,AT&T Bell Laboratories, Colorado State University,and Digital Productions. In 1985, four new super-computer centers were established with NSFsupport—the John von Neumann Center atPrinceton University, the San Diego Supercomputer

From Modest BeginningsIn the 1970s, the sharing of expensive computingresources, such as mainframes, was causing abottleneck in the development of new computerscience technology, so engineers developed net-working as a way of sharing resources.

The original networking was limited to a fewsystems, including the university system that linkedterminals with time-sharing computers, early busi-ness systems for applications such as airlinereservations, and the Department of Defense’sARPANET. Begun by the Defense AdvancedResearch Projects Agency (DARPA) in 1969 as anexperiment in resource-sharing, ARPANET providedpowerful (high-bandwidth) communications linksbetween major computational resources andcomputer users in academic, industrial, and gov-ernment research laboratories.

Inspired by ARPANET’s success, the CoordinatedExperimental Research Program of the ComputerScience Section of NSF’s Mathematical andPhysical Sciences Directorate started its own net-work in 1981. Called CSNET (Computer ScienceNetwork), the system provided Internet services,including electronic mail and connections toARPANET. While CSNET itself was just a startingpoint, it served well. “Its most important contri-bution was to bring together the U.S. computerscience community and to create the environmentthat fostered the Internet,” explains LarryLandweber, a professor at the University ofWisconsin and a CSNET principal investigator. In addition, CSNET was responsible for the firstInternet gateways between the United States andmany countries in Europe and Asia.

Satellites, the Hubble Space Telescope,

and observatories around the world

provide data to Earth-bound scientists.

Once received, the data are sent, via

the Internet, to researchers around

the country. At the University of

Illinois-based National Center for

Supercomputing Applications, Frank

Summers of Princeton University

used the data to create a model of

a galaxy formation.

The Internet — 11

Center on the campus of the University of Californiaat San Diego, the National Center for Super-computing Applications at the University of Illinois,and the Cornell Theory Center, a production andexperimental supercomputer center. NSF laterestablished the Pittsburgh Supercomputing Center,which was run jointly by Westinghouse, Carnegie-Mellon University, and the University of Pittsburgh.

In 1989, funding for four of the centers, SanDiego, Urbana-Champaign, Cornell, and Pittsburgh,was renewed. In 1997, NSF restructured thesupercomputer centers program and funded thesupercomputer site partnerships based in SanDiego and Urbana-Champaign.

A fundamental part of the supercomputing ini-tiative was the creation of NSFNET. NSF envisioneda general high-speed network, moving data morethan twenty-five times the speed of CSNET, andconnecting existing regional networks, which NSFhad created, and local academic networks. NSFwanted to create an “inter-net,” a “network ofnetworks,” connected to DARPA’s own internet,which included the ARPANET. It would offer usersthe ability to access remote computing resourcesfrom within their own local computing environment.

NSFNET got off to a relatively modest start in 1986 with connections among the five NSF university-based supercomputer centers. Yet itsconnection with ARPANET immediately put NSFNETinto the major leagues as far as networking wasconcerned. As with CSNET, NSF decided not torestrict NSFNET to supercomputer researchersbut to open it to all academic users. The other wide-area networks (all government-owned) supported mere handfuls of specialized contrac-tors and researchers.

The flow of traffic on NSFNET was so great inthe first year that an upgrade was required. NSFissued a solicitation calling for an upgrade and,

equally important, the participation of the privatesector. Steve Wolff, then program director forNSFNET, explained why commercial interests eventually had to become a part of the network,and why NSF supported it.

“It had to come,” says Wolff, “because it was obvious that if it didn’t come in a coordinated way,it would come in a haphazard way, and the academiccommunity would remain aloof, on the margin.That’s the wrong model—multiple networks again,rather than a single Internet. There had to becommercial activity to help support networking,to help build volume on the network. That wouldget the cost down for everybody, including theacademic community, which is what NSF wassupposed to be doing.”

To achieve this goal, Wolff and others framedthe 1987 upgrade solicitation in a way that wouldenable bidding companies to gain technical expe-rience for the future. The winning proposal camefrom a team including Merit Network, Inc., a con-sortium of Michigan universities, and the state ofMichigan, as well as two commercial companies,IBM and MCI. In addition to overall engineering,management, and operation of the project, theMerit team was responsible for developing user

In 1996, researchers and artists envi-

sioned the information superhighway

this way. Since then, individuals, schools

and universities, organizations, and

government agencies have added

millions of connections to the Internet.

12 — National Science Foundation

support and information services. IBM providedthe hardware and software for the packet-switchingnetwork and network management, while MCIprovided the transmission circuits for the NSFNETbackbone, including reduced tariffs for that service.

Merit Network worked quickly. In July 1988,eight months after the award, the new backbonewas operational. It connected thirteen regionalnetworks and supercomputer centers, representinga total of over 170 constituent campus networksand transmitting 152 million packets of informa-tion per month.

Just as quickly, the supply offered by theupgraded NSFNET caused a surge in demand.Usage increased on the order of 10 percent permonth, a growth rate that has continued to this dayin spite of repeated expansions in data commu-nications capacity. In 1989, Merit Network wasalready planning for the upgrade of the NSFNETbackbone service from T1 (1.5 megabits per sec-ond or Mbps) to T3 (45 Mbps).

“When we first started producing those trafficcharts, they all showed the same thing—up andup and up! You probably could see a hundred ofthese, and the chart was always the same,” saysEllen Hoffman, a member of the Merit team.“Whether it is growth on the Web or growth of traf-fic on the Internet, you didn’t think it would keepdoing that forever, and it did. It just never stopped.”

The T3 upgrade, like the original networkimplementation, deployed new technology underrigorous operating conditions. It also required aheavier responsibility than NSF was prepared toassume. The upgrade, therefore, represented anorganizational as well as a technical milestone—the beginning of the Internet industry.

In 1990 and 1991, the NSFNET team wasrestructured. A not-for-profit entity called AdvancedNetworks and Services continued to providebackbone service as a subcontractor to MeritNetwork, while a for-profit subsidiary was spun offto enable commercial development of the network.

The new T3 service was fully inaugurated in1991, representing a thir tyfold increase in thebandwidth on the backbone. The network linkedsixteen sites and over 3,500 networks. By 1992,over 6,000 networks were connected, one-third ofthem outside the United States. The numberscontinued to climb. In March 1991, the Internetwas transferring 1.3 trillion bytes of informationper month. By the end of 1994, it was transmitting17.8 trillion bytes per month, the equivalent ofelectronically moving the entire contents of theLibrary of Congress every four months.

An End and a BeginningBy 1995, it was clear the Internet was growingdramatically. NSFNET had spurred Internet growthin all kinds of organizations. NSF had spent approx-imately $30 million on NSFNET, complemented byin-kind and other investments by IBM and MCI.As a result, 1995 saw about 100,000 networks—both public and private—in operation around thecountry. On April 30 of that year, NSF decommis-sioned the NSF backbone. The efforts to privatizethe backbone functions had been successful,announced Paul Young, then head of NSF’s CISEDirectorate, and the existing backbone was nolonger necessary.

From there, NSF set its sights even higher. In1993, the Foundation offered a solicitation callingfor a new, very high per formance BackboneNetwork Service (vBNS) to be used exclusively forresearch by selected users. In 1995, Young andhis staff worked out a five-year cooperative agree-

Mosaic: The Original BrowserBy 1992, the Internet had becomethe most popular network linkingresearchers and educators at thepost-secondary level throughout theworld. Researchers at the EuropeanLaboratory for Particle Physics,known by its French acronym, CERN,had developed and implemented theWorld Wide Web, a network-basedhypertext system that let usersembed Internet addresses in theirdocuments. Users could simply clickon these references to connect tothe reference location itself. Soonafter its release, the Web came tothe attention of a programming team at the National Center forSupercomputing Applications(NCSA), an NSF-supported facility at the University of Illinois.

The history of NSF's supercom-puting centers overlapped greatlywith the worldwide rise of the per-sonal computer and workstation. Itwas, therefore, not surprising that

software developers focused on creat-ing easy-to-use software tools fordesktop machines. The NSF centersdeveloped many tools for organizing,locating, and navigating throughinformation, but perhaps the mostspectacular success was the NCSAMosaic, which in less than eighteenmonths after its introduction becamethe Internet “browser of choice” forover a million users, and set off anexponential growth in the number ofdecentralized information providers.Marc Andreessen headed the teamthat developed Mosaic, a graphicalbrowser that allowed programmers to post images, sound, video clips,and multifont text within a hypertextsystem. Mosaic engendered a widerange of commercial developmentsincluding numerous commercial versions of Web browsers, such as Andreessen's Netscape and Microsoft’s Internet Explorer.

14 — National Science Foundation

ment with MCI to offer the vBNS. That agreementwas recently extended to keep the vBNS operatingthrough March 2003. The vBNS has met its goalof pushing transmission speed from its startingpoint of 155 Mbps to speeds in excess of 2.4billion bits per second by the turn of the century.

The vBNS originally linked the two NSF super-computing leading-edge sites that are part of the Foundation’s Partnerships for AdvancedComputational Infrastructure (PACI) program. NSFsoon tied in another thir teen institutions. By2000, the network connected 101 institutions,including 94 of the 177 U.S. universities thathave received high-per formance computingawards from the Foundation.

“In ensuring that vBNS will be available at leastthrough March 2003, NSF is living up to its Next-Generation Internet commitments while chartingthe course for new research applications thatcapitalize on that infrastructure,” says NSF'sStrawn. “The new Information Technology Researchprogram—begun in fiscal year 2000—has spurredan overwhelming response of proposals from theacademic community, which proves that thesetools have become critical to research in scienceand engineering.”

Research on Today’s Internet For researchers, the expanding Internet meansmore—more data, more collaboration, and morecomplex systems of interactions. And while notevery university and research institution ishooked up to the vBNS, all forms of the Internethave brought radical changes to the way researchis conducted.

Ken Weiss is an anthropologist at Penn StateUniversity and an NSF-supported researcherstudying the worldwide genetic variability of humans.While he is not actively seeking a hook-up to the

vBNS, he says the Internet has had a significantimpact on his research. For example, he usesemail constantly, finding it more convenient thanthe phone ever was. And he has access to muchmore data. He can download huge numbers ofgene sequences from around the world and doresearch on specific genes.

Weiss is an active user and enthusiast, but hedoes not necessarily agree that more is alwaysbetter. “The jury is still out on some aspects, suchas the exponential growth of databases, whichmay be outpacing our ability for quality control.Sometimes the data collection serves as a sub-stitute for thought,” says Weiss.

Other disciplines are seeing an equally phenom-enal surge of information. Researchers can nowget many journals online when they once neededto work geographically close to a university library.The surge of data is both a boon and a problemfor researchers trying to keep on top of their fields.But no one is asking to return to the pre-Internetdays, and no one is expecting the informationgrowth to end.

On a more profound level, the Internet ischanging science itself by facilitating broaderstudies. “Instead of special interest groupsfocusing on smaller questions, it allows peopleto look at the big picture,” says Mark Luker, who was responsible for high-performance net-working programs within CISE until 1997 and isnow vice president at EDUCAUSE, a nonprofitorganization concerned with higher educationand information technology.

Routers, sometimes referred to asgateways or switches, are combina-tions of hardware and software thatconvey packets along the right pathsthrough the network, based on theiraddresses and the levels of conges-tion on alternative routes. As withmost Internet hardware and soft-ware, routers were developed andevolved along with packet switchingand inter-network protocols.

NSFNET represented a major new challenge, however, because itconnected such a diverse variety of networks. The person who didmore than anyone else to enablenetworks to talk to each other wasNSF grantee David L. Mills of the

University of Delaware. Mills developedthe Fuzzball software for use onNSFNET, where its success led toever broader use throughout theInternet. The Fuzzball is actually apackage comprising a fast, compactoperating system, support for theDARPA/NSF Internet architecture,and an array of application programsfor network protocol development, testing, and evaluation.

Why the funny name? Mills beganhis work using a primitive version ofthe software that was already knownas the “fuzzball.” Nobody knows whofirst called it that, or why. But every-one appreciates what it does.

Fuzzball: The Innovative Router

16 — National Science Foundation

By “special interest groups,” he means themore traditional, individualistic style of sciencewhere a researcher receives a grant, buys equip-ment, and is the sole author of the results. Thecurrent trend is for multiple investigators to con-duct coordinated research focused on broad phenomena, according to Tom Finholt, an organi-zational psychologist from the University ofMichigan who studies the relationship betweenthe Internet and scientists.

This trend, Finholt and others hasten to add,has existed for a long time, but has been greatlyenhanced by the Internet’s email, Web pages, andelectronic bulletin boards. In addition, formal collaboratories—or virtual laboratories of collabo-rators—are forming around the globe. The SpacePhysics and Aeronomy Research Collaboratory(SPARC) is one of these. Developed as the UpperAtmospheric Research Collaboratory (UARC) inAnn Arbor, Michigan, in 1992 and focused onspace science, this collaboratory has participantsat sites in the United States and Europe. Scientistscan read data from instruments in Greenland,adjust instruments remotely, and “chat” with col-leagues as they simultaneously view the data.

“Often, space scientists have deep but nar-row training,” says Finholt. SPARC allows them to fit specialized perspectives into a bigger pic-ture. “Space scientists now believe they havereached a point where advances in knowledgewill only be produced by integrating informationfrom many specialties.”

Furthermore, the collaboration is no longer asphysically draining as it once was. Now, scientistslike Charles Goodrich of the University of Marylandand John Lyon of Dartmouth College can continuecollaborating on space-weather research, evenwhen one person moves away. While Goodrichadmits that the work might be done more easily if it could be done in person, he is sure that boththe travel budget and his family life would sufferif he tried. “You can put your data on the Web, butnot your child,” he quips.

Expectation for the Internet of TomorrowIf the past is a guide, the Internet is likely tocontinue to grow at a fast and furious pace. Andas it grows, geographic location will count lessand less. The “Information Superhighway” is notonly here, it is already crowded. As Luker says, itis being divided, as are the highways of manycities, allowing for the equivalent of HOV lanesand both local and express routes. The electronichighway now connects schools, businesses, homes,universities, and organizations.

And it provides both researchers and businessleaders with opportunities that seemed like sci-ence fiction no more than a decade ago. Even now,some of these high-tech innovations—includingvirtual reality, computer conferencing, and tele-manufacturing—have already become standardfare in some laboratories.

Tele-manufacturing allows remote researchers to move quickly from computer drawing boards toa physical mock-up. At the San Diego SupercomputerCenter (SDSC), the Laminated Object Manufacturing(LOM) machine turns files into models using eitherplastic or layers of laminated paper. The benefitsare especially pronounced for molecular biologistswho learn how their molecules actually fit together,or dock. Even in a typical computer graphicsdepiction of the molecules, the docking processand other significant details can get lost among

A networked virtual laboratory allows

scientists to work remotely on a design

project. This virtual laboratory and lab

worker were created by specialists at

the University of Illinois at Chicago.

The Internet — 17

the mounds of insignificant data. SDSC’s modelscan better depict this type of information. They arealso relevant to the work of researchers studyingplate tectonics, hurricanes, the San Diego Bayregion, and mathematical surfaces.

To make the move from the vir tual to thephysical, researchers use the network to sendtheir files to SDSC. Tele-manufacturing lead sci-entist Mike Bailey and his colleagues then createa list of three-dimensional triangles that bound thesurface of the object in question. With that infor-mation, the LOM builds a model. Researchers caneven watch their objects take shape. The LOMcamuses the Web to post new pictures every forty-fiveseconds while a model is being produced.

“We made it incredibly easy to use so thatpeople who wouldn’t think about manufacturingare now manufacturing,” says Bailey. For someresearchers, the whole process has become soeasy that “they think of it no differently than youdo when you make a hard copy on your laserprinter,” he adds. SDSC’s remote lab has movedout of the realm of science fiction and into thearea of everyday office equipment.

While other remote applications are not as faralong, their results will be dramatic once the bugsare ironed out, according to Tom DeFanti of theUniversity of Illinois at Chicago and his colleagues.DeFanti and many others are manipulating thecomputer tools that provide multimedia, interac-tion, virtual reality, and other applications. Theresults, he says, will move computers into anotherrealm. DeFanti is one of the main investigators ofI-WAY, or the Information Wide Area Year, ademonstration of computer power and networkingexpertise. For the 1995 Supercomputer Conferencein San Diego, he and his colleagues, Rick Stevensof the Argonne National Laboratory and LarrySmarr of the National Center for SupercomputingApplications, linked more than a dozen of thecountry’s fastest computer centers and visualiza-tion environments.

The computer shows were more than exercisesin pretty pictures; they demonstrated new ways ofdigging deeply into the available data. For example,participants in the Virtual Surgery demonstrationwere able to use the National Medical Library’sVisible Man and pick up a “virtual scalpel” to cut“vir tual flesh.” At another exhibit, a researcherdemonstrated tele-robotics and tele-presence.While projecting a cyber-image of himself into theconference, the researcher worked from a remoteconsole and controlled a robot who interacted withconference attendees.

Applications such as these are just the begin-ning, says DeFanti. Eventually the Internet willmake possible a broader and more in-depth experience than is currently available. “We’retaking the computer from the two-dimensional‘desktop’ metaphor and turning it into a three-dimensional ‘shopping mall’ model of interaction,”he says. “We want people to go into a computerand be able to perform multiple tasks just as theydo at a mall, a museum, or even a university.”

NSF Directorate for Computer and Information Scienceand Engineering www.cise.nsf.gov

National Computational Science Alliance www.access.ncsa.uiuc.edu/index.alliance.html

National Partnership for Advanced ComputationalInfrastructure www.npaci.edu

Pittsburgh Supercomputing Centerwww.psc.edu

vBNS (very high performance Backbone Network Service)www.vbns.net

Defense Advanced Research Projects Agencywww.darpa.mil

25th Anniversary of ARPANETCharles Babbage Institute at University of Minnesotawww.cbi.umn.edu/darpa/darpa.htm

To Learn More