INFORMATION TECHNOLOGY

KEY CONCEPTS
Corporate applications are well under way,
A wide variety of online
Semantic Web applications are and consumer uses are emerging
emerging, from Vodafone Live!s
mobile phone service to Boeings

S
system for coordinating the work
of vendors. ix years ago in this magazine, Tim Ber- that would make this vision come true: a com-
ners-Lee, James Hendler and Ora Las- mon language for representing data that could
Scientic researchers are devel-
sila unveiled a nascent vision of the be understood by all kinds of software agents;
oping some of the most advanced
Semantic Web: a highly interconnected network ontologies sets of statements that translate
applications, including a system
that pinpoints genetic causes of of data that could be easily accessed and under- information from disparate databases into
heart disease and another system stood by any desktop or handheld machine. common terms; and rules that allow software
that reveals the early stages of in- They painted a future of intelligent software agents to reason about the information de-
uenza outbreaks. agents that would head out on the World Wide scribed in those terms. The data format, ontol-
Web and automatically book ights and hotels ogies and reasoning software would operate
Companies and universities,
working through the World Wide for our trips, update our medical records and like one big application on the World Wide
Web Consortium, are developing give us a single, customized answer to a partic- Web, analyzing all the raw data stored in online
standards that are making the Se- ular question without our having to search for databases as well as all the data about the text,
SLIM FILMS

mantic Web more accessible and information or pore through results. images, video and communications the Web
easy to use. The Editors They also presented the young technologies contained. Like the Web itself, the Semantic

90 SCIENTIFIC AMERICAN December 2007

20 07 SCIENTIFIC AMERIC AN, INC.
BY :: Lee Feigenbaum, Ivan Herman, Tonya Hongsermeier,
Eric Neumann and Susie Stephens

Web would grow in a grassroots fashion, only Web to enhance business-to-business interac-
this time aided by working groups within the tions and to build the hidden data-processing
World Wide Web Consortium, which helps to structures, or back ends, behind new consumer
advance the global medium. services. And like an iceberg, the tip of this
Since then skeptics have said the Semantic large body of work is emerging in direct con- semantic web
Web would be too difcult for people to under- sumer applications, too. [si-man-tik web]

stand or exploit. Not so. The enabling technol- noun
ogies have come of age. A vibrant community Just below the Surface A set of formats and lan-
of early adopters has agreed on standards that The Semantic Web is not different from the guages that nd and ana-
have steadily made the Semantic Web practical World Wide Web. It is an enhancement that gives lyze data on the World
to use. Large companies have major projects the Web far greater utility. It comes to life when Wide Web, allowing con-
sumers and businesses to
under way that will greatly improve the ef- people immersed in a certain eld or vocation,
understand all kinds of use-
ciencies of in-house operations and of scientic whether it be genetic research or hip-hop music, ful online information.
research. Other rms are using the Semantic agree on common schemes for representing

w w w. S c i A m . c o m SCIENTIFIC AMERICAN 91
20 07 SCIENTIFIC AMERIC AN, INC.
 [CONSUMER APPLICATIONS]
information they care about. As more groups

Combining Concepts develop these taxonomies, Semantic Web tools

allow them to link their schemes and translate
Search engines on the World Wide Web cannot provide a single answer to a broad- their terms, gradually expanding the number of
ranging question such as Which television sitcoms are set in New York City? people and communities whose Web software
But a new Semantic Web engine called pediax can, by analyzing different concepts can understand one another automatically.
(top, in approximated form) found on Wikipedias seven million online pages. Perhaps the most visible examples, though
Pediax, which grew from the DBpedia project to extract information from Wiki- limited in scope, are the tagging systems that
pedia, provides a clean result (bottom) that merges text and images. have ourished on the Web. These systems in-
clude del.icio.us, Digg and the DOI system used
Sitcoms set in NYC Find by publishers, as well as the sets of custom tags
available on social sites such as MySpace and
Concept 1: Concept 2: Concept 3: Flickr. In these schemes, people select common
terms to describe information they nd or post
Sitcoms Set In NYC
on certain Web sites. Those efforts, in turn, en-
able Web programs and browsers to nd and
crudely understand the tagged information

1
such as nding all Flickr photographs of sunris-
es and sunsets taken along the coast of the Pa-
cic Ocean. Yet the tags within one system do
not work on the other, even when the same term,
such as expensive, is used. As a result, these
systems cannot scale up to analyze all the infor-
mation on the Web.

2 The World Wide Web Consortium an ad
hoc organization of more than 400 companies
and universities co-hosted by the Massachu-
setts Institute of Technology in the U.S., the Eu-

3 ropean Consortium for Informatics and Math-
ematics in France, and Keio University in Ja-

LUCY READING-IKKANDA ( illustration); AGILE KNOWLEDGE ENGINEERING AND SEMANTIC WEB ( screen shot); COURTESY OF APPLE, INC. ( laptop)
pan has already released the Semantic Web

4 languages and technologies needed to cross
such boundaries, and large companies are ex-
ploiting them. For example, British Telecom has
built a prototype online service to help its many
vendors more effectively develop new products
together. Boeing is exploring the technologies
to more efciently integrate the work of part-
ners involved in airplane design. Chevron is ex-
perimenting with ways to manage the life cycle
of power plants and oil reneries. MITRE Cor-

1
poration is applying Semantic Web tool kits to
help the U.S. military interpret rules of engage-
ment for convoy movements. The U.K.s nation-

2 al mapping agency, Ordnance Survey, uses the
Semantic Web internally to more accurately and
inexpensively generate geographic maps.

3
Other companies are improving the back-end

4 operations of consumer services. Vodafone
Live!, a multimedia portal for accessing ring
tones, games and mobile applications, is built on
Semantic Web formats that enable subscribers
to download content to their phones
much faster than before. Harpers
Magazine has harnessed semantic

92 SCIENTIFIC AMERICAN December 2007

20 07 SCIENTIFIC AMERIC AN, INC.
 ontologies on its Web site to present annotated
timelines of current events that are automatical-
ly linked to articles about concepts related to FRIEND OF A FRIEND
those events. Joost, which is putting television Users of a grassroots, semantic social net-
on the Web for free, is using Semantic Web soft- work system Friend of a Friend have cre-
ware to manage the schedules and program ated a vocabulary that describes the person-
guides that viewers use online. al information they want to post and nds common interests. The network
Consumers are also beginning to use the data (logo shown) can also integrate information from isolated, commercial
language and ontologies directly. One example systems such as MySpace and Facebook. See www.foaf-project.org
is the Friend of a Friend (FOAF) project, a de-
centralized social-networking system that is
growing in a purely grassroots way. Enthusiasts Nokia are promoting open-source Semantic Web
have created a Semantic Web vocabulary for de- frameworks common tools for crafting pol-
scribing peoples names, ages, locations, jobs ished programs. Oracles agship commercial
and relationships to one another and for nding database, 10g, used by thousands of corpora-
common interests among them. FOAF users can tions worldwide, already supports RDF, and the
post information and imagery in any format upgrade, 11g, adds further Semantic Web tech-
they like and still seamlessly connect it all, which nology. The latest versions of Adobes popular
MySpace and Facebook cannot do because their graphics programs such as Photoshop use the
elds are incompatible and not open to transla- same technologies to manage photographs and
tion. More than one million individuals have al- illustrations. Smaller vendors among them
ready interlinked their FOAF les, including us- Aduna Software, Altova, @semantics, Talis,
ers of LiveJournal and TypePad, two popular OpenLink Software, TopQuadrant and Soft-
Weblog services. ware AG offer Semantic Web database pro-
As these examples show, people are moving grams and ontology editors that are akin to the
toward building a Semantic Web where rela- HTML browsers and editors that facilitated the
tions can be established among any online piec- Webs vibrant growth. Semantic Web sites can
es of information, whether an item is a docu- now be built with virtually all of todays major
ment, photograph, tag, nancial transaction, computer programming languages, including
experiment result or abstract concept. The data Java, Perl and C++.
language, called Resource Description Frame- We are still nding our way toward the grand
work (RDF), names each item, and the relations vision of agents automating the mundane tasks
among the items, in a way that allows comput- of our daily lives. But some of the most ad-
ers and software to automatically interchange vanced progress is taking place in the life scienc-
the information. Additional power comes from es and health care elds. Researchers in these
ontologies and other technologies that create, disciplines face tremendous data-integration
query, classify and reason about those relations challenges at almost every stage of their work.
[see box on page 95]. Case studies of real systems built by these pio- [THE AUTHORS]
The Semantic Web thus permits workers in neers show how powerful the Semantic Web All ve authors have participated
different organizations to use their own data la- can be. in various projects to develop
bels instead of trying to agree industry-wide on Semantic Web technologies.
one rigid set; it understands that term X in da- Case Study 1: Drug Discovery Lee Feigenbaum, formerly at IBM,
is vice president of technology and
tabase 1 is the same as term Y in database 2. The traditional model for medicinal drugs is
standards at Cambridge Semantics,
What is more, if any term in database 1 changes, that one size ts all. Have high blood pressure? Inc. Ivan Herman leads the Seman-
the other databases and the data-integration Take atenolol. Have anxiety? Take Valium. But tic Web Activity initiative at the
process itself will still understand the new infor- because each person has a unique set of genes World Wide Web Consortium.
mation and update themselves automatically. Fi- and lives in a particular physical and emotional Tonya Hongsermeier is corporate
manager of clinical knowledge
nally, the Semantic Web enables the deployment environment, certain individuals will respond
management and decision support
of reasoners software programs that can dis- better than others. Today, however, a greater at Partners Healthcare System.
cover relations among data sources. understanding of biology and drug activity is Eric Neumann is executive direc-
Just as the HTML and XML languages have beginning to be combined with tools that could
WWW.FOAF-PROJECT.ORG

tor of Clinical Semantics Group

made the original Web robust, the RDF lan- predict which drugs and what doses will Consulting. Susie Stephens was
principal product manager at
guage and the various ontologies based on it are work for a given individual. Such predictions
Oracle Corporation and has recent-
maturing, and vendors are building applications should make custom-tailored, or personalized, ly become principal research scien-
based on them. IBM, Hewlett-Packard and medical treatments increasingly possible. tist at Eli Lilly and Company.

w w w. S c i A m . c o m SCIENTIFIC AMERICAN 93
20 07 SCIENTIFIC AMERIC AN, INC.
 Personalized The challenge, of course, is to somehow meld
a bewildering array of data sets: all sorts of his-
four or ve databases for each one, trying to dis-
cern which genes (or the proteins they encode)
medicine will toric and current medical records about each have features most likely to affect the biology of
become person and all sorts of scientic reports on a
number of drugs, drug tests, potential side ef-
the disorder a painstaking task. In the end, in-
vestigators often cannot afford the hours, and
possible only fects and outcomes for other patients. Tradi- the work falters.
when semantics tional database tools cannot handle the com-
plexity, and manual attempts to combine the
The Cincinnati team, which includes a Se-
mantic Web consultant, began by downloading
makes medical databases would be prohibitively expensive. into a workstation the databases that held rele-
databases Just maintaining the data is difcult: each time
new scientic knowledge is incorporated into
vant information but from different origins and
in incompatible formats. These databases in-
smarter and one data source, others linked to it must be re- cluded Gene Ontology (containing data on genes
easier to use. integrated, one by one.
A research team at Cincinnati Childrens
and gene products), MeSH (focused on diseases
and symptoms), Entrez Gene (gene-centered in-
Hospital Medical Center is leveraging semantic formation) and OMIM (human genes and ge-
capabilities to nd the underlying genetic causes netic disorders). The investigators translated the
of cardiovascular diseases. Traditionally, re- formats into RDF and stored the information in
searchers would search for genes that behave a Semantic Web database. They then used Pro-
differently in normal and diseased tissues, as- tg and Jena, freely available Semantic Web
suming that these genes could somehow be in- software from Stanford University and HP Labs,
volved in causing the pathology. This exercise respectively, to integrate the knowledge.
could yield tens or hundreds of suspect genes. The researchers then prioritized the hun-
Researchers would then have to pore through dreds of genes that might be involved with car-

CHANDRA GUDIVADA, ANIL JEGGA, ERIC BARDES, SCOTT TABAR AND BRUCE ARONOW Cincinnati Childrens Hospital Medical Center (screen shot); COURTESY OF APPLE, INC. (monitor)
diac function by applying a ranking algorithm
[ANALYZING DATABASES] somewhat similar to the one Google uses to
rank Web pages of search results. They found
Which Genes Cause Heart Disease? candidate genes that could potentially play a
causative role in dilated cardiomyopathy, a
Hundreds of genes could potentially contribute to heart disease. Researchers at
weakening of the hearts pumping ability. The
Cincinnati Childrens Hospital Medical Center are using Semantic Web tools to nd
team instructed the software to evaluate the
the most likely culprits by analyzing numerous online databases and scientic refer-
ranking information, as well as the genes rela-
ences (left, on screen), revealing possible causative connections (right, on screen).
For example, they have pinpointed suspect genes related to a chromosomal region tions to the characteristics and symptoms of the
linked to dilated cardiomyopathy, a weakening of the hearts pumping ability. condition and similar diseases. The software
identied four genes with a strong connection
to a chromosomal region implicated in dilated
cardiomyopathy. The researchers are now in-
vestigating the effects of these genes mutations
as possible targets for new therapeutic treat-
ments. They are also applying the semantic sys-
tem to other cardiovascular diseases and expect
to realize the same dramatic improvement in ef-
ciency. The system could also be readily ap-
plied to other disease families.
Similarly, senior scientists at Eli Lilly are ap-
plying Semantic Web technologies to devise a
complete picture of the most likely drug targets
for a given disease. Semantic tools are allowing
them to compile numerous incompatible biolog-
ical descriptions into one unied le, greatly ex-
pediting the search for the next breakthrough
drug. Pzer is using Semantic Web technologies
to mesh data sets about protein-protein interac-
tion to reveal obscure correlations that could
help identify promising medications. Research-
ers there are convinced that these technologies

94 SCIENTIFIC AMERICAN December 2007

20 07 SCIENTIFIC AMERIC AN, INC.
 [HOW IT WORKS]
will increase the chance for serendipitous dis-
coveries, accelerate the speed of delivering new
drugs to market and advance the industry as a Making the Semantic Web Tick
S
whole toward personalized medicine. This is everal formats and languages form the building blocks of the Semantic Web. They
where the Semantic Web could help us, says extend similar software technologies that underlie the World Wide Web itself and
Giles Day, head of Pzers Research Technology have been published as standards by the World Wide Web Consortiums Semantic
Center informatics group in Cambridge, Mass. Web Activity initiative.
In each of these cases, the Semantic Web en-
:: RDF FORMAT. The most fundamental building block is Resource Description
hances drug discovery by bringing together vast
Framework (RDF), a format for dening information on the Web. Each piece of data,
and varied data from different places. New con-
and any link that connects two pieces of data, is identied by a unique name called
sumer services are being built in similar fashion. a Universal Resource Identier, or URI. (URLsthe common Web addresses that we
For example, the British rm Garlik uses Se- all use, are special forms of URIs.) In the RDF scheme, two pieces of information, and
mantic Web software to compare previously in- any notation indicating how they are connected, are grouped together into what is
compatible databases to alert subscribers that called a triple. For example, an online reference to the famous television animal
they might be the target of an identity thief. Flipper, a reference to the relationship is a, and a reference to the concept of
Garlik culls disparate personal identity infor- dolphin could be joined in the triple shown below.
mation from across the Web, integrates it using
common vocabularies and rules, and presents < uri for Flipper > < uri for Is A> < uri for Dolphin >
subscribers with a clear (and sometimes surpris-
ing) view of their online identity. URIs can be agreed on by standards organizations or communities or assigned by
individuals. The relation is a is so generally useful, for example, that the consortium
Case Study 2: Health Care has published a standard URI to represent it. The URI http://en.wikipedia.org/
The health care industry confronts an equally wiki/Dolphin could be used by anyone working in RDF to represent the concept of
dense thicket of information. One initiative that dolphin. In this way, different people working with different sets of information
has been deployed since 2004 was developed at can nonetheless share their data about dolphins and television animals. And people
the University of Texas Health Science Center everywhere can merge knowledge bases on large scales.
at Houston to better detect, analyze and respond
to emerging public health problems. The system,
:: ONTOLOGY LANGUAGES. Individuals or groups may want to dene terms
called SAPPHIRE (for situational awareness and data they frequently use, as well as the relations among those items. This set
of denitions is called an ontology. Ontologies can be very complex (with thousands
and preparedness for public health incidences
of terms) or very simple. Web Ontology Language (known as OWL) is one standard
using reasoning engines), integrates a wide
that can be used to dene ontologies so that they are compatible with and can be
range of data from local health care providers,
understood by RDF.
hospitals, environmental protection agencies
and scientic literature. It allows health of- :: INFERENCE ENGINES. Ontologies can be imagined as operating one level
cials to assess the information through different above RDF. Inference engines operate one level above the ontologies. These software
lenses, such as tracking the spread of inuenza programs examine different ontologies to nd new relations among terms and data
or the treatment of HIV cases. in them. For example, an inference engine would examine the three RDF triples below
Every 10 minutes in the greater Houston and deduce that Flipper is a mammal. Finding relations among different sources is
area, SAPPHIRE receives reports on emergency an important step toward revealing the meaning of information.
room cases, descriptions of patients self-re-
ported symptoms, updated electronic health rec- < uri for Flipper > < uri for Is A> < uri for Dolphin >
ords, and clinicians notes from eight hospitals
that account for more than 30 percent of the re-
gions emergency room visits. Semantic technol- < uri for Dolphin > < uri for Subclass Of > < uri for Mammal >
ogies integrate this information into a single
view of current health conditions across the
area. A key feature is an ontology that classies < uri for Flipper > < uri for Is A > < uri for Mammal >
unexplained illnesses that present ulike symp-
toms (fevers, coughs and sore throats) as poten- :: OTHER TECHNOLOGIES. The Web consortium is crafting inference
tial inuenza cases and automatically reports engines as well as many other technologies. One is SPARQL, a query language that
them to the Centers for Disease Control and allows applications to search for specic information within RDF data. Another is
Prevention. By automatically generating reports, GRDDL, which allows people to publish data in their traditional formats, such as
SAPPHIRE has relieved nine nurses from doing HTML or XML, and species how these data can be translated into RDF. For more,
such work manually, so they are available for see www.w3.org/2001/sw
active nursing. And it delivers reports two to

w w w. S c i A m . c o m SCIENTIFIC AMERICAN 95
20 07 SCIENTIFIC AMERIC AN, INC.
 If two three days faster than before. The CDC is now
helping local health departments nationwide to
Health and Human Services epidemiologists in
the eld. SAPPHIRE succeeded in identifying
databases implement similar systems, replacing tedious, gastrointestinal, respiratory and conjunctivitis
joined by the inconsistent and decades-old paper schemes.
The nimbleness of Semantic Web technolo-
outbreaks in survivors of the disaster much
sooner than would have been possible before.
Semantic Web gies allows SAPPHIRE to operate effectively in SAPPHIREs exibility showcases an impor-
have different other contexts as well. When Hurricane Katrina
evacuees poured into Houstons shelters, public
tant lesson about Semantic Web systems: once
they are congured for a general problem in this
privacy criteria, health ofcials quickly became concerned about case, public health reporting they can quickly
the software the possible spread of disease. Within eight
hours after the shelters were opened, personnel
be adapted to a variety of situations within that
eld. Indeed, the CDC would like to roll out a
will have to at the University of Texas Health Science Center single, integrated, SAPPHIRE-style illness alert
honor both configured SAPPHIRE to help. They armed
public health ofcials with small handheld com-
system nationwide.
SAPPHIRE succeeds because it can unify in-
sets of rules. puters loaded with health questionnaires. The formation from many places, which can then be
responses from evacuees were then uploaded to used for different goals. This same attribute is fu-
the system, which integrated them with data eling FOAFs grassroots growth. By using an
from the shelters emergency clinics and surveil- agreed-on Semantic Web vocabulary, the FOAF
lance reports from Houston Department of system nds common interests among friends

COURTESY OF THE CENTER FOR BIOSECURITY AND PUBLIC HEALTH INFORMATICS RESEARCH, UNIVERSITY OF TEXAS HEALTH SCIENCE CENTER AT HOUSTON ( screen shot) ; COURTESY OF DELL, INC. ( laptop)
and acquaintances, even if they do not belong to
[UNIFYING INFORMATION] the same social-networking sites such as MySpace
or Facebook. FOAF enthusiasts are also now de-
Is a Flu Outbreak Under Way? veloping semantic trust networks white lists of
trusted senders as a way to ght e-mail spam.
Public health ofcials take longer than they would like to recognize new disease out-
breaks, because they must manually compare disparate reports in incompatible for-
mats from many hospitals and doctors ofces. Researchers at the University of Texas
Crossing Boundaries
The success of SAPPHIRE and other applica-
Health Science Center have built a Semantic Web system that quickly and automatical-
ly tracks and analyzes these online data across the Houston area. It presents ofcials tions has prompted calls for more Semantic
with clear trends, such as the incidence of u symptoms across different age groups Web integration in health care. The Food and
over time (center, on screen); a sharp rise would indicate early signs of outbreaks. Drug Administration and the National Insti-
tutes of Health have both recently declared that
a shift toward research into translating data
across boundaries is necessary for improving
the drug development and delivery process.
The same work will enhance the traditional
computerized clinical decision support (CDS)
systems that medical professionals use knowl-
edge bases that contain the latest wisdom on
therapeutic treatments. Each hospital, physi-
cians network and insurance company has had
to custom-design its own system, and all of
them are struggling mightily to stay current.
Every time an advance is made about diagnoses,
clinical procedures or drug safetywhich is of-
ten administrators must rework their systems.
The personnel time required is usually far great-
er than most of these organizations can afford.
Furthermore, because the custom systems are
frequently incompatible, making industry-wide
insights or deciphering best practices is slow
and cumbersome. What is more, we are
investigating Semantic Web technol-
ogies because traditional approach-
es for data integration, knowledge
management and decision support

96 SCIENTIFIC AMERICAN December 2007

20 07 SCIENTIFIC AMERIC AN, INC.
 will not scale to what is needed for personal-
ized medicine, says John Glaser, chief infor-
mation ofcer at Partners HealthCare System BLOG ANALYZER
in Boston. Oracle Technology Network has demonstrated a Semantic Web site that can analyze
To remedy this situation, Agfa HealthCare blogs, podcasts and discussion groups to nd related commentary about specic
has constructed a prototype CDS system based topics. It also can produce visualizations of its ndings, such as tag clouds (below)
that show whose blogs are drawing the most trafc (larger names) and bar charts
on Semantic Web technologies. When a person
that identify the most concentrated discussions. Project details are available at
inputs a change into one part of a system, rec- http://otnsemanticweb.oracle.com
ords that should be altered in other parts of the Abhinav Agarwal Aditya Agarkar Alejandro Varg
system or in the systems of another institution Christopher Jones Clemens Utsching David Allen Didier Laura D
are automatically updated. For example, Agfas Burns)
nominated for a Best Picture AcademyGaurav
Award Verma Hari Jake jean-pierre dijcks Jonath
prototype transforms standard radiology proto- before 1990 that ran longer than three hours.
Katheryn. potterf Kris Rice mark Mark Rittman m
cols into Semantic Web notation and integrates As applications develop, they will dovetail
them with other common knowledge bases, such
[email protected] ([email protected]) St
with research at the Web consortium and else-
Blogs Pankaj Chandriramani Pat Shuff Phil Hunt Ramakumar Me
as clinical guidelines from medical societies. In- where aimed at fullling the Semantic Web vi-
Shmeltzer Stellan Tim Dexter Robert Smyth...
stitutions can maintain their own internally sion. Reaching agreement on standards can be
standardized content, yet end users such as hos- slow, and some skeptics wonder if a big compa-
pitals can readily integrate new content, greatly ny could overtake this work by promoting a set
reducing the labor hours required. of proprietary semantic protocols and browsers.
As systems such as Agfas are implemented Perhaps. But note that numerous companies
across the health care network, medical knowl- and universities are involved in the consortiums
edge bases will become smarter, easier and less semantic working groups. They realize that if
expensive to use. Imagine a patient who is prone these groups can devise a few well-designed
to blood clots and has a genetic mutation that, protocols that support the broadest Semantic
according to current medical literature, will re- Web possible, there will be more room in the fu-
spond well to a new anticlotting medication. ture for any company to make money from it.
Over the ensuing months, however, new studies Some observers also worry that peoples pri-
show that particular variants of this mutation vacy could become compromised as more data
actually cause that same drug to increase clot- about them from disparate sources is interlinked.
ting. This patients clinician must be notied to But Semantic Web advocates argue that the pro-
change the therapy for anyone with this variant. tections are the same as those used in the non-
How could notications such as this be effective- linked world. If two databases joined by the Se-
ly carried out given that thousands of genes are mantic Web have different privacy criteria, then
involved in hundreds of diseases across millions the software will have to honor both sets of rules MORE TO
of patients? Meeting this challenge will not be or create a set that covers both. When SAPPHIRE EXPLORE
possible without robust semantic approaches. joins patient databases, it adheres to the privacy
The Semantic Web. Tim Berners-
requirements of both or it wont proceed; the
Lee, James Hendler and Ora Lassila
Daily Life, Too nurses who had formerly performed the same in Scientic American, Vol. 284,
The same Semantic Web technologies that are mergers manually imposed the same practice. No. 5, pages 3443; May 2001.
transforming drug discovery and health care The Semantic Web will probably operate more
are being applied to more general situations. behind the scenes than the World Wide Web Books about the Semantic Web are
One example is Science Commons, which helps does. We wont see how it helps Eli Lilly create described at http://esw.w3.org/
topic/SwBooks
researchers openly post data on the Web. The personalized drugs; well just buy them. We
nonprot organization provides Semantic Web wont know how Vodafone makes cool ring tones Case studies of how companies and
tools for attaching legally binding copyright so readily available, but well appreciate how research groups are applying the
and licensing information to those data. This easy they are to download. And yet, soon enough Semantic Web can be found at
www.w3.org/2001/sw/sweo/
capability allows a scientist, for example, to the Semantic Web will give more direct power to
public/UseCases
instruct a software applet to go nd informa- us, too, allowing us to go on eBay and not just say
tion about a particular gene but only informa- nd me the Toyota Priuses for sale but nd me Guides to RDF are indexed at
tion that comes with a free license. only used, red Priuses for sale for less than http://planetrdf.com/guide, and
DBpedia is an effort to smartly link informa- $14,000 by people who are within 80 miles of my tools to develop Semantic Web
pages are available at http://esw.
COURTESY OF ORACLE

tion within Wikipedias seven million articles. house and make them an offer. Grand visions
w3.org/topic/SemanticWebTools
This project will allow Web surfers to perform rarely progress exactly as planned, but the Se-
detailed searches of Wikipedias content that are mantic Web is indeed emerging and is making Related blogs and RSS feeds can be
impossible today, such as, Find me all the lms online information more useful than ever. g accessed at http://planetrdf.com

w w w. S c i A m . c o m SCIENTIFIC AMERICAN 97
20 07 SCIENTIFIC AMERIC AN, INC.