COMPUTING RESEARCH NEWSA Publication of the Computing Research Association

September 2003 Vol. 15/No. 4

Title IX and Women in Academics

This fall the athletic fields ofAmerica’s elementary and secondaryschools, colleges and universities willresound with the voices of girls andyoung women who choose to includesports as part and parcel of theireducational experience. Those girlsand young women will not only betaking physical exercise; they’ll beexercising their rights to equalopportunity under a law known asTitle IX.

Title IX states a simple principle.The entire statute reads: “No personin the United States shall on the basisof sex, be excluded from participationin, be denied the benefits of, or besubject to discrimination under anyeducational program or activityreceiving Federal financial assistance.”

Many Americans know theenforcement of that common-senserule has brought women much closerto parity in high school and collegesports opportunities. But in my view,what Title IX has achieved on theplaying field remains undone in theclassroom, where the promise of thislaw was originally directed. Par-ticularly, I believe that Title IX hasyet to be applied stringently enough

in traditionally male-dominatedfields such as the hard sciences, mathand engineering—disciplines whereour nation needs competent workersnow more than ever before.

We can all agree that fairnessimplores us to create and enforceequal opportunity for women inmath, science and technology. Thatis a compelling argument in itself,but it is not the only argument. Areport from the Hart-RudmanCommission on National Security to2025 warned that America’s failureto invest in science and to reformmath and science education was thesecond biggest threat to our nationalsecurity, greater than that from anyconceivable conventional war.

America will not remain thepower it is in the world today, norwill our people be as healthy, aseducated, or as prosperous as theyshould be, if we do not lead theworld in scientific research andengineering development. To makeour country better, to improve ournational security and quality of life,we need to encourage people to gointo these disciplines. Women repre-sent a largely untapped resource inachieving this vital goal. Encourage-ment through Title IX is more than

the right thing to do; it is the smartthing to do.

The numbers reveal a strikinginequity when it comes to genderrepresentation in the math, scienceand technology fields. A NationalScience Foundation study found thatwomen accounted for only 23 per-cent of physical scientists and 10percent of engineers. The percent-ages of women on faculties in theseareas are even lower, with 14 percentof science faculty members beingwomen and a mere 6 percent inengineering departments. Moreover,the numbers are getting worse insome areas. The percentage of recipi-ents of computer and informationsciences bachelor’s degrees who werewomen, which peaked at 37 percentin 1984, had decreased to just 28percent in 1999. That is a movementin the opposite direction from athlet-ics, where Title IX has been ade-quately enforced. Before Title IX, onein 17 girls in school played sports.Now it is one in 2.5. This countryneeds that kind of progress in math,science, and technology. But it willnot happen as long as subtle and not-so-subtle discriminations persist in oureducational institutions.

Studies show that women oftenhave trouble advancing in math, sci-ence and technology due to a lack ofequal access to financial aid and alack of access to child care in gradu-ate school. Additionally, evidenceindicates discrimination toward not

only aspiring students, but towardmembers of university faculties aswell. Women in science and mathoften find themselves pushed intotraditional female roles, such asteaching, while their male counter-parts receive almost all the researchfellowships that pay more completelyfor graduate school. Without aresearch background, women are lesslikely to obtain tenure-track facultypositions, which carry higher pay and prestige.

Just as America’s schools weresent a clear message that they wouldlose Federal funding unless womenwere given parity in sports, it’s timefor our institutions to understandthat there will be consequences ifTitle IX does not become a guiding

By Senator Ron Wyden

Inside CRN

The “productivity paradox” ofmissing organizational payoffs frominvestments in information tech-nology has finally been put to rest.Recent research has demonstrated amajor surge in U.S. productivitybetween 1995 and 2000 due almostentirely to IT. While investment inIT is essential to this improvement,the key to achieving payoffs from ITinvestments lies in changing thenature of work processes to exploitwhat IT offers.

The productivity paradox beganin early 1986 when economistStephen Roach demonstrated thatthe huge increase in organizationalexpenditures on IT (computers,peripheral devices, software, andrelated services) between 1975 and1985 was accompanied by virtuallyno gains in organizational productiv-ity. Within weeks, Fortune magazine’scover story was about “The PunyPayoff” from computers, and the restof the business trade press soon fol-lowed. Nobel Prize winning econo-mist Robert Solow quipped, “We see

computers everywhere except in theproductivity statistics.”

This news did not make sense topeople in the IT fields. Computerscould do many things far faster andfar better than people could. Theirapplication had dramaticallyimproved performance in all kinds oftasks, from payroll processing to airtraffic control. It did not seem possi-ble that such task-level performancewould fail to show up in the produc-tivity statistics. Yet the analyses weregrounded in the best data availableat that time, and the story of theproductivity paradox was establishedas fact.

Several objections to the storyarose immediately. Most were aimedat the problems with the data used inRoach’s analysis. As good as theywere, Roach’s data came from theNational Income and ProductAccounts (NIPA) data maintainedby the U.S. Bureau of EconomicAnalysis. The NIPA system wasinstalled in the 1930s, long beforethe era of modern IT application. Itwas not clear whether the NIPA data measured the right things. In

addition, the NIPA data measuredeffects at the level of whole indus-tries, not at the level of individualfirms where the vaunted task-levelperformance of computers wouldappear. The NIPA data could not beused to account for differences in thequality of IT implementation effortsamong the organizations measured.The successes and the failures mightbalance each other out, with zeroproductivity gain as a result.

A different objection was raisedby economic historians. Theypointed out that flat or even declin-ing productivity was a common fea-ture during transitions from old tonew regimes for doing complicatedthings. Paul David of StanfordUniversity compared the replace-ment of steam engines by electricpower in U.S. factories in the late19th and early 20th centuries to thereplacement of older informationmanagement practices by IT duringthe late 20th century. Productivityremained flat for several decades aselectricity replaced steam, but then

IT Responsible For Most Productivity Gains

NO

NP

RO

FIT

OR

G.

U.S

. PO

STA

GE

PAID

WA

SH

ING

TON

, DC

PE

RM

IT N

O. 9

93

CR

A11

00 S

even

teen

th S

tree

t, N

WS

uite

507

Was

hing

ton,

DC

200

36-4

632

Title IX Continued on Page 8

Productivity GainsContinued on Page 6

Best Practices Memo ........................3Report from Canada ..........................4NSF Update .......................................4

Chair’s Report ....................................5CRA Members List ............................7Professional Opportunities ..............11

Senator Ron Wyden

By John L. King

Guest Column

COMPUTING RESEARCH NEWS September 2003

Page 2

Affiliate Societies

Computing ResearchAssociation

Board OfficersJames FoleyChairGeorgia Institute of Technology

Janice CunyVice ChairUniversity of Oregon

Kathleen McKeownSecretaryColumbia University

Philip A. BernsteinTreasurerMicrosoft Research

Board MembersWilliam AsprayIndiana University

Randal BryantCarnegie Mellon University

Doris CarverLouisiana State University

Lori ClarkeUniversity of Massachusetts

Carla EllisDuke University

Timothy FininUniversity of Maryland,Baltimore County

Barbara GroszHarvard University

James HorningNAI Labs

Leah H. JamiesonPurdue University

Michael JonesMicrosoft Research

John KingUniversity of Michigan

Edward LazowskaUniversity of Washington

Guylaine M. PollockSandia National Laboratories

Daniel ReedUniversity of Illinoisat Urbana-Champaign

Robert SchnabelUniversity of Colorado at Boulder

Marc SnirUniversity of Illinoisat Urbana-Champaign

Lawrence SnyderUniversity of Washington

Mary Lou SoffaUniversity of Pittsburgh

Eugene SpaffordPurdue University

Alfred SpectorIBM Corp.

John StankovicUniversity of Virginia

Frank TompaUniversity of Waterloo

Moshe VardiRice University

Jeffrey VitterPurdue University

David WaltzColumbia University

Richard WatersMitsubishi Electric Research Labs

Elaine WeyukerAT&T Labs - Research

Stuart ZwebenOhio State University

Executive DirectorAndrew Bernat

CACS/AIC

In 2002, CRA sponsored its first“Grand Research Challenges inComputer Science and Engineering.”This was the first in a series of highlynon-traditional conferences wherethe goal is to define important ques-tions rather than expose currentresearch. Grand Challenges meetingsseek “out-of-the-box” thinking toexpose some of the exciting, deepchallenges yet to be met in comput-ing research. Because of its clearimportance and pressing needs,CRA’s second “Grand ResearchChallenges Conference” will bedevoted to defining technical andsocial challenges in informationsecurity and assurance. We are seek-ing scientists, educators, businesspeople, futurists, and others who

have some vision and understandingof the big challenges (and accompa-nying advances) that should shapethe research agenda in this field overthe next few decades. These meet-ings are not structured as traditionalconferences with scheduled presenta-tions, but rather as highly participa-tory sessions exposing importantthemes and ideas. As such, this isnot a conference for security special-ists alone: We seek to convene adiverse group from a variety of fieldsand at all career stages—we seekinsight and vision wherever it mayreside.

The organizing committee ischaired by Eugene Spafford, PurdueUniversity, and co-chaired byRichard DeMillo, Georgia Institute

of Technology. Attendance is limitedto 50 people and is by invitationonly. Individuals invited must com-mit to attending for the entire three-day conference (beginning Sunday at6 pm, ending after lunch onWednesday.) If you are interested inattending, please submit a two-page(or less) statement of two or threeexamples of a “grand research chal-lenge” problem in the IS/IA area.The deadline for submission isSeptember 17, 2003.

For additional details, includinginstructions for submitting state-ments and a list of the organizingcommittee members, seehttp://www.cra.org/

Visionaries Needed for CRA Conference Grand Research Challenges in Information Security & AssuranceAirlie House, Warrenton, VirginiaNovember 16-19, 2003

The Computing Research Asso-ciation is pleased to announce the10th annual CRA OutstandingUndergraduate Awards Program,recognizing undergraduate studentswho show outstanding researchpotential in an area of computingresearch.

Nominees must attend a univer-sity or college located in the UnitedStates or Canada, and must be nomi-nated by the department chair or afaculty member.

A cash prize of $1,000 will beawarded to each of two undergradu-

ate student winners, one female andone male, who are majoring in com-puter science, computer engineering,or an equivalent program. A numberof other outstanding candidates willbe recognized with HonorableMention. The awards will be pre-sented at one of the major comput-ing research conferences sponsoredby CRA, ACM, the IEEE ComputerSociety, SIAM, AAAI, or USENIX.The two first-prize winners willreceive financial assistance towardtheir travel to the conference. CRAencourages home departments to

provide similar assistance to otherstudents who are recognized.

CRA gratefully acknowledges thesupport of Microsoft Research andMitsubishi Electric Research Labswho sponsor the OutstandingUndergraduate Awards Program inalternate years. Mitsubishi ElectricResearch Labs is this year’s sponsor.

Additional information about thenomination procedure and criteriafor selection are available on theCRA website: http://www.cra.org.All nominations must reach CRA byOctober 20, 2003.

CRA Outstanding Undergraduate AwardsDeadline October 20

The CRA Board of Directors hasrecently approved a change in CRA’sprocedures for disseminating theresults of the annual Taulbee Survey.

One reason for the change is toreward departments that submit thesurvey on time by disseminating thepreliminary salary data to them inDecember rather than in the JanuaryCRN. Another is to provide finalsurvey results to CRA members, as abenefit of their membership, beforethey are made publicly available.

Previously, the Taulbee prelimi-nary faculty salary results have beenpublished in the January issue of

Computing Research News, followedby the complete survey results in theMarch edition of CRN.

Beginning with the 2002-03survey, which will be circulated tochairs of Ph.D.-granting departmentsin mid-September 2003, onlydepartments that have submittedtheir surveys by the November dead-line will receive the preliminarysalary results. These results will beprovided to those departments bymid-December; they will no longer bepublished in the January CRN or madepublic at that time.

In mid-February, final TaulbeeSurvey results will be provided todepartments that participated in the survey and to all CRA members.Once again, this is earlier than inthe past. Final results will no longerappear in the March CRN, nor will they be publicly available at thattime. Instead, they will be publishedin the May issue, and will be postedon the CRA website at that time.

Change in Taulbee Survey Reporting

Note toDepartment

Chairs

Taulbee Survey 2002-03

Coming Soon

Richard Tapia Celebration of Diversityin Computing Conference 2003

Co-Sponsored by ACM and CRAOctober 15-18, 2003 in Atlanta, Georgia

http://www.ncsa.uiuc.edu/Conferences/Tapia2003/

September 2003 COMPUTING RESEARCH NEWS

Page 3

Computing Research Association

Best Practices Memo

University-Industry SponsoredResearch Agreements

Universities and businesses have considerable incentive to cooperate in thedevelopment of intellectual property (IP). Businesses recognize universities fortheir rich talent pool and enthusiastic graduate students, while universities rec-ognize businesses as a source of real-world problems, technical know-how, andfunding. There are numerous examples of successful research collaborations incomputer science, computer engineering, and electrical engineering. Mindfulthat some IP such as gene splicing and human growth hormone have produced“IP goldmines,” many university administrators (and some students and fac-ulty) are eager to establish strong safeguards to protect their rights to intellec-tual property.

While such safeguards are perhaps essential in biomedical, pharmaceutical,and agricultural research, they are not appropriate in Information Technology(IT). They can be difficult and time-consuming to negotiate, and because con-siderations such as time-to-market are so important in IT, the delay can frus-trate beneficial cooperation. Moreover, patent safeguards are unnecessarybecause of the role of IP in IT products and the complications involved indeploying IT IP. Formulating university-industry cooperative agreements mustbe sensitive to these issues. This document describes the best practices foruniversity-industry agreements in IT, particularly the IP aspects of suchagreements.

Context and SettingResearch and development in IT-related university departments is funded

largely by two mechanisms: federal grants and university-industry sponsoredresearch agreements (SRAs). Commercializing intellectual property derivedfrom federal grants is (when appropriate) required by the Bayh-Dole Act. Thelaw specifies the conditions of ownership and defines “standard practice” forgrant-receiving institutions. Practices surrounding university-industry SRAs,however, vary widely, being governed mostly by the needs of the agreeingparties. These agreements can take a variety of forms, as explained in the next section.

Research universities typically have two offices, variously named, that areconcerned with funding and intellectual property. The Office of SponsoredProjects (OSP) is generally responsible for negotiating funding agreementswith granting agencies, foundations, and companies. The Office of TechnologyTransfer (OTT) is generally responsible for patenting and licensing technologycreated at the university. In rough terms, the OSP is largely involved beforethe intellectual property is created, and the OTT is involved afterwards. (Asanother generalization, the OSP is typically less familiar with industry’s needsthan is the OTT.) For all research covered by the Bayh-Dole Act, the univer-sity is stipulated as the (initial) owner of the intellectual property. For SRAs,ownership and rights to the intellectual property resulting from research arethe subjects of negotiations prior to funding.

The motivation for establishing best practices guidelines is the potential forthe conflicting interests of universities and industry to impede their negotia-tions. An important “best practice” is for the OSP and the OTT to cooperatein establishing the practices described below.

ExpectationsThe possibility of producing a much-needed revenue stream by licensing

their intellectual property has motivated some administrators, regents, andchancellors to require OSPs to exact strong protection for the university’srights. Patent protection, which is generally required for biomedical, pharma-ceutical, and agricultural IP, is very slow to obtain and can be expensive tosecure and to defend. Almost all patenting expenditures do not recover theirinvestment. As a general rule, universities that successfully generate revenuefrom IP do so with a tiny number (< 10) of significant patents. (In 2001, theUniversity of California system generated 77 percent of all revenue from 25licenses, and none was IT.) There are no known “goldmine” IT licenses.

Managing IT IP using the traditional patent/licensing mechanism is inap-propriate for the following reasons. First, patent protection is rarely the bestform of IT IP protection. Copyright is usually better, since it can be used to

control an embodiment of the ideas, such as a software implementation.Second, time-to-market is often a significant consideration in making a prod-uct a success, so both the university and industry are best served by rapidaction. Third, products like software often contain many “key” ideas (e.g.,algorithms), and it is difficult to assess how any specific idea contributes to theoverall worth of the product, say for the purpose of assessing royalties. Fourth,unlike patents, which are published in enough detail that someone can repro-duce the art, effective transfer of IT IP such as software often requires partici-pation by the creators. Finally, many IT ideas can be implemented by “thoseschooled in the art” once they have seen the technology in operation. Thus,companies have a risk of using IP inadvertently, increasing the value of mecha-nisms that lower that risk.

The implications of these considerations are: a) universities can introducesignificant barriers to cooperation by forcing IT into a standard patent-centricform, and b) agreement principles customized to IT will focus on rapid action.

Consulting and InternshipsThe most valuable part of intellectual property is the intellect that pro-

duced it. Accordingly, IT businesses understand that working with faculty andgraduate students is at least as valuable as licensing the IP that they produce.Because IT requires only modest facilities, and to avoid complex negotiationswith universities about who owns the resulting IP, many firms have opened labsnear universities as a venue for faculty consulting and student internships.Performed on their premises with their equipment and staff, the companiesown all of the IP. It is an efficient scheme for the businesses, and it can provideprofessionally valuable experience for both faculty and students. But it cuts outthe university.

A Model for Sponsored Research Confronted with the aforementioned facts, several universities have

adopted approaches that reflect the best practice. In these cases, an industrialpartner funds research with the understanding that it will receive a non-exclu-sive, non-transferable, worldwide, royalty-free license to any IP created by theorganization. In one model the partner funds (annually) a specific team with aspecific research direction, and the arrangement is seen as ongoing. In anothermodel (e.g., Stanford’s EPIC program) industry partners join a consortium for amodest annual fee, and then have “pay-per-view” privileges for any specific IPat a specified rate for a non-exclusive, royalty-free license. The universityretains ownership of the IP, and the option to negotiate an exclusive license isavailable. Standardized terms and conditions regularize the process for rapidand predictable execution.

Although it is assumed that industry wants and needs exclusive licenses, ingeneral this seems not to be the case. Since companies in IT do their owndevelopment, licenses protect them from being sued for infringing on others’IP. For that purpose a non-exclusive license suffices.

In this model, industry supports the effort with its funds “up front,” with theassurance that there will be a license “if anything useful comes out of theresearch.” Not charging royalties has the advantage that the uses of the IP donot have to be accounted for. There is risk on both sides: It is possible the uni-versity might do better by negotiating more favorable terms for a specific prom-ising technology, and industry is gambling that the investigators’ discoveriesget to them early enough for rapid deployment. However, this model has thevalue of promoting and accelerating cooperation. There is generally an under-standing among the participants that if the IP turns out to be a “home run,”then the company will return later to the question of what it owes the univer-sity for the IP, typically in the form of generous donations or longer-termresearch contracts.

End NotesThe assertions in this memo are documented in “Model Language for Patent

and Licensing Agreements for Industrially Sponsored University Research InInformation Technology,” by J Strother Moore, University of Texas, Austin.See: http://www.cra.org/reports/ip/. Additional information and sample wordingfor agreements are also provided.

Approved by theComputing Research AssociationBoard of DirectorsJuly 2003

Prepared by:J Strother Moore (University of Texas, Austin)Lawrence Snyder (University of Washington) Philip A. Bernstein (Microsoft Research)

Special Report

COMPUTING RESEARCH NEWS September 2003

Page 4

CISE FY2004 Update

Changes in CISEThe CISE Directorate is

approaching the end of its seconddecade as a major organizational unitwithin NSF. CISE is unique amongNSF directorates in its dual responsi-bility for the health of the CISEresearch communities and for thesupport of a national computationand communication infrastructurefor all of NSF’s research and educa-tional disciplines. The past few yearshave seen unprecedented growth inthe CISE budget, a continuing diver-sification of CISE research, educa-tion and infrastructure programs, andan attendant, dramatic increase inproposal pressure and staff workload.Today, CISE is a primary source forthe funding of fundamental academicresearch in computing, communica-tion, and information, and ofresearch at the interface of informa-tion technology and other disci-plines. Since the last CISEreorganization in 1997, the direc-torate’s budget has grown by 113 per-cent (60% for NSF) and the numberof proposals received has grown bymore than 125 percent (16% forNSF).

While the opportunities for ourcommunity have increased with theintroduction of NSF-wide priorityareas, such as the CISE-led Infor-mation Technology Research pro-gram and a growing number ofprogram announcements and solici-tations for traditional programs andareas of special emphasis, successrates have diminished dramatically.We believe investigators haveresponded to these opportunities andthe falling success rates by submit-ting more proposals. This is counter-

productive, as it raises the workloadin CISE and in the research commu-nity. We want the communityinvolved in research and teaching,not proposal-writing, and we want aCISE staff with the time andresources to effectively manage grantselection and oversight.

The proposed new CISE will havefour divisions, working as highlyintegrated units, coordinating tomanage the broad research portfoliofor which CISE is responsible. Thefour divisions are: Computing &Communication Foundations (CCF);Computer and Network Systems(CNS); Information and IntelligentSystems (IIS); and DeployedInfrastructure (DDI). The slides from the announcement of the reor-ganization provide a good introduc-tion and can be found at:http://www.cise.nsf.gov/news/cise_all_hands_meeting_files/frame.htm

The new divisional structure willbe more closely aligned with thedemands of the science. CCF, CNS,and IIS are principally “research”divisions, which represent a naturalprogression from core foundationsthrough systems to basic and focusedresearch in the context of complexinformation and intelligent systems.We believe this organization reflectsa common grouping of the researchcommunities served by CISE, andthat individual investigators will eas-ily find a “home” for their researchinterests. We also recognize the highdegree of interactivity across thesedelineations and will work to ensurea strong, collaborative environment.

DDI represents the next logicalstep in the progression with manage-ment responsibility for nationalcyberinfrastructure activities. This

division merges the national infra-structure programs currently man-aged by Advanced ComputationalInfrastructure and Research (ACIR)and Advanced NetworkingInfrastructure and Research (ANIR),and expands the scope of CISEcyberinfrastructure to include com-putation, communication, informa-tion, distributed sensing,instrumentation, and middleware.DDI will also be responsible for theeducation, outreach, and training(EOT) activities related to thenational cyberinfrastructure. Inchoosing to put all of the planning,design, construction, operation, andupgrading of centers and facilitiesinto one division, we are recognizingthe highly integrated and diverseresources in modern cyberinfrastruc-ture, the need for coordinating theseactivities with all NSF directorates,and the unique management chal-lenges associated with large and dis-tributed facilities. In addition, eachof the other three CISE divisionswill support fundamental and focusedresearch that will enable future gen-erations of cyberinfrastructure.

Within each division there will bea small number of clusters of techni-cal and administrative staff, eachresponsible for a single program or asmall group of programs. Individualprogram managers may be designatedas the point of contact for specificsub-disciplines within a cluster, buteach manager will be part of a team.This will add breadth to the narrowconfines of current programs, andallow investigators to seek supportalone or as collaborators for researchwith a broader scope.

CISE will introduce a series ofthemes in order to address national

priorities and broadly applied CS&Eresearch priorities. Individuals fromeach of the divisions will develop,manage, and integrate the portfolioof projects that address these multi-disciplinary topics. Initial themes areexpected to include cyber trust, sci-ence of design, information integra-tion, and education and workforce.Information Technology Research(ITR) funds for FY2004 will includeCISE investments in these thematicareas.

Cyber TrustCybersecurity is a clear example

of a national priority to which theCISE themes are designed torespond. Congress expects NSF lead-ership on cybersecurity research andeducation. Fiscal year 2003 researchactivities in cybersecurity were han-dled through four different programannouncements. In 2004, these fourprograms will be handled by the fourprogram managers working as a teamunder one program solicitation called“Cyber Trust.” The title reflects ourunderstanding that the public notonly wants their information systemsto be secure, they want to trust themin all kinds of situations. As a simpleexample, they need to be able totrust that data will be kept private.

By centralizing the research incyber trust around a key theme,CISE will ensure that the area ofcybersecurity receives increased, con-certed attention that builds on thesignificant work that has alreadybeen devoted to it. In addition torestructuring the program, we expectto allocate more funds to researchand education in this area, which is

By Sean Jackson, National Science Foundation

This column will serve as an update on some of the changes in CISE as we start a new fiscal and academic year.

NSF Update Continued on Page 10

The Northern Report: Computer Science News fromCanada

Now that the Canadian Asso-ciation of Computer Science/Association informatique canadienne(CACS/AIC) is officially an affili-ated society of the ComputingResearch Association, it seems rea-sonable to provide an annual reporton important issues within theCanadian computer science commu-nity that may also be of interest toCRA members and its other affili-ated societies. This year I would liketo briefly focus on several issues:research initiatives, current enroll-ment trends, accreditation of pro-grams, and the software engineeringdispute in Canada.

Research InitiativesThe research climate for

Canadian computer science (andother science and engineeringdisciplines) continues to improve.

For several years the Canadian fed-eral government has been increasingits investments in research anddevelopment as it tries to positionCanada to succeed in the emergingknowledge economy. At the sametime, many provincial governmentshave invested heavily in computerscience and computer engineering,stimulating considerable growth inmany computer science departmentsin the country.

The result of these investments isa huge growth in demand from com-puter scientists for research funding.In fact, for each of the past 4 years,the number of new computer scienceapplicants for “discovery grants” fromCanada’s Natural Sciences andEngineering Research Council (thegrants that fund professors’ curiosity-driven research programs) has led allNSERC disciplines. In 2003 alonesome 114 of 900 new applicants werecomputer scientists (with many more

computer engineers applying to engi-neering committees).

Fortunately, NSERC has contin-ued to get increased funding to han-dle this influx of “news,” with some$12.5 million (Canadian) in newfunds having recently been allocatedfor first-time applicants over thenext 3 years. The next challenge will be to make sure that even more money enters the system tofund these new applicants as theyprogress through the system, seeking increased renewal grantsdown the road. And a modest start has been made: in a recentNSERC re-allocations exercise thebase budget for funding computerscience was increased in recognitionof the importance of the field and its incredible growth.

NSERC has also added consider-ably more money to its CanadaGraduate Scholarships program, cre-ating some 600 new scholarships

paying $17.5K (Can) for one year toM.Sc. students and $35K (Can) for 3years to Ph.D. students, while theexisting program of postgraduatescholarships continues. In addition,NSERC has created several new pro-grams, including a networks programto encourage “distributed centers ofcritical mass” of scientists at differentinstitutions interacting with oneanother on cutting-edge research,and a special opportunities programfor high-risk (but potentially high-payoff) research.

Beyond NSERC, over the past fewyears the Canadian government hasannounced funding of some 2,000Canada Research Chairs positions(across all academic disciplines) atuniversities across the country. It alsocontinues its Canadian Foundationfor Innovation (CFI) program thathas begun to rebuild university

By Gord McCalla

The Northern Report Continued on Page 9

September 2003 COMPUTING RESEARCH NEWS

Page 5

Past is Prologue: View from the Chair

What can you, our members,expect from CRA? Well, becausepast is indeed prologue, I encourageyou to read our 2003 Annual Report,enclosed with this issue of CRN.

CRA’s biennial Conference atSnowbird 2004 for department chairsand directors of labs and centers isalready being planned by co-chairsMoshe Vardi and Dick Waters andan able committee. The dates areJuly 11-13, 2004. Plenary sessionswill cover such topics as CSE educa-tion apres le crash; the vast and per-suasive impact of IT on the U.S.economy; and issues affecting womenand minorities in CSE.

The ever-popular AcademicCareers Workshop for new facultyand nearly finished Ph.D. students,organized by Lori Clark, is scheduledfor February 23-24, 2004 inWashington, DC. Space fills up very

quickly for this workshop, so advisepotential attendees (graduate stu-dents and junior faculty choosing orbeginning their careers) of the datesnow. Information will be posted athttp://www.cra.org. Encourage newfaculty to use start-up funds toattend; scholarships will be availablefor grad students.

The second CRA Grand ResearchChallenges Conference, onInformation Security and Assurance,will be held November 16-19, 2003.The organizing committee, chairedby Gene Spafford and co-chaired byRich DeMillo, will issue invitationsbased on papers submitted bySeptember 17, 2003 (seehttp://www.cra.org for details). Thisis your opportunity to help shape thefield. We are also seeking ideas for atopic for Grand Research ChallengesIII, currently planned for 2005.

Growth in our academic member-ship slowed to just two net new

members last year. Facing uncertainmembership growth for this year andwithout the surplus generated by aSnowbird conference, the CRAboard adopted a very conservative2003-04 budget at our Februarymeeting. After learning of our situa-tion, two industry members recentlystepped up to the plate and increasedtheir membership levels—Microsoftbecame a Sustaining Member andSun a Supporting Member. Otherindustrial members are planningincreases as well. These acts of cor-porate good citizenship, along withtight cost controls at CRA, haveallowed us to continue all of our pro-grammatic efforts.

Funding for the National ScienceFoundation this year will probablyincrease by about 6 percent—nowhere near what is needed for thebudget-doubling approved last yearby Congress. Improving on this forthe following fiscal year will be a

challenge. This is particularly criticalto continuing progress, given thedecreasing base of industrial research.We will need all of our members tohelp carry this message toWashington.

How can you help CRA achieveour four goals of community build-ing, human resources development,information dissemination, and influ-encing policy? Very simply: suggestadditional ways that CRA canachieve our goals—and volunteer tomake it happen! Become one of the55 non-board members of CRA com-mittees (see list) who help makeCRA work. You can be one of them!Contact me at foley@cc.gatech.eduto volunteer.

Jim Foley, CRA’s board chair, isProfessor and Stephen Fleming Chair inTelecommunications at the GeorgiaInstitute of Technology.

By Jim Foley

SNOWBIRD 2004 ALERT - Department Chairs and Directors of Labs/CentersMark your calendars now for CRA’s Conference at Snowbird 2004!

This biennial event is a “must” for department chairs and directors of labs and centers. The planning committee is putting together astimulating program, including the always-popular workshop for new chairs. The dates are July 11, 12, and 13, 2004 in Snowbird, Utah.Refer to future issues of CRN and the CRA website (http://www.cra.org) for updates, including program details and registration/accommo-dation instructions as they become available.

If you would like to suggest a topic for the program, please contact either of the program co-chairs:

Academic Co-Chair: Moshe Y. Vardi, Rice University (vardi@cs.rice.edu) Labs/Centers Co-Chair: Dick Waters, Mitsubishi Electric Research Labs

(dick@merl.com)

Committee Members:James Aylor (University of Virginia) Mary Lou Soffa (University of Pittsburgh)Lori Clarke (UMass, Amherst) Alfred Spector (IBM Corp.)Scot Drysdale (Dartmouth College) Valerie Taylor (Texas A&M)John King (University of Michigan) Frank Tompa (University of Waterloo)Dan Reed (University of Illinois, UC)

Al Aho, Columbia University

Richard Alo, University of Houston

Nancy Amato, Texas A&M

Sandra Johnson Baylor, IBM

Ron Brachman, AT&T (now at DARPA)

Eric Brittain, MIT

Carla Brodley, Purdue University

Sheila Castaneda, Clarke College

Theresa Chatman, Rice University

Allison Clark, NCSA

Joanne Cohoon, University of Virginia

Anne Condon, University of British Columbia

Carla Ellis, Duke University(board member eff. July 1, 2003)

Gerald Engel, University of Connecticut

Faith Fich, University of Toronto

Joan Francioni, Winona State University

Oscar Garcia, Wright State University

C.W. Gear, NEC Research Institute

Barbara Gutek, University of Michigan

Mary Jane Harrold, Georgia Institute ofTechnology

Juris Hartmanis, Cornell University

Tom Henderson, University of Utah

Raquell Holmes, Boston University

Andrew Hume, AT&T Labs - Reseasrch

John Hurley, Clark Atlanta University

Charles Isbell, Georgia Institute of Technology

Jolene Jesse, AAAS

Sid Karin, UC San Diego

Randy Katz, UC Berkeley

John King, University of Michigan(board member eff. July 1, 2003)

Willis King, University of Houston

Charlotte Kuh, National Research Council

Cynthia Lanius, Rice University

Andrea Lawrence, Spelman College

Phoebe Lenear, NCSA

Clayton Lewis, University of Colorado, Boulder

Ran Libeskind-Hadas, Harvey Mudd College

Monica Martinez-Canales, Sandia National Lab

J Strother Moore, University of Texas, Austin

James Morris, Carnegie Mellon University

David Novick, University of Texas, El Paso

Christos Papadimitriou, UC Berkeley

Lori Pollock, University of Delaware

Ann Redelfs, NPACI/SDSC

Eric Roberts, Stanford University

Gabby Silberman, IBM

Barbara Simons, ACM

Robert Sproull, Sun

Valerie Taylor, Texas A&M

Patricia Teller, University of Texas, El Paso

Henry Walker, Grinnell College

Roger Webb, Georgia Institute of Technology

Pamela Williams, Sandia National Lab

Ellen Yoffa, IBM

Wei Zhao, Texas A&M University

The computing research community thanks the following non-board members who served on CRA committees from July 2002 through June 2003.

COMPUTING RESEARCH NEWS September 2003

Page 6

CRA awarded its 2003 Outstanding Undergraduate Awards at the ACM awards banquet in San Diego inJune. Pictured above are: Front row (l-r) Omar Khan (Cornell University), Outstanding Male Award; JulieThornton (Kansas State University), Female Runner-Up; Bistra Dilkina (Simon Fraser University), OutstandingFemale Award; Noah Snavely (University of Arizona), Male Runner-Up; Chand John (University of Texas atAustin), Finalist. Back row (l-r) Jim Foley, CRA's board chair; Anton Morozov (Hunter College), HonoraryMention; Vijay Reddi (Santa Clara University), Honorary Mention; Phillipe Loher (North Carolina StateUniversity), Honorary Mention; Mahdi Mekic (Lamar University), Honorary Mention; Yuli Ye (University ofWaterloo), Finalist; and Andy Bernat, CRA's executive director.

REGISTER EARLY!

CRA AcademicCareers Workshop

fornew faculty and

advanced graduatestudents

in computing-relateddisciplines

February 23-24, 2004Washington, DC

Details:http://www.cra.org/

accelerated rapidly in the late 1920sas the last of the steam factoriesclosed and the full benefit of electricpower took hold. Electricity provedto be far superior to steam, but thetransition took a long time. The ITrevolution would play out in a simi-lar way, David argued, with produc-tivity rising as the full effect of ITuse was felt.

More than fifteen years havepassed since Roach’s work appeared.IT application never slowed down.In fact, the mid and late 1990sbrought the dot-com era, an extraor-dinary boom in IT application thatcoincided with one of the longestperiods of economic expansion inU.S. history. If productivity were tochange, the late 1990s was the timeto see it. In addition, and partly inresponse to the productivity paradox,economic researchers were develop-ing new measures focusing on thefirm level and incorporating animproved suite of productivity meas-ures. The results of new research onthe late 1990s are coming out, andthey tell a remarkable story.

Productivity grew from 1.33 per-cent to 2.07 percent between theperiods 1975-1993 and 1995-2000,according to Dale Jorgensen ofHarvard University, Mun Ho ofResources for the Future, and KevinStiroh of the Federal Reserve Bankof New York. This is the largest gainin many years. The gain was duemainly to “capital deepening,” whichmeans providing more effective capi-tal investment to leverage the effortsof workers. When IT-capital deepen-ing was separated from other capitaldeepening, the results are even moreimpressive. IT capital deepeningjumped from 0.37 percent to 0.87percent, while other capital deepen-ing actually dropped from 0.43percent to 0.37 percent in the

1995-2000 period. In other words, ITcapital input accelerated while othercapital input decelerated.

Capital deepening is not the onlyfactor that can affect productivity:the quality of the work force alsoplays a major role. During these twotime periods, labor quality actuallydeclined due to the heated economyand the easy availability of jobs asthe less productive workers enteredthe work force. Thus, the majorproductivity increase occurred inspite of a decline in labor qualityduring that period. When productiv-ity gains are separated into IT-relatedproductivity and other-related pro-ductivity, the conclusions becomeobvious. IT-related productivity rosefrom 0.21 percent to 0.45 percent,while other-related productivity overthe same periods grew from only 0.05percent to 0.17 percent. In short,there was a doubling of U.S. produc-tivity in the period 1995-2000,which effectively means that thewealth of the country was buildingtwice as quickly. Nearly all of thatimprovement was due to IT.

The dot-com crash and the slug-gish economy following 2000 put adamper on the celebration. But thisslowdown also provided a neededrespite to step back from the “irra-tional exuberance” of that era toreflect on the changes underway.This is particularly important whenarticles with titles like “IT Doesn’tMatter” appear in Harvard BusinessReview, and academic leaders castnervous glances at noticeabledeclines in applications and enroll-ments in undergraduate computerscience and graduate informationsystems programs. Paul David’shopeful story suggests that IT will con-tinue to bring major improvements inproductivity over the coming years,especially given the continued paceof improvement in the underlyingtechnology itself. The challenge now

lies in better understanding the waysin which IT affects organizationalperformance for the better.

The effort to demonstrate IT’srole in productivity has beenmatched by research into the “hid-den assets” that complement ITinvestments in the quest forproductivity improvement. ErikBrynjolffson of MIT and Lorin Hittof the University of Pennsylvaniadefine hidden assets as those that arenot counted by standard economicmeasurement systems. Many of thesehidden assets have grown up aroundIT implementation, including spe-cialized software and utilities, revisedwork practices, new control systems,and improved analytical capacity toaid in management decision making.The ratio of hidden assets is highestin the most productive firms, reach-ing as much as 10:1. The value cre-ated by IT assets and hidden assetsworking together goes beyond stan-dard economic measurements,affecting things such as customerconvenience and service quality.Previous research not only missedthe productivity impacts of IT, butfailed to recognize the complexity ofthe mechanisms involved in improv-ing productivity.

The story of IT and productivityhas been odd. The productivity pay-offs of IT have been elusive for years,yet organizations kept spending onIT in spite of this news, and dramati-cally accelerated IT investments inthe 1990s. When econometriciansfinally teased out the payoffs of IT,the dot-com boom collapsed and theeconomy turned down. It seems

ironic that it was impossible to seethe evidence of productivity whenthe economy was roaring, but it iseasy now that things have sloweddown. In fact, this is to be expectedbecause understanding almost alwayslags the things to be understood. Ittakes time to understand change ofthis magnitude.

The revolution launched by IT isfar too large and complex to play outin just a few decades. The revolutioninvolves economic, political, socialand cultural systems that oftenchange slowly and in subtle andsometimes ambiguous ways. Theadvances in IT that gave the revolu-tion its power were accomplished byexpertise in narrow areas of special-ization in computer engineering,computer science, and informationsystems. An important role for suchspecialization remains, but it isincreasingly important to look acrossand not just within these specializa-tions, and to join with economists,psychologists, operations manage-ment specialists, library and informa-tion specialists, and others to graspthe full magnitude of the changesunderway. Without this broad effort,many of those who helped to launchthe revolution will see it default tothose with far less sense of how it allhappened, and only a distant notionof what is at stake.

Dr. John L. King (jlking@si.umich.edu)is Dean and Professor in the School ofInformation at the University ofMichigan. He was recently elected to athree-year term on the CRA Board ofDirectors.

CRA Welcomes New MembersAcademic Departments

Loyola University Chicago (CS)University of Montana (CS)

University of South Florida (CSE)

Productivity Gains from Page 1

COMPUTING RESEARCH NEWS September 2003

Special Insert

IntroductionThe Computing Research

Association (CRA) hosted a work-shop in September 2002 to developrecommendations that willstrengthen the research infrastruc-ture in areas of critical importance tonational security. It was funded bythe National Science Foundation.

The workshop focused on threegeneral topics:

1. how to facilitate technologytransfer from research to practice;

2. how to foster research andinfrastructure support for best prac-tices in security and informationfusion; and

3. strategies for funding researchin this area.

Participants were invited from theareas of computer security, real-timesystems, and information fusion, andincluded representatives from acade-mia, industry, and government.Twenty people attended (see list atthe end). The workshop was co-located with two workshops onInformation Technology Research forCritical Infrastructure—hosted byUC Berkeley, Vanderbilt University,and the University of Virginia—inorder to facilitate attendance at allevents and to cross-pollinate ideasfrom different groups of people.

The aims of this workshop onresearch related to national securityare directly relevant to CRA’s mis-sion. CRA seeks to strengthenresearch and advanced education incomputing and allied fields. It doesthis by working to influence policythat impacts computing research,encouraging the development ofhuman resources, contributing to thecohesiveness of the professional com-munity, and collecting and dissemi-nating information about theimportance and the state of comput-ing research. Each plays an impor-tant role in achieving theorganizational objectives for the ben-efit of the country.

In the following sections, we dis-cuss the recommendations that theworkshop produced for each of thetopic areas.

Research AreasThe workshop focused on three

research areas—security, informationfusion, and critical infrastructure—based on recommendations made bya National Research CouncilCommittee on Science and Tech-nology for Countering Terrorism.That committee called for researchin information and network security,new information technology foremergency response, and new infor-mation technology for the detection,remediation, and attribution ofattacks (information fusion).

Security includes authentication,availability, containment, detectionand identification, privacy, recovery,and new security models.Information technology for emer-gency response includes a variety ofproblems—most notably problemsfor critical infrastructure.Information fusion includes researchin data and text mining, data inte-gration, language technologies,image and video processing, and evi-dence combination. Rather thanconsider again the question of whichresearch areas are important fornational security, the CRAWorkshop took the results of theNRC committee’s report as a startingpoint.

To provide common grounds fordiscussion, the workshop began withpresentations by experts, who out-lined the current state of the art andactive research topics in each ofthese three areas.

Recommendations on Howto Facilitate Interactionbetween Research andPractice in Security andInformation Fusion

Top priority should be given tomethods for facilitating interactionbetween research and practice. It isespecially important that researchershave the ability to base their workon real problems with connections toreal data. Because of concerns aboutnational security and privacy, thiscan be particularly problematic.Nonetheless it is important thatresearchers and technologists haveaccess to scenarios and data that arerecognized as realistic and as repre-sentative of the challenges beingfaced by practitioners. If this is notthe case, research results face the riskof being dismissed as irrelevant orimmature.

In addition to providing access tothe problems and the data, programsmust be developed that facilitate anunderstanding of their counterpartsby both researchers and practitioners.Researchers need a deeper under-standing of the complex processes inwhich practitioners, such as govern-ment analysts, participate. Theyneed to be able to observe practi-tioners and their processes in action.Practitioners need an understandingof the potential of new technology.Most are not comfortable with newtechnology, and novel methods forintroducing technology must bedeveloped so that people can becomefamiliar with and test new systems—all while continuing to makeprogress on their real-world tasks.

Given these needs, the workshoprecommended that the followingactions be taken:

1. Create testbeds of open data.The workshop recommends estab-lishing a center that will make it eas-ier for both government and industryto provide data. In general, it is hardto generate synthetic data withenough scale; this is a research proj-ect worthy of its own funding. Issuesinclude development of new tech-niques for automatic scrubbing,agreement between researchers andintelligence agencies on what consti-tutes good, normal operational andattack scenarios, and a long-termfocus on establishing and maintain-ing the testbed. It was suggested thatdifferent research groups focus ondifferent aspects of the testbed; onegroup might focus on generatingoperational data, while anotherfocuses on generating attacks. Initialmodels for such a testbed are beingexplored under the NSF KDD pro-gram, a joint program with theNational Security Agency. Fundingfor such models involving these andother agencies should be provided.

2. Establish structures thatfacilitate interactions. Theworkshop recommended a variety ofstructures that could address theproblem of connecting research withpractice. Grants focusing solely onthe transfer of technology for ashort-term period should be estab-lished. Funding programs that stressinteractions between the intelligencecommunity and industry andresearch groups are also needed.National laboratories that focus onissues of security and data miningwould allow researchers and practi-tioners to come together for longerperiods of time. Such laboratoriescould provide the ability to generatelarge-scale simulations in whichexperiments could be carried out.Organizations such as In-Q-Telshould seek to encourage technolo-gies driven by needs and not by themarket, with special effort placed onremoving bureaucratic difficulties.Google provides a good model formoving from research to practicethat could be used as the basis fornew structures.

3. Adopt human factors methodsfor modeling and improving securityprocesses. It was recognized thatoften the security processes thatpractitioners follow are cumbersomeand error-prone. To facilitate under-standing of the tasks and the humanactivities involved, research shouldconsider and incorporate cognitiveapproaches such as scenario-drivenexercises, workflow modeling, cogni-tive think-aloud protocols, andexpert panels. Increasing theautomation of many of these securityprocesses, combined with rigorousanalysis, would eliminate many

opportunities for security breaches.4. Reconsider research para-

digms. Researchers and funders mustlook to long-term efforts that includethe continual development andimprovement of realistic testbeds andcareful evaluation of technologybased on those testbeds. The work-shop recommended that multiplecycles of evaluation are needed. Inthe first cycle, researchers mightwork with end-users to see how theyreact to initial tool functionality anddesign. In later cycles, after respond-ing to initial concerns, more rigorousevaluation could be undertaken. Thisis a process that may go throughmany cycles and takes time. Fundingagencies and users must recognizethat long time periods are needed forthis process to work well.

5. Create measures of effective-ness. If practitioners are to under-stand which technologies are worthbeing deployed, they need measuresof effectiveness that can help themdistinguish and choose amongoptions. Such measures shouldprovide qualitative assessments offunctionality and usefulness, as wellas the more typical quantitativemetrics.

Best Practices in Security,Real-Time Systems, andInformation Fusion

The workshop addressed the ques-tion of best practices primarilythrough breakout groups that focusedon each research area separately. Itbecame clear, however, that therewere commonalities across all areas.Unfortunately, it was agreed thatthere are not very many best prac-tices within individual areas. It wasdifficult enough to define ‘best prac-tice,’ let alone the appropriate prob-lems for which best practices shouldbe developed. Furthermore, bestpractices change so quickly that itwould be difficult to create a staticlist.

Instead, the subgroups looked tomechanisms and processes that couldbe put in place to dynamically trackbest practices. We report on recom-mendations separately for eachresearch area.

Information FusionWhile researchers are very often

focused on tools and methods, weagreed that what needs to be dissem-inated and described to the moregeneral community are the best toolsfor given tasks. We need a focus onthe problem, not the tool. Thus, asummary of what different searchengines do is not helpful; instead,practitioners need to know how itbehaves in a specific context.

CRA Workshop on Research Related to National Security:Report and RecommendationsBy Kathleen McKeown, Lori Clarke, and John Stankovic (Organizers)

September 2003 COMPUTING RESEARCH NEWS

This subgroup recommended thedevelopment of a playground for toolevaluation. The playground woulddefine scenarios and data againstwhich tools could be tested. Such aplayground might be set up on theWeb, allowing researchers to posttools and practitioners to specifyproblems against which they couldevaluate multiple tools. In order forthis to work, researchers must agreeon an annotation scheme for markupof data and common APIs for tightlycoupled or distributed architectures.In addition to tasks, games shouldalso be explored as a motivatingmechanism for exploring the best fitof a tool.

In summary, this subgroup did notthink it appropriate for any organiza-tion to develop a list of best prac-tices; rather, it thought it would bebetter to define an environment fordetermining best practice, given aparticular task. Best practice dependson context. This environmentshould be used to capture lessonslearned. It should be developed as aglass-box scenario, logging behaviorand allowing observation of end-users to see how well tools work, par-ticularly when personal preferencesplay a role. It is possible that anorganization such as the LinguisticData Consortium at the Universityof Pennsylvania would be appropri-ate for setting up and maintainingsuch an environment, if providedwith adequate funding.

Real-Time and Embedded SystemsThe few best practices in exis-

tence include formal methods usedfor core algorithms, real-time analy-sis, and quality of service guarantees.In addition, there are common mod-eling and analysis tools in use, aswell as integrated development envi-ronments. However, most of thesetools are limited to idealized systemsand situations. They also do not ade-quately address security and informa-tion fusion issues. Extensions tothese tools and best practices areneeded for all of these issues.

The most critical areas for whichbest practices are needed includemethods to deal with the integrationof constraints, dynamic real-timeaspects of the system, dependablesoftware development for real-timesystems, computer security, and moreprincipled development of large-scaledistributed systems, which typicallyare still ad hoc.

This subgroup recommended thedevelopment of a set of CriticalInfrastructure Protection (CIP) cen-ters that focus on science and pro-vide industry/research consortiums.Such centers could provide diversityon any given problem and will allowfor integration of security with real-time issues. Different centers mightfocus on different problems—emer-gency response systems; wireless sen-sor networks for security ofinfrastructure systems for power,water, and transportation; or cybersecurity on the Internet—but coop-erate with others.

SecurityThe security subgroup had the

least agreement on what constitutesbest practice, opting for the term“plausible practice” instead. Evensecurity itself encompasses manypossible areas, such as cryptography,network security, computer security,and security administration. Thesubgroup focused on securityadministration.

Recommendations include theneed for more quantitative researchon good security and evaluation. Forimprovement in security practice,the subgroup pointed out the needfor creating novel forms of attacks onexisting methods. Best practice isoften limited due to the installedoperating system and software, which are often decades behind thetechniques put forth by the researchcommunity. This dichotomy betweenresearch and practice in securitymeans that different recommenda-tions must be developed for differentsituations. Given that all of our sys-tems have vulnerabilities, it is unre-alistic to expect that any system canever be entirely secure. Instead, weneed to move toward strategies thatprovide security components that areself-configurable and, in the case ofattacks, self-healing.

Recommendations onStrategies for FundingResearch in These Areas

There are a number of programsalready in place at the different fund-ing agencies to address issues ofnational security.*

The workshop recommends that amix of funded programs targetingissues of national security be estab-lished and maintained. In particular,it is important that both short- andlong-term efforts be supported; eithertype of effort alone is not sufficient.Four critical issues were identified askey to development of new technol-ogy for national security:

1. Improve mechanisms for fund-ing technology transfer. We needbetter methods for funding efforts todeploy mature research into applica-tions. Possibilities include 12-monthfunding augmentation at the end ofexisting grants or short-term grantsfocused entirely on technology trans-fer. Improvement of the SmallBusiness Innovation Research(SBIR) model should also be consid-ered. The conversion rate from PhaseI to Phase II SBIR grants in the cur-rent model is fairly large, but manysmall companies are never weanedoff of SBIR grants; when these grantsend, the company also ends. A moregradual move between phases isneeded. In addition to short-termefforts, a study of mechanisms thatfacilitate tech transfer is needed.

2. Establish support for longer-term research on national security.The problems will not be addressedby deployment of existing researchalone. Many of the problems facingthe intelligence community are hardones and existing solutions are notavailable. Nonetheless, there are

research efforts underway that areapplicable to these problems thatcould be focused on this area.Funding programs that allow for thecreation of centers and focusedresearch over a long time period areneeded. The need for open and real-istic testbed data sources, comparableto the data used by the intelligencecommunity, is one example of anarea where new research is needed.These testbeds would in turn be usedfor other research. The NSF KDDand ARDA programs provide goodmodels for this type of funding.Additional programs such as these inmore areas are encouraged.

3. Create new programs thatfacilitate interactions betweenpractice and research. Such pro-grams could include a faculty centerwhere faculty are given clearances, ora scholar-in-residence program whereresearchers spend a sabbatical or ashorter period of time at one of theintelligence agencies or national lab-oratories where researchers and prac-titioners could be brought together.Programs that embody cognitivemethods for observing end-userneeds and the use of demo andemployed systems are particularlyimportant. Flexibility and experi-mentation with new models for pro-totyping, testing, and redesigningsystems are needed.

4. The research communitymust get involved. There is a needfor more participation by theresearch community in funding pro-grams. DARPA has a need for newprogram managers, and without themnew programs will not be initiated.NSF also has a need for rotators whoare willing to serve time at NSF tooversee funding programs. Researchrecommendations from the commu-nity are also influential in startingnew programs at both DARPA andNSF. For example, in order to estab-lish a cross-institutional workshop ona topic of relevance, NSF needs aWhite Paper from a university.Similarly, DARPA is open to sugges-tions from the community on newprograms.

*The National Science Foundation (NSF)has created at least one program jointly withthe National Security Agency under theKnowledge Discovery and Dissemination(KDD) program, and plans others. Themission of the Advanced Research andDevelopment Activity (ARDA) is to workclosely with the intelligence agencies and hasseveral programs (e.g., AQUAINT,NIMD) where researchers and intelligenceanalysts are teamed to work together onproblems and solutions. DARPA has no set-aside to address problems in informationsecurity, but initiatives can come through theprogram managers. That said, there areseveral ongoing security-related programswithin DARPA that bring together researchfrom different sites.

Workshop Participants James AllanUniversity of Massachusetts,Amherst

Kelcy AllweinDefense Intelligence Agency

Chris BuckleySabir Research

Jagdish ChandraGeorge Washington University

Yvo DesmedtFlorida State University

Helen GillNational Science Foundation

Virgil GligorUniversity of Maryland

Sally HoweNational Coordination Office

Andrew HumeAT&T

Rob KolstadSAGE Executive ProducerUSENIX

Jay LalaDARPA

Carl LandwehrNational Science Foundation

Elizabeth D. LiddySyracuse University

Stephen R. MahaneyNational Science Foundation

Kathleen R. McKeownColumbia University

Bert MiuccioCenter for Benchmark Services

Al MokUniversity of Texas at Austin

Salim RoukosIBM

Shankar SastryUniversity of California, Berkeley

Jonathan SmithUniversity of Pennsylvania

John StankovicUniversity of Virginia

Gary StrongNational Science Foundation

Kathleen McKeown (ColumbiaUniversity), John Stankovic (Universityof Virginia), and Lori Clarke(University of Massachusetts, Amherst)are all members of CRA’s Board ofDirectors.

September 2003 COMPUTING RESEARCH NEWS

Page 7

Arizona State University - CSEAuburn University - CSSEBall State University - CSBoston College - CS*Boston University - CSBowdoin College - CS*Bradley University - CSBrandeis University - CSBrigham Young University - CSBrown University - CSBryn Mawr College - MCSBucknell University - CSCalifornia Institute of Technology -

CSCalifornia Polytechnic State

University - CSCalifornia State University, Chico -

CSCalifornia State University,

Hayward - MCSCarnegie Mellon University - CSCarnegie Mellon University - ECECase Western Reserve University -

EECSClemson University - CSColgate University - CS*College of Charleston - CSCollege of William & Mary - CSColorado School of Mines - MCSColorado State University - CSColumbia University - CSCornell University - CSCornell University - ECE*Dalhousie University - CSDartmouth College - CSDePaul University - CSDrexel University - CSDuke University - CSEmory University - MCS*Florida Atlantic University - CSEFlorida Institute of Technology - CSFlorida International University - CSFlorida State University - CSFlorida State University - ISGeorge Washington University - CSGeorgia Institute of Technology - CSGeorgia Institute of Technology -

ECEGeorgia Southern University - ITGeorgia State University - CISGeorgia State University - CSGrinnell College - MCSHarvard University - CSHarvey Mudd College - CSIllinois Institute of Technology - CSIllinois State University - ACSIndiana University - CSIndiana University - IIndiana University - LIS*Iowa State University - CSJohns Hopkins University - CSJohns Hopkins University - SI*Juniata College - IT*Kansas State University - CISKent State University - CSLehigh University - CSELong Island University - ICSLouisiana State University - CSMassachusetts Institute of

Technology - AA*Massachusetts Institute of

Technology - EECSMiami University - CS*Michigan State University - CSMichigan Technological University -

CSMississippi State University - CSMontana State University - CSNaval Postgraduate School - CS*New Jersey Institute of Technology -

CCSNew Mexico State University - CSNew York University - CSNorth Carolina State University -

CSNortheastern University - CSNorthwestern University - CS

Northwestern University - ECEOakland University - CSEOhio State University - CISOhio University - EECSOklahoma State University - CSOld Dominion University - CSOregon Health & Science

University - CSEOregon State University - CSPace University - CSISPennsylvania State University - CSEPennsylvania State University - ISTPolytechnic University - CISPortland State University - CSPrinceton University - CSPurdue University - CSPurdue University - ECERensselaer Polytechnic Institute - CSRice University - CSRochester Institute of Technology -

CSRoosevelt University - CS&T*Rutgers University, Busch Campus -

CSSanta Clara University - CESimon Fraser University - CSSingapore Management University -

IS*Southern Illinois University,

Carbondale - CSSouthern Methodist University -

CSESouthern Polytechnic State

University - CSEStanford University - CSState University of New York,

Albany - CSState University of New York,

Binghamton - CSState University of New York, Stony

Brook - CSStevens Institute of Technology - CSSwarthmore College - CSSyracuse University - ISTemple University - CISTexas A&M University - CSTexas Tech University - CSTrinity College - CS*Tufts University - CSUniversity at Buffalo - CSEUniversity at Buffalo - ISUniversity of Alabama,

Birmingham - CISUniversity of Alabama, Tuscaloosa -

CSUniversity of Arizona - CSUniversity of Arkansas - CSCEUniversity of Arkansas at

Little Rock - IUniversity of California,

Berkeley - EECSUniversity of California,

Berkeley - IMSUniversity of California, Davis - CSUniversity of California, Irvine - ICSUniversity of California,

Los Angeles - CSUniversity of California,

San Diego - CSEUniversity of California,

Santa Barbara - CSUniversity of California,

Santa Cruz - CEUniversity of California,

Santa Cruz - CSUniversity of Central Florida - CSUniversity of Chicago - CSUniversity of Cincinnati - ECECSUniversity of Colorado, Boulder - CSUniversity of Delaware - CISUniversity of Denver - CSUniversity of Florida - CISEUniversity of Georgia - CSUniversity of Hawaii - ICS*University of Houston - CSUniversity of Idaho - CSUniversity of Illinois, Chicago - CS

University of Illinois, Urbana Champaign - CS

University of Iowa - CSUniversity of Kansas - EECSUniversity of Kentucky - CSUniversity of Louisiana at

Lafayette - CACSUniversity of Louisville - CECSUniversity of Maine - CSUniversity of Manitoba - CSUniversity of Maryland - CSUniversity of Maryland,

Baltimore Co - CSEEUniversity of Massachusetts,

Amherst - CSUniversity of Michigan - EECSUniversity of Michigan - IUniversity of Minnesota - CSEUniversity of Mississippi - CISUniversity of Missouri, Rolla - CSUniversity of Nebraska, Lincoln -

CSEUniversity of Nevada, Las Vegas -

CSUniversity of Nevada, Reno - CSUniversity of New Brunswick - CSUniversity of New Hampshire - CSUniversity of North Carolina at

Chapel Hill - CSUniversity of North Carolina,

Charlotte - ITUniversity of North Dakota - CSUniversity of North Texas - CSUniversity of Notre Dame - CSEUniversity of Oklahoma - CSUniversity of Oregon - CISUniversity of Pennsylvania - CISUniversity of Pittsburgh - CSUniversity of Puget Sound - MCSUniversity of Redlands - MCSUniversity of Rochester - CSUniversity of South Alabama - CISUniversity of South Carolina - CSEUniversity of Southern California -

CS

University of Tennessee, Knoxville -CS

University of Texas, Arlington - CSEUniversity of Texas, Austin - CSUniversity of Texas, Dallas - CSUniversity of Texas, El Paso - CSUniversity of Toledo - EECSUniversity of Toronto - CSUniversity of Toronto - ECEUniversity of Tulsa - MCSUniversity of Utah - CSUniversity of Vermont - CSUniversity of Virginia - CSUniversity of Washington - CSEUniversity of Washington - IUniversity of Washington, Bothell -

CSUniversity of Washington,

Tacoma - CSSUniversity of Waterloo - CSUniversity of Wisconsin, Madison -

CSUniversity of Wisconsin,

Milwaukee - EECSUniversity of Wyoming - CSUtah State University - CSVanderbilt University - EECSVirginia Commonwealth

University - CS*Virginia Tech - CSWake Forest University - CSWashington State University - EECSWashington University in St. Louis -

CSWayne State University - CSWest Virginia University - CSEEWestern Michigan University - CSWilliams College - CSWorcester Polytechnic Institute - CSWright State University - CSEYale University - CSYork University - CS

Affiliate Professional Society Members

American Association for Artificial Intelligence

Association for Computing Machinery

Canadian Association of Computer Science (CACS/AIC)

IEEE Computer Society

Society for Industrial and Applied Mathematics

USENIX Association

2002-03 Computing Research Association Members

Accenture Technology Labs*AT&T LabsAvayaComputer Science Research

Institute at Sandia National Labs

Fraunhofer Center for Research in Computer Graphics

Fujitsu Laboratories of AmericaHewlett-Packard Co.Honda R&D AmericasIBM ResearchIDA Center for Computing Sciences*Institute for Human & Machine

CognitionIntel Corp.Lawrence Berkeley National

Laboratory*

Los Alamos National Laboratory*Lucent Technologies, Bell LabsMicrosoft Corp. (Sustaining Member)Mitsubishi Electric Research LabsNational Center for Atmospheric

ResearchNational Center for Supercomputing

ApplicationsNEC Laboratories AmericaPalo Alto Research Center Panasonic Information & Networking

Technologies LabRicoh Innovations, Inc.San Diego Supercomputer CenterSun Microsystems, Inc.

(Supporting Member)Telcordia Technologies

(Supporting Member)

Labs and Centers Members

*New members 2002-03

Undergraduate Awards

Pictured above receiving the female CRA Outstanding UndergraduateAward for 2003 is Bistra Dilkina, Simon Fraser University. The awardwas presented by CRA's executive director, Andy Bernat, at ACM'sawards banquet in San Diego in June. Looking on is Julie Thornton,Kansas State University, who received the female runner-up award.

Pictured above receiving the male CRA OutstandingUndergraduate Award for 2003 is Omar Khan, CornellUniversity. The award was presented by CRA's executivedirector, Andy Bernat, at ACM's awards banquet in SanDiego in June.

COMPUTING RESEARCH NEWS September 2003

Page 8

COMPUTING RESEARCH NEWSVol. 15/No. 4

Computing Research News(ISSN 1069-384X) is published five times per year, in January, March, May,September, and November. Copyright 2003 by the Computing ResearchAssociation (CRA), 1100 Seventeenth Street, NW, Suite 507, Washington, DC20036-4632; tel. 202-234-2111. All rights reserved. Material in CRN is notendorsed by CRA nor intended to reflect any official positions of CRA or its board.

Subscriptions: Call 202-234-2111, send e-mail to crn@cra.org, or mail subscriptioninquiries to CRA, 1100 17th Street, NW, Suite 507, Washington, DC 20036-4632.A free subscription is available to qualified subscribers. One-year paid subscriptionsare $30 in the United States, $45 (U.S.) in Canada, and $54 (U.S.) elsewhere.

Change of Address: Note that a change of address must include the old and newaddresses with ZIP+4. Please include a street address or PO Box number.

Postmaster: Send address changes to: CRA, 1100 17th Street, NW, Suite 507,Washington, DC 20036-4632. Postage paid at Washington, DC.

CRA Communications CommitteeJan Cuny (University of Oregon), ChairDoris Carver (Louisiana State University)Michael Jones (Microsoft Research)Frank Tompa (University of Waterloo)Stuart Zweben (Ohio State University)

Computing Research Association StaffAndrew Bernat, Executive DirectorPeter Harsha, Director of Government AffairsDana Neill, Business ManagerSusanne Pile, Administrative AssistantCarla Romero, Director of ProgramsJean Smith, Sr. Communications Associate and CRN EditorDrew Sutter, Program AssociateJay Vegso, Manager of Membership and Information Services

principle in hiring, tenure, scholar-ships, and the provision of lab spaceand equipment. It is the law on thebooks, and schools that are notfollowing it now should be put onnotice that Title IX will be enforcedas vigorously in the halls of academiaas it is on athletic fields.

Applied more comprehensively,Title IX can serve as a valuable toolin not only breaking down formalbarriers to entry, but in actuallyensuring that more women succeedat math, science and engineering—or any other disciplines they choose.

This will require a sea change inthe attitudes of many people at thenation’s educational institutions. Inmy view, the Federal governmentshould move now to bring about thatchange in a number of ways—partic-ularly by helping to clearly definejust how pervasive and how institu-tionalized discrimination in thesefields has become.

Despite a great deal of anecdotalevidence, there are few studies aboutjust how often women are discour-aged from studying math, scienceand engineering, and how regularlywomen who do enter these fieldsface discouragement from theirsupervisors and colleagues. That iswhy Senator Barbara Boxer (D-CA)and I have asked the GeneralAccounting Office to find out whatis being done, if anything, to ensureTitle IX enforcement in math,science, and engineering.

I also passed legislation requiringa review of whether the NationalScience Foundation is meeting itsgoals to expand opportunities inthese disciplines for women, minori-ties, and people with disabilities, par-ticularly in faculty hiring,promotion, tenure, and allocation oflab space. Another section of myamendment required a study to lookat gender differences in the distribu-tion of Federal research and develop-ment funds.

Shifting awareness in the govern-ment’s scientific culture may be justas critical as changes at our educa-tional institutions. As chair of theScience, Technology, and SpaceSubcommittee in the previousCongress, I held several hearings onthe topic of encouraging women toenter math and science fields. Icalled on NASA AdministratorSean O’Keefe and his agency to use anew education initiative to helptriple the number of women graduat-ing with math, science, and engi-neering degrees by 2012. With clearevidence of inequity, there was noreason a Federal agency launching aneducation program should not do sowith an eye to closing the gendergap.

I am committed to continuing topush government agencies and insti-tutions receiving Federal funds toabide by and actively consider thetext and spirit of Title IX. That iswhy I have signed on as a co-sponsorto two resolutions, one introduced bySenator Joe Biden (D-DE) and theother by Senator Patty Murray (D-WA), both of whom reaffirm our

commitment to the principles setforth in the Title IX law.

After I began advocating publiclyon this issue, I received an e-mailfrom a professor on a search commit-tee for a chemistry professor. He waslamenting the fact that out of 80applicants for the position, only sixwere women. This frustrated educa-tor suggested that gender inequityhad to be attacked much earlier inthe process, and I agree—it must beattacked much earlier, and in somecases even outside the scope of TitleIX.

In addition to the barriers womenface in the educational arena, cul-tural stereotypes discourage girls frommath and science at a young age.Young girls in their formative yearstoo often receive the message thatmath and science are not meant forthem. In fact, one popular talkingdoll on the market a few years agoactually spouted catch phrases like“math is hard” and “shopping is fun.”Inside the classroom, a lack of expec-tations and a shortage of female rolemodels frequently perpetuate theproblem.

The good news is that the stereo-types can be overcome. NancyStueber, the president of the OregonMuseum of Science and Industry, hastold me stories of young girls whowalked into the museum thinkingthat science and math were for boys.When the girls were asked to drawpictures of a scientist, they all drewan older white man in a lab coat.However, after participating in pro-grams at the museum, those samegirls drew pictures of women in labcoats. They had begun to imaginethemselves as mathematicians, scien-tists and engineers.

My goal is to make sure thatwhen those young women choosetheir careers, this nation’s educa-tional institutions are fully compli-ant with the law that guaranteesthem equal access. Careers in mathand the hard sciences are theirright—and it is in our nation’s inter-est to encourage them. The enforce-ment of Title IX may well beAmerica’s best hope to maintain ourposition at the forefront of key scien-tific disciplines and our leadership inthe world community.

Senator Wyden (D-OR) was elected tothe Senate in 1996 and is a former chairof the Senate Commerce Subcommitteeon Science, Technology and Space.[The Senator’s views do not necessarilyrepresent the views of CRA.]

“Expanding the Pipeline” is a regularfeature in CRN, prepared by the CRAWomen’s Committee (CRA-W). Itserves as a vehicle for describing projectsand issues related to women in CSEand a source of information for issuesfaced by underrepresented groups inCSE.

Title IX from Page 1

September 2003 COMPUTING RESEARCH NEWS

Page 9

research infrastructure, often withmatching funds from provincial gov-ernments and industry. While noneof these programs is focused on com-puter science specifically, computerscience departments have certainlytaken at least their fair share of thesefunds. Overall, the general climate inCanada for research seems to behighly positive, with both levels ofgovernment actually supporting theirinnovation agenda with real funding.

Enrollment TrendsA recent survey of Canadian

computer science departments hasrevealed a general leveling off inundergraduate enrollments with someuniversities experiencing consider-able decline, especially in the earlyyears of their programs. It is alsoincreasingly difficult to get intern-ship and co-op placements for stu-dents, and the job market generallyfor computer science graduates issoft. This seems to be at least aNorth America-wide phenomenonin the wake of the dot-com crash.

On the other hand, graduateenrollments continue to surge, anatural enough result of increasednumbers of faculty members andincreased research funding. Unlessthe job market turns around, anobvious crunch point lies ahead,with many freshly-minted Ph.D.graduates and not many new facultyand industrial research positions forthem. This is certainly worrisome.One trend that may help to offsetthis worry is a widespread movementamong Canadian computer sciencedepartments to forge interdiscipli-nary programs with a variety of otheracademic disciplines. New facultypositions may become available tosupport these interdisciplinary pro-grams, and many new students maywell be attracted to such programs,especially those who have not beenattracted to traditional notions ofcomputer science.

AccreditationThe Computer Science Accredi-

tation Council of Canada(CSAC)—a joint body of theCanadian academic computerscience community and theCanadian Information ProcessingSociety (CIPS), the Canadian organ-ization of software professionals—hasdrafted revised standards for bothcomputer science and software engi-neering programs. At a recent meet-ing of computer science departmentchairs, however, there was someunrest with the conservative natureof these standards, and a wish on thepart of many departments to expandthe notion of computer science toinclude much more broadly interdis-ciplinary perspectives and to havestandards for a much wider variety ofprograms. The new standards andthese broader issues are now beingdiscussed by CSAC and CACS/AIC.

Software EngineeringSoftware engineering professional-

ism is still a hot issue in Canada.Starting in the mid-90s, the engi-neering profession in Canada beganto assert what it saw as its exclusiveright to practice in the area of soft-ware engineering, an assertion thathas been hotly contested by theCanadian computer science commu-nity. There are three interrelatedfronts to this dispute.

The first front has (nominally)been an argument about the use ofthe term “engineering.” In 1997, theengineering profession launchedlegal action against MemorialUniversity of Newfoundland (MUN)for offering a software engineeringprogram within their computer sci-ence department in the faculty ofscience. The basis for this lawsuitwas a 1989 trademark filed by theengineering profession on the terms“engineer” and “engineering.”Eventually, MUN was joined in itsdefense by the Association ofUniversities and Colleges of Canada(AUCC, the organization of

Canadian universities, which wasconcerned about academic freedom);CIPS (which was concerned aboutits members’ right to practice); andCACS/AIC (which was concernedabout both). In 1999, the lawsuit wassuspended for a period of 5 years toallow both sides to work out a com-promise. Despite several early com-promise proposals, it now appearsthat no reconciliation will beachieved, even though the morato-rium expires in another year.

The second front is the creationand accreditation of software engi-neering programs within and outsideof engineering. For the past severalyears, engineering schools acrossCanada have been creating theirown software engineering programs,variations on standard engineeringprograms and accredited by theCanadian Engineering AccreditationBoard (CEAB). Most engineeringschools now have such programs,although the software-specificcourses are often provided by com-puter science departments. So far,seven of these programs have beenaccredited by CEAB. At the sametime, a number of computer sciencedepartments have created their ownsoftware engineering programs, varia-tions on standard science programs,five of which have been accreditedby CSAC.

The third front involves the rightto practice. Several years ago, engi-neering associations across Canadabegan to successfully lobby forchanges to their provincial and terri-torial Engineering Acts (underwhich the engineering profession isgranted exclusive right to practiceengineering) to make it easier toclaim “emerging areas of engineeringpractice.” These changes tended toreplace the existing description ofthe scope of engineering practice(which heretofore had included alengthy list of engineering works)with an essentially circular definitionof engineering as that discipline thatdraws on engineering principles.

Such a broad definition would cer-tainly make it easier to “claim” soft-ware engineering (as well as manyother areas not traditionally consid-ered to be engineering). CIPS,CSAC, and CACS/AIC are attempt-ing to see if further revisions can bemade to these Engineering Acts toensure that the right to practice soft-ware engineering is clearly protectedfor the wide variety of people whohave the appropriate skills to do so,whether or not they are professionalengineers. Negotiations withengineering societies and govern-ments have not yet resulted in such protection.

Given the size and importance ofthe software industry, and the impli-cations for everything from the train-ing of software professionals to thefinal certification of software quality,the final reconciliation of the soft-ware engineering dispute is of greatimportance not only to Canadiancomputer science departments, butto the entire Canadian economy.Unfortunately, at the current time, itappears that it will be up to thecourts to make the final decisionsabout the issues involved in the soft-ware engineering dispute, since noappropriate compromise positionsseem to be on the horizon.

So, that’s it for this year’s report.Hopefully I will be able to updateyou next year with lots of positivenews from the far north.

Note: The word “department” isused generically to refer to department,school, or faculty of computer science.

Gord McCall