The profession - The computing profession at a crossroads

  • Published on

  • View

  • Download


92 ComputerB y any measure, the computingfield enjoyed a golden era dur-ing the past decade. Abouttwo years ago, however, arather dramatic meltdownbegan as harsh economic realitiesunfolded, causing an attendant up-heaval in the field. Suddenly, there wasa glut of seasoned computing profes-sionals in the job market, and busi-nesses became disenchanted with thevalue of their computing investments. This newfound wisdom promptedseveral questions and concerns aboutreturn on investments, softwaresappalling quality, the negative impactof immature and fleeting technologies,and the deeply entrenched practice ofsoftware development as art, not science.ACADEMIC IRRELEVANCEEmpirical evidence reveals thatfewer than 20 percent of computingprofessionals in the USs financial ser-vices industry have any formal acade-mic background or training in com-puting systems. Over the past decade,project managers ignored fresh com-puting-discipline graduates, insteadhiring crash-course training-center orliberal-arts college graduates with acouple of years experience in the indus-try. These managers held the view thatrecent computing graduates needed togo through long company-sponsoredtraining programs and apprenticeshipsbefore they could perform useful work.On the academic and training side,institutions from universities to com-munity colleges to street corner mom-and-pop training centers expandedexisting computing programs orstarted new ones. These programscame in many forms, partly to addressthe allegedly increased demand forcomputing professionals. Others sawthis trend as an opportunity to realizetheir cherished dream of becoming aprogram director, chair, or dean.Academia, however, failed in its edu-cational mission in the name of funda-mentals versus fleeting practicalknowledge. Neither academic leadersnor professional organizations such asthe IEEE Computer Society and theACM expressed concern about theenormous impact and attendant prob-lems caused by mushrooming com-puting academic programs. The fallout from this trend includeda curriculum irrelevant to overallindustry needs and a lack of pragmaticand enforceable curriculum standards,quality control, resources, and an iden-tity for the discipline itself.COMPUTING RENAISSANCETodays computing profession offersfour broad employment categories: design and manufacture of com-puter hardware; development of system software,embedded systems, and general-purpose software such as databaseservers, toolkits, and applicationframeworks; applications development withminimal programming, using con-figurable and interoperable third-party commercial off-the-shelfcomponents; and application implementation usingpackaged software such as enter-prise resource planning systems.Only employers for the first two cate-gories seem to attach significance toformal academic degrees in the com-puting discipline. The third categoryemploys more people than the others.The fourth category demands bothcomputing technical skills and a deeperunderstanding of each applicationdomains business processes.In addition to programming exper-tise, employers expect future comput-ing-discipline graduates to mastermany other skills, including systemsengineering and end-to-end systemarchitecture and its elements: security,performance, scalability, availability,reliability, supportability, manageabil-ity, and maintainability. Programming services will emerge asThe ComputingProfession at aCrossroadsVenkat N. Gudivada, University of Michigan, DearbornT H E P R O F E S S I O NContinued on page 90Industry and academia mustwork together to chart thebest course for theprofessions future.90 ComputerT h e P r o f e s s i o ncomputer science is inherently an inter-disciplinary subject with little coreknowledge to justify its distinct iden-tity. Yet others compare computer sci-ence with mathematics in that, just aswe need applications to extract practi-cal usefulness from mathematics the-ory, computer science provides theunderpinning for academic disciplinesin engineering and the sciences.Therefore, these disciplines can betterfill the role of educating computing stu-dents and designing and developingmore relevant curricula that betterreflect these students domain-specificneeds. The emergence of computer sci-ence courses transformed to suit anapplication-oriented context in disci-plines such as engineering manage-ment, systems engineering, biology,and physics vindicates this argument.COMPUTING VERSUS ENGINEERINGIn a previous column (Jobs, Trades,Skills, and the Profession, Computer,Sept. 2002, pp. 104, 102-103), NevilleHolmes provided an illuminating lookat how the computing profession dif-fers from the traditional engineeringbranches. According to Holmes, engi-neering clearly demarcates the rolesplayed by its practitioners, tradespeo-ple, and the resulting products endusers.The computing field lacks any suchdemarcation, however. Industry viewsthe computing profession as a collectionof variegated and vacillating skills withno clear structure. Holmes argues thatwe can correct some of the issues thediscipline faces if we accept computingsstatus as a secondary profession, onethat aids and abets other professions.Meanwhile, software engineering isemerging as a distinct discipline whosea commodity item, with most pro-gramming outsourced to the lowestbidder for economic reasons. The mar-kets current development toolkits andframeworks already sport high-levelabstractions that even some smart highschool students can use to develop sim-ple yet useful applications.Thus, an industry renaissance infavor of developing functionally exten-sible applications has led developers todesign applications based on an archi-tectural framework with built-in exten-sibility points. These designs allow theimplementation of new functionalitywithout writing any code. Functionallyextensible systems lie somewherebetween approaches like parameter-ized and aspect-oriented programmingand enterprise resource planning sys-tems. This trend will further lessen theneed for people with no skills beyondprogramming.RevitalizationIn addition to the recent demise ofmany dot-com companies, further con-solidation in the computing industry willresult in the disappearance of some busi-nesses and the reemergence of those withclear vision, distinctive competence, andan effective strategy. This revitalized computing industrywill demand high quality from employ-able graduates who can understand andsolve real problems and orchestrate end-to-end solutions. The responsibility forsupplying this expertise rests primarilywith academia, while industry providesan active advisory and support role.Identity challengesThe academic front faces challengesto the identity of the computing disci-pline in both general and computer sci-ence. The ascent to prominence ofinterdisciplinary research in bioinfor-matics, computational biology andphysics, library science, and digitallibraries has only made defining, estab-lishing, and asserting the true identityof the discipline more imperative thanever.Some circles have even asserted thatroots lie in computer science. DavidLorge Parnas offers compelling rea-sons to support that perspective inSoftware Engineering Programs AreNot Computer Science Programs(IEEE Software, Nov./Dec. 1999, pp.19-30). Parnas cogently argues for andsubstantiates the merits of developingsoftware engineering programs inengineering departments, especially inelectrical and computer engineering. THE UNCERTAIN FUTUREComputing professionals may soonbe held legally liable for a defectivesoftware system. Accountability andresponsibility should thus be at theheart of any professional practice.Therefore, we face a compelling needto address issues related to curriculumstructure and its standardization, itscurrency and relevance to overallindustry needs, its program quality,and the extent to which it fosters collaboration between industry andacademia.Wake-up callAcademia must awaken to the hardrealities the computing professionalfaces in the field and then become moreresponsive and accountable. Now that the body of knowledgehas matured and stabilized, we need todefine curricula with much more rigor,then standardize it across all institu-tions irrespective of their ability toacquire the needed resources to imple-ment that curriculum. This approachcontrasts sharply with todays sug-gested computer science curriculamodels and accreditation standards.We must strike a balance betweenfocusing on fundamentals and impart-ing practical skills. Unfortunately,many departments lack the resourcesfor accomplishing this task.Further, in computer science instruc-tion, a dichotomy exists between the-ory and practice. Professors frequentlytake a concept-oriented approach toteaching, compartmentalizing anypractice-based learning. Often, cap-stone courses in computer science seekContinued from page 92The academic front faceschallenges to the identity of the computing disciplinein both general andcomputer science.May 2003 91sionals with advanced degrees, espe-cially those who have gained theirpractical insight by working in thetrenches. Similarly, few opportunitiesexist for faculty to spend their sabbat-ical time working with industry tosolve real problems.At least part of the faculty in soft-ware engineering programs shouldpossess substantive practical experi-ence and insighttraits somewhatmore valued in engineering disciplinesthan in computer science. Institutionscommonly expect the engineering fac-ulty to possess substantial industryexperience because designing anddeveloping a real-world system posemuch greater challenges than writinga research paper. In contrast, com-puter science departments typicallyfavor publications over practical expe-rience.Academia alone cannot extricateitself from the current situation with-out the active leadership of industryand professional organizations such asthe IEEE Computer Society and theACM. For better or worse, academi-cians dominate professional societies,thus these societies primarily reflectacademics needs: organizing confer-ences and publishing magazines, jour-nals, and conference proceedings.Introspection is paramount.Mission, vision, goals, and objec-tives must all be closely examined andaligned with reality, while balancingtactical needs with long-term strategy.Some energy should be channeledaway from introducing more confer-ences and publications and into reduc-ing the explosion of more or less thesame information packaged differ-ently and appearing in a plethora bring coherence and consummationto the knowledge students have gainedin various coursesyet theory andpractice should be interwoven acrossthe curriculum.This fundamental difference ininstruction style raises the question ofcomputer science and software engi-neering programs coexisting within thesame academic department. A coexis-tence scenario is desirable because itlets software engineering programsleverage computer science programscore competency. But several factorsargue for separating the two disci-plines: curriculum structure and stan-dardization, accreditation and licens-ing, and culture clash. However,resource-related issues will force manysoftware engineering programs todebut in the coexistence scenario; theywill thus assert their own identity onlygradually.Meaningful collaborationAcademia and industry should col-laborate more meaningfully and pro-ductively. Although this statement hasbeen made by many, time and again,collaboration has yet to go beyondobtaining grant money and establish-ing inactive industry advisory com-mittees.Industry views the professor as anidealist detached from reality. But sim-ply blaming academia for the fieldscurrent state doesnt ameliorate the sit-uation. Working through professionalsocieties to shape the curriculum, pro-viding real-world case studies for class-room use, establishing programsthrough which faculty can spend sum-mer and sabbatical time in industry,and exploring other avenues tostrengthen cooperation would benefitindustry.EXPERIENCE MATTERSGenerally, academia doesnt valueindustry experience. Worse, industrysimply ignores academic experience.This hostile attitude doesnt help theprofessions advancement. It preventsattracting seasoned industry profes-CERTIFICATION AND LICENSINGProfessional societies should cham-pion both the need for and administra-tion of certification and licensing. Theyshould also liaise with government,business, and industry to educate thesesectors about the significance of hiringlicensed professionals and the processinvolved in that licensing.Currently, the market treats the ITworkforce as disposable labor. Societiesshould actively seek to improve thestature and public image of computingprofessionals and address issues thataffect the careers of computer scienceand software engineering graduatesand practicing professionals.The recently initiated IEEE ComputerSocietys Certified Software Develop-ment Professional credential is a goodstep in this direction (Leonard L. Tripp,Benefits of Certification, Computer,June 2002, pp. 31-33). Counter-arguments to certification have beenadvanced, however, on the grounds thatpoor software quality stems from morefundamental problems than softwareengineers not all learning the samematerial (Adam Kolawa, CertificationWill Do More Harm than Good,Computer, June 2002, pp. 34-35).E ducators must recognize the pro-fessions current reality (SpencerJohnson, Who Moved My Cheese?Putnam, 1998), forge a vision forimproving it, devise a strategy for real-izing this vision, then provide soundleadership for its implementation. If wedo not, survival of the fittest alonewill determine the professionsDarwinian future. Venkat N. Gudivada is a visiting asso-ciate professor of computer and infor-mation science at the University ofMichigan, Dearborn. Contact him Neville Holmes, School of Computing, University of Tasmania, Locked Bag 1-359, Launceston 7250; societiesshould champion both theneed for and administrationof certification andlicensing. Index: CCC: 0-7803-5957-7/00/$10.00 2000 IEEEccc: 0-7803-5957-7/00/$10.00 2000 IEEEcce: 0-7803-5957-7/00/$10.00 2000 IEEEindex: INDEX: ind:


View more >