Transcript
  • 7/28/2019 Bridging the Gap Between Academia and Industry - Germany

    1/9

    BridgingtheGapbetween

    AcademiaandIndustry

    StevensAwardLecture

    byHarryM.Sneed

    attheCSMR2009Kaiserslautern,Germany

    March25th,2009

    CSMR2009

    StructureoftheLecture

    1)Thecurrentstatusofcooperationbetween

    AcademiaandIndustry

    2)Thereasonsbehindthegapbetween

    AcademiaandIndustry

    3)Theconsequencesofthegap

    4)Whatcanbedonetoclosethegap

    5)Suggestionsforfuturecooperation

    betweenAcademiaandIndustry

    CSMR20091

  • 7/28/2019 Bridging the Gap Between Academia and Industry - Germany

    2/9

    CurrentStatusofCooperationbetweenAcademiaandIndustry

    Asseenfromthesideofindustry:

    IndustrialUsersaredrivenbytheirvendorsandtheITmedia(newspapersandmagazines)

    Industrialparticipationinacademicconferenceshasbeendecliningforyears(ICSE,ICSMandCSMRaregoodexamplesofthis).

    Industryfeelsthatthereislittleitcanlearnfromacademia

    especiallyinthefieldsofmaintenanceandreengineering. IndustrialITpersonneldonotaccessacademicjournals.

    IndustrialITusersarenotawareofwhatacademicshavetooffer

    ITmanagersdistrustacademicians,theybelievetheyonlywanttoexperimentattheirexpense.

    CSMR20092

    CurrentStatusofCooperation

    betweenAcademiaandIndustry

    AsseenfromthesideofAcademia

    Academicsaredrivenbytheirconferencesandtechnical

    journalsandtheirneedtopublish.

    Academicsseldomattendindustrialconferencesastheyfeel

    thisisbelowtheirstandard.

    Academicslookdownuponindustrialnewspapersand

    magazines.

    Academicsarenotawareoftheproblemsandconstraintsof

    industry.

    AcademicsdistrustITmanagers,whotheybelieveonlywant

    toexploitthemtoachievetheirbusinessgoals.

    CSMR20093

  • 7/28/2019 Bridging the Gap Between Academia and Industry - Germany

    3/9

    CounterExamplesExceptionsconfirmtheRules

    Whereindustryandacademiaworkbesttogetheriswheretheyareone.

    Thishappenswhenacademicsfoundindustrialcompaniestoexploittheirideas.Googleisagoodexampleofthis.

    Italsohappenswhenindustryfinanceslaboratoriesinanacademicenvironment.TheUBSLab attheETHZrichisagoodexampleofthat.

    ThereareinstitutionswhichcombinebothsidesliketheFrauenhofer InstituteinKaiserslauternandMaryland.

    TheTUMunichismakingaconcentratedefforttogetinvolvedinindustrialprojectsinmaintenance,testandreengineering.

    Unfortunately,mostofthecurrentsoftwareindustryleadersgrewupwithoutacademicinvolvement(IBM,MicroSoft,Oracle,SAP).Latertheytriedtocompensatethatbyinstallingtheirownresearchlabs,butthespiritoftheseorganizationsisbasicallyantiacademic.

    CSMR20094

    ReasonsbehindtheGapbetween

    AcademiaandIndustry

    AcademicsandIndustrialistshaveadifferentmindset,theyare livingin

    differentworlds.

    AcademicsandIndustrialistsarepursuingdifferentgoals.TheAcademicis

    strivingforrecognitionfromhispeers.TheIndustrialistisstrivingto

    survive.

    Industrythinksintermsofshortrangegoals,Academiahasalongrange

    perspective. Industryprefersprovensolutionswithalowrisk,Academiaisinterested

    increatingnewsolutionswithahighinnovationrate.

    Industryseeksforaminimumsolutiontominimizetheirriskwhereas

    academiastrivesforamaximumsolutiontomaximizetheirrecognition.

    Industryismainlyconcernedwithcosts.Academiacouldcarelessabout

    costs,itismainlyinterestedinthebenefits.

    CSMR20095

  • 7/28/2019 Bridging the Gap Between Academia and Industry - Germany

    4/9

    Reverse&ReengineeringasagoodExampleof

    theGapbetweenAcademiaandIndustry

    Industryhastomaintainahugeamountoflegacysoftware(over75%ofallsoftwareislegacysoftware).Itisnotparticularlyconcernedwiththequalityofthatsoftware.Itismainlyconcernedwithfindingpersonneltomaintainitandwithmigratingittonewplatforms.

    Academicsareconcernedwithimprovingthequalityoflegacysoftware,i.e.inreengineeringratherthan

    migration.Partofthatefforttoimprovethequalityoflegacysystemsisreverseengineering,atermconfusingtoindustrialusers,whodontperceivewhatitisgoodfor.Contrarytotheacademicbelief,reverseengineeringisnotaprerequisitetotransformingorwrappingcode,sincethisisalmostalwaysanautomatedprocess.

    CSMR20096

    WhyindustryrejectsReverse&

    ReengineeringTechnology

    Industrialusersareunwillingtobuyintoreengineering,i.e.raisingthequalityoftheirexistingsoftware,withoutknowingbeforehandwhatsavingsthatwillbringthem.Academiaisnotabletoguaranteethemthat.

    IndustrialusersarenotwillingtofinanceareengineeringprojectwithoutknowingbeforehandwhattheROIwillbe.

    AcademicsarenotabletocalculateaROIbecausetheydonothavethedatatoquantifythebenefits.

    Asaresult,itisdifficulttosellreengineeringandreverseengineeringprojects.Theseactivitiescanonlybecarriedoutwithinthescopeofthemaintenancebudgetwhichishardlyenoughtocoverthechangesandcorrections.Reverseandreengineeringare,incontrasttomaintenanceandmigration,oftenneglectedactivitiesinindustry,whereasinAcademia,theyreceivealotofattention.

    CSMR20097

  • 7/28/2019 Bridging the Gap Between Academia and Industry - Germany

    5/9

    WhatIndustryislookingforconcerningtheirlegacysystems

    Thereisaneedforbrowserstosearchthroughexistingcodetofind

    dependencies.

    Thereisaneedforimpactanalysistoidentifywhathastobechanged.

    Thereisaneedformetricstoestimatethecostsofchange.

    Thereisaneedformetricstoidentifyerrorpronecomponents.

    Thereisaneedfortoolstoautomaticallyconvertoldlanguages intonew

    oneswhilepreservingthefunctionalityofthesystem.

    Thereisaneedfortoolstotransformdatafromonedatabaseto another.

    Thereisaneedfortoolstowrapexistingcomponents,dataand

    proceduresforreuseinotherenvironments.

    Thesuccessofamigrationproject transformationorwrapping iseasily

    measured.Eitheritworksinthenewenvironmentoritdoesnt.Thisisnot

    trueofareverseorreengineeringproject,wherethegoalisaquality

    improvement.Theremustbeawaytomeasurequalityimprovement.

    CSMR20098

    OntheDiscrepancybetweenSupply

    andDemandinSoftwareMaintenance

    Itisobviousthattheacademiccommunityspendstoomucheffortonsubjectsthatarenotrelevanttoindustryandtoolittleeffortonrelevantsubjects.Thus,theofferingsofacademiadonotmatchtothedemandsofindustry.Toavoidthismismatch,academiashouldstudytherequirementsofindustry.

    Ontheotherhand,industryisoftennotawareofwhattheyreallyneed.Industrymustbeeducatedastotheinfluencethatcomplexityandqualityhaveonthecostsofmaintenanceandmigration.Theymustalsolearnwhatispossibleandwhatisnot.Itwouldbethetaskofacademiatoeducateindustry.

    CSMR20099

  • 7/28/2019 Bridging the Gap Between Academia and Industry - Germany

    6/9

    ConsequencesoftheGap

    AcademiaandIndustrytendtodriftapart

    Academicsliveintheirownworld,theyfirstinventsolutions,thenthey

    searchforaproblemthatfitstoit.Theyarereluctanttodeal withreal

    worldproblems,sincetheseareusuallywicked.Asaconsequence much

    oftheireffortiswastedontoyproblems.

    Academiaisnotteachingstudentswhattheyneedtoknowabout

    maintainingandevolvingexistingsystems.

    Industryusersletthemselvesbemanipulatedbyscruplesvenderswho

    likedrugdealersarebentonmakingtheircustomersdependent.Their

    inabilitytodecideforthemselvesisduetotheirlackofeducationand

    theirlackoftrustintheacademiccommunity.

    Industrysuffersfromalackofawarenessastohowtheycouldapproach

    theirproblems.Theydonotknowwhattoolsandtechniquesfrom

    academiacouldhelpthem.Theyalsosufferfromalackofstandardization

    andtheabilitytoobjectivelycomparealternatesolutions.

    CSMR200910

    ClosingtheGapbetween

    IndustryandAcademia

    MeasuresforclosingthegapbetweenAcademiaand

    Industry Academiciansshouldtakepartinindustrialprojects

    Professorsshouldbegivensabbaticalstoworkasmaintenanceprogrammers

    Practitionersshouldattendcoursesattheuniversitiestorefresh

    theirknowledge Thereshouldbecommonconferencestoappealtobothsidesin

    whichtherearetwodifferentstandardsforselectingpapers

    AccordingtoapollbytheIEEESoftware,bothsidesseekastrongercollaboration,howeverforthatweneedabettermutualunderstanding.

    Thereneedstobeacommonlanguage,i.e.setofterms

    Thereneedstobeastandardsetofmetricstomeasuresoftwaresystemsandprojects.

    CSMR200911

  • 7/28/2019 Bridging the Gap Between Academia and Industry - Germany

    7/9

    TowardsaCommonLanguageStepstowardabettermutualunderstanding

    Thelanguageusedbyacademiciansshouldbe

    understandabletopractitioners

    Reengineering&ReverseEngineeringaregoodexamples

    ReverseEngineeringmeanstoapractitionerpost

    documentinghissystemasopposedtopredocumentingit.

    Reengineeringcouldmeanmanythings renovating,

    restructuring,refactoring,etc.Oneshouldbespecific. Migrationneedstobekeptapartfromreengineering.

    Itwouldbebetternottousethetermsreengineeringand

    reverseengineeringatallsincetheyareconfusing.Theyhave

    beenmisusedtodenotetoomanyquitedifferentactivities,

    likebusinessreengineering,businessprocessreengineering

    architecturalreengineeringandcodereengineering.

    CSMR200912

    TeachingSoftwareMaintenance Softwaremaintenanceshouldbetaughtonrealsoftware

    systemsfromindustry

    Studentshavetocorrecterrorsinbothproceduraland

    objectorientedprograms

    Studentshavetomakechangestobothproceduraland

    objectorientedprograms

    Studentshavetoredocumentbothproceduralandobjectorientedsystems

    Studentshavetorestructureproceduralprogramsand

    refactorOOcomponents

    Studentshaveestimatethecostsofmaintenancetasks

    Studentshavetodesignproblemreportsandchangerequests

    Studentshavetodesignamaintenanceprocessforeachtype

    ofmaintenancetask

    CSMR200913

  • 7/28/2019 Bridging the Gap Between Academia and Industry - Germany

    8/9

    TeachingSoftwareMigration

    SoftwareMigrationshouldbetaughtonareallegacysoftwareapplication

    Studentsshouldmeasurethelegacyapplicationandcalculatethecostsofalternatemigrationstrategies

    StudentsshouldhavetotransformlegacyproceduralcodetoamodernOOlanguage

    Studentsshouldhavetowraplegacymodulesandconvertthemoverintowebservices

    Studentsshouldhavetoconvertoldmasterfilesoverintorelationaldatabases

    StudentsshouldhavetotransferolddatainterfacefilesoverintoXMLfiles

    Studentsshouldhavetoperformaregressiontestuponthemigratedsystem

    CSMR200914

    ExpressingSoftwarewithNumbersTheroleofmeasurementinincreasingmutual

    understanding

    Weshouldstrivetoexpresssoftwareintermsofnumbersunderstandabletobothacademiciansandpractitioners

    Thereneedstobeacommonunderstandingofwhatis

    size,whatiscomplexityandwhatisqualityinsoftware. Thereneedstobeametricstandardformeasuring

    quality,complexityandproductivity

    Softwaresystems,developedindifferentmodesandlanguageshavetobecomparable.

    Thereshouldbeestimationtechniquesaimeddirectlyatestimatingmaintenanceandmigrationprojects.

    CSMR200915

  • 7/28/2019 Bridging the Gap Between Academia and Industry - Germany

    9/9

    SuggestionsforfutureCooperation

    Conductcommonconferenceswithadoublestandardforpaperacceptance.

    Firmsshouldinvitelocaluniversitiestoparticipateinindustrialprojects.

    Universitiesshouldoffercontinualeducationprogramsoneveningsandweekends.

    Academiaandindustryshouldstrivetofindacommonvocabularybymeansofstandardization.

    Academiashouldofferindustryasimplesetofmetricswithwhichtheycanmeasuretheirquality,complexityandproductivity.

    CSMR200916


Top Related