Second Advisor oUniv-Prof Dr Drhcmult Hermann Maur

AAuuggmmeennttiinnggkknnoowwlleeddggee--iinntteennssiivveessyysstteemmsswwiitthhDDyynnaammiiccPPeerrssoonnaalliizzaattiioonnccoonncceeppttssDollyKandpalAugmentingknowledge-intensivesystemswithDynamicPersonalizationconceptsDissertationfortheAwardoftheAcademicDegreeDoctorofTechnicalSciencesatGrazUniversityofTechnologysubmittedbyDollyKandpalKnow-CenterGrazUniversityofTechnologyAustriaDecember2003©Copyright2003byDollyKandpalFirstAdvisor:Univ.-Doz.Dr.KlausTochtermannSecondAdvisor:o.Univ.-Prof.Dr.Dr.h.c.mult.HermannMaurerAnreicherungwissensintensivSystemedurchdynamischePersonalizierungskonzepteDissertationzurVerleihungdesakademischenGradesDoktorderTechnischenWissenschaftenanderTechnischenUniversitätGrazvorgelegtvonDollyKandpalKnow-CenterTechnischeUniversitätGrazÖsterreichDezember2003©Copyright2003byDollyKandpalDieseDissertationistinenglischerSpracheverfasst.ErsterBegutachter:Univ.-Doz.Dr.KlausTochtermannZweiterBegutachter:o.Univ.-Prof.Dr.Dr.h.c.mult.HermannMaurerIntroduction11IntroductionKnowledgeanditsefficientmanagementconstitutethekeytosuccessandsurvivalfororganizationsinthehighlydynamicandcompetitiveworldoftoday.Efficientacquisition,storage,transfer,retrieval,application,andvisualizationofknowledgeoftendistinguishsuccessfulorganizationsfromtheunsuccessfulones,andareessentialformanagementofknowledge.Eachoftheaboveactivityisinherentlycomplexandposesanumberofchal-lengesfordesignersofknowledge-intensivesystems.Inthecontextofthedynamicityandtheabundanceofknowledgeavailableovervaryingchannelsandmediums,onlysomeofwhichmayberelevanttothecurrentcontextoftheknowledgeworkers,thechallengesbecomeveryimportantandshouldbeaddressedinordertoprovidetherelevantknowl-edgetoknowledgeworkersinordertoenablethemtoperformtheirworkefficiently.Ofallthechallengesfacedbytheknowledgeworkers,twostandoutintermsofrele-vancetothecurrentcontext.Thesetwochallengesfacedbytheknowledgeworkersaredescribedbelow.Thefirstchallengeisposedbythephenomenonofknowledgegap,whichisamajorproblemfacedbyknowledgeworkerswhileusingknowledge-intensivesystems.Asmen-tionedin(Tochtermannetal.,2002),knowledgeisonlyofvalueifitcanbecontextual-izedandappliedinagivencontext.Inthesamearticle,knowledgegaphasbeendefinedasthedifferencebetweentherequiredandsuppliedknowledge.Thisgapbetweentherequiredandsuppliedknowledgeiscausedbyanumberoffactors,thethreemainonesamongthembeing:•Themeta-informationrelatedtotheknowledgeentities,andprovidedbythecontentprovidersistoogenericanduniversal,forexample,differenceinterminologyusedbythecontentprovidersandtheknowledgeworkers,•Outdatedandirrelevantknowledge,causedbytoomuchknowledgeaccumulatinginrelativelyshorttimeperiods,and,moreimportant,toomuchknowledgebeingout-datedinevershortertimeperiods,thusaffectingthequalityandactualityoftheknowledge,Introduction2•Lackofcontextcausedbyveryfewsystemseitherautomaticallycontextualizingtheknowledge,orallowingknowledgeworkerstoarbitrarilycontextualizetheknowl-edgeinthewaytheyneedit.Thesecondchallengeisposedbytheemergenceandincreasingpopularityofubiqui-touscomputing.Theusageofdeviceslikemobilephones,personaldigitalassistants,lap-topandnotebookcomputers;andglobalizationaswellassocialchangesintheworkingenvironmentliketheadventofteleworking,virtualteamsworkingovervarioustime-zonesandgeographicallocationshaveledtotherequirementofprovisionofknowledge-baseaccesstoknowledgeworkersatallplacesandatalltimes.Althoughubiquitouscomputingallowsknowledgeworkerstoaccessknowledgesourcesatalltimes,usingdifferentchannelsandmediums;italsoposesanumberofchallenges.Thedevicesusedbytheknowledgeworkersareveryvariedintermsoftheirdisplaycharacteristics,com-putingpower,memoryavailability,networkcharacteristics,andotheruserinterfaces.Thus,thecontentprovidedtotheknowledgeworkerscannotbeconstant,andmustvarydependingonthetypeofthedeviceonwhichitistoberenderedintermsofdisplaychar-acteristics,typeandsizeofthecontent,inordertobemeaningfultotheknowledgework-ers.Thisinitselfposesanumberofchallenges.Thedevicesorchannelsattachedmustbedynamicallydetected,theircharacteristicsmustbeisolated,andthecharacteristicsofthecontentprovidedmustbedynamicallymatchedtothedevicecharacteristics;andalltheseactivitiesmustideallytakeplacewithoutmajordisruptionofthenormalfunctioningofthesystem.Thechallengesposedbytheabovetwophenomenahavemotivatedthisresearch.Thisthesisarguesthatdynamicpersonalizationprovidesconceptsandsolutionsthataddresstheabovechallenges.Personalization,i.e.theadaptationofasystemaccordingtoauser'sneeds,hasrecentlyattractedalotofattentionnotonlyinresearchbutalsoinindustry.Todayanumberscientificapproachesdealwithstaticpersonalizationbutverylittleworkhasbeendonetoaddressdynamicpersonalizationataconceptuallevel.Also,aninten-sivestudyofexistingknowledge-intensivesystemsinvariousdomainsrevealsthatmanyofthemalreadyofferfeaturesthatsupportdynamicpersonalization.Butfromascientificpointofview,nosoundandpropermodelfordynamicpersonalizationexist.Theobjectiveofthisworkistwofold:thefirstobjectiveistoestablishahomogeneousunderstandinganddefinitionofwhatdynamicper