levelssoftmod

NewSupportTechnologiesforEnterpriseIntegrationH.
TedGoranson1,(Ed.
),RolandJochem2,JamesG.
Nele,HervePanetto4,ChrisPartridge5,FrancescaSempereRipol16,DavidShorter7,PeterWebb8,andMartinZelm910/dDominionUniversity,USA;2FhG-IPK,Germany;3NationallnstituteofStandardsandTechnology,USA;4CRAN-ResearchCenterForAutomaticControl,France;'LADSEB-CNR-BORO,Italy;6UniversidadPolitecnicadeValencia,Spain;7/TFocus,UK;8BMTDefenceServicesLtd,UK;9CIMOSAAssociation,Germany,tedg@sirius-beta.
comAbstract:seeQuadChartonpage21INTRODUCTIONThefollowingQuad-Chart(Table1)summarizestheworkofthegroup.
Itidentifiestheapproachtakentoresolvetheissuesandproposesseveralprojectsandideasforfutureworkfortestingandenhancingtheproposedsolutions.
2BACKGROUNDEnterprisemodelingisastrangebeast,becausebyitsnatureitsynthe-sizestwomajortypesofviews:localviewsandaglobalone.
Historically,modelingwasalocalactivity,focusedonindividualprocessesandindivid-ualdomains(activity,information,role,etc.
).
Theseprocessmodels,suitablyconstrained,werethenaggregatedinanorderedfashiontoforma"model"ofthewholeenterprise,usedforinsight,analysisandoptimization.
Theoriginalversionofthischapterwasrevised:Thecopyrightlinewasincorrect.
Thishasbeencorrected.
TheErratumtothischapterisavailableatDOI:IFIPInternationalFederationforInformationProcessing2003K.
Kosankeetal.
(eds.
),EnterpriseInter-andIntra-OrganizationalIntegration10.
1007/978-0-387-35621-1_43348Goranson,HT.
eta/TableI:WorkingGroupQuad-ChartEI3-ICWorkshop4CommonRepresentationofEnterpriseModelsWorkgroup2NewSupportTechnolo-giesforEnterpriseInte-gration2002-February-20/22Berlin,GermanyAbstract:Theworkgroupfocusedonradicalbutprac-ticalstrategiesforgreatlyimprovingenter-prisemodelingandprocessmodelinginanenterprisecontext.
Thegroup'sworkcen-teredonimprovinguserbenefitsinthecon-textofcommonmodels,enterprisecontextandenterpriseviews.
Majorproblemsad-dressedwere:multi-worldviews,softmod-elingandmeta-modelingtheories.
Severaldiscreteresearchprojectswereproposed.
Approach:-UsetheGERAMasreferenceandex-tendwherenecessary-Designfocused,professionallyfacili-tatedworkshopsforaselectedmixofexperts-Seekparticipantswithexpertiseinlogic;situation,type,andgraphtheories;knowledgerepresentation;agentsys-tems;andinternationalstandards-Trytoexploitpriorresearchinrelevantareasnotnonnallyencounteredintheenterprise-modelingcommunityMajorproblemsandissues:-Enterprisemodelingencompassestwomajortypesofviews,localandglobal-Theseviewstendtousedifferenttermi-nology,modelingmethods,andontolo-gies;andoccupydifferentspaceinvir-tualenterprises-Globallymodeledthingstendtobe"soft"andnon-deterministic-Toengineerenterprisesneedsmethod-ologytomergethesoftanddeterminis-ticaspects-TheinfinityoftacitknowledgeneedstobeclassifiableintoaproQoschunksResultsandFuturework:Fivetosixprojectswereidentified:-Softmodeling:suchas,non-determinism,uncertainty,socialandcul-turaldynamics,andtacitknowledge-Introspectivemodeling:sothatmodelscancontrolothermodels-Multi-worldmodeling:Domainssuchaslegal,financial,andproductioncontainconceptualdiscontinuities.
-Multi-levelmodeling:Improvethein-teractionbetweenprocessmodelsandenterprisemodels-Meta-toolsformodeling:tobetterac-complishmulti-levelandmulti-worldmodels-Additiveabstraction:modeltheproc-essesdirectlyintotheenterprisecontextAstimeprogressed,thedisciplineallowedforactualmodelingofenter-prisesasacongruentactivitywiththeaggregationofprocessmodels.
Manyinthefieldbelievethatthefutureofthedisciplineisinthispowertoexplorebothways:tweaktheenterprisemodelandseetheimplicationsattheproc-esslevel,andtheotherwayaround.
Butwiththisvision,oneinheritsacollectionofsuperficiallyrelatedproblemsassociatedwiththedifferentsortsofheterogeneityonefindsinanenterprise.
Someofthesehavebeenwidelydiscussedandpossiblesolutionsareforthcoming:NewSupportTechnologiesforEnterpriseIntegration349-Differentmodelingmethods-Differentontologies-Differentcorporate"zones"inavirtualenterprise(A"zone"maybeafunctionlikemarketing,oraprofitcenterorsupplier.
)-Tosomeextent,differentrolesandgoals.
Butthereareahostofadditionalproblemsthatrealenterprisespresent,whichareapparentlyhardtoaddressbutwhichwouldyieldimmensebenefitifsolvedinsomeway.
Thesearebrieflydiscussedbelow.
Suggestedprojectstoaddresstheseproblemsfollow.
2.
1SoftModelingApreviousICEIMTworkshopaddressedthissetofproblems.
["Ontolo-giesasaNewCostFactorinEnterpriseIntegration"thisvolume]butsub-stantiallymoreabouttheproblemisunderstoodnow.
Severalkindsof"softness"existinanenterprise,andmustbeaccommo-datedinsomewayattheenterpriselevel.
Allofthesesharethequalityofnotbeingrepresentablebydeterministicmodelsbasedonfirstorderlogic.
Nondeterminism(includingnonlinearity):Traditionally,aprocessismodeledpreciselybecauseitisdeterministic,andonewishestounderstandandengineertherevealedmechanics.
Butmanytypicalenterprisesaresuffi-cientlycomplexthattheyincludenon-deterministicprocesses,orwouldliketoeitherincludesuchprocessesormodelthemiftheycouldbemanagedinsomeway.
TheproblemofmodelingNondeterminismhasseveraldimen-sionsthatcanbecollectedunderthenotionof"apparentdeterminism.
"Inthisnotion,somekeyelementsoftheprocessappeardeterministicattheen-terpriselevel(forinstancecost,qualityortimeconstraints)butthedetailsoftheprocessgetmoreunclearasonezoomsdown.
Thisfliesinthefaceofcurrentapproaches,whichdependonorderedaggregationstrategies.
Atsomepointinthedecomposition,anexplicit,semanticallyrich"placeholder"ofthenon-deterministiccausaldynamicsneedstoenterthepicture.
Suchaplaceholderwouldrepresentthatyouknowsomethingbutnotevery-thingaboutanelement,relationshipormechanism.
Uncertainty:Quiteapartfromthesoftnessofnondeterminism(whichconcerns"how"),thereisaanissueofuncertaintyconcerning"what.
"Thisappearsatalllevels,butoftenattheprocesslevelitishandleablebyre-coursetoprobabilities.
Thisisbecauseprocessesareusuallyrepeatableandoftenhavehistoriesthatcanbeincorporatedstatistically.
Butattheenter-priselevel,thereareuncertaintiesofamoreprofound,disruptiveandunpre-dictablenature,suchasnaturaldisasters.
Mostoftheimportantofthesecan-notbeusefullyrepresentedstatistically.
Thereneedstobeawaytorepresent350Goranson,HT.
eta/andreasonaboutobjectsorelementsthatarepartiallyundefinedorlargelyspeculative.
Social/culturaldynamics:Thisisabroadclassthatappearsinseveralformsandappearsatalllevelsoftheenterprise.
Itencompassesthephe-nomenoncoveredbythe"soft"sciences:sociology,anthropology,psychol-ogyandthelike.
Atalllevelsoneoftenneedstounderstandandtosomeex-tenttoengineercollaborativedynamics.
Attheprocesslevelthisisusuallyseenasteamdynamics:attheenterpriselevel,corporateculture.
Alsoin-cludedarepersonnelqualificationandcertificationwithinthecollaborativefabricoftheenterprise.
Butthereismore.
Increasingly,theenterpriseisen-gagedinprovidingcustomersatisfactionthroughvaluesthatthemselvesaresoft,suchasproductstyling,lifestylebrandingordirectsocialupliftinginaserviceorproductinteraction.
Thesesoftdynamicsofcustomermotivationsareahigherorderofsoftnessthatneedscarefulconsideration.
Relatingthesemotivestoproductfeaturesandprocesselementsisarealchallenge.
Tacitknowledge:Thetacitknowledgeproblemisthemostrecognizedofthisclass,whichmakesthelackofsupportatthislatestageratherfrustrat-ing.
Tacitknowledgeisalltheimpliedknowledgethatpartiessharewhentheycollaborate.
Itisamajorbugabooinmodelingbecausethetacitknowl-edgeislargelyunrecognizedandextraordinarilyexpensivetomakeexplicit.
Lackofdoingsomakesmodelsandmodeltransactionsonlyanincompleteshadowoftherealworld.
Inmanycases,thelimitsarefatal.
(Notethatcur-rentmethodsofmanagingthenon-determinisminvolvethecorrectuseoftacitknowledgeto"convert"non-determinismintodeterminism.
)Effectivesoftnessthroughcombinatoricsorothermodeling"holes":Tacitknowledgeisonekindofinformationthatisusuallynotcapturedbymodels,leavingthemincomplete.
Butthereisawholeclassofotherincom-pletenessthatisarealityinmodelinglargeenterprises.
Unlikethethreeitemsabove,theinformationismissingnotbecauseitisunrepresentable.
Butitjustfellthroughthecracks,wasoverlooked,orsomemodelscontainerrors,orsomeintegrationortranslationprocesswasmisapplied,orthemodelissimplyoutofdate.
Or,evenwhenthemodelingoperationisper-fect,sometimesthecombinatorialsofthesituationmakeitimpracticaltocarryeverything,sopruningoccursasamatterofexpediency.
Whateverthecause,therearealmostcertainlylargeholesintheenter-prisemodelofanyreasonablyinterestingenterprise.
Onewouldhopethatinthespiritofmakingthingsexplicit,onecouldrepresentthese"holes"insomewayandreasonoverthem.
Perhapstheonlyutilityisafidelitymetricmeasuringtheaccuracyandcompletenessofthemodel.
Butperhapsagreaterfacilityforself-correctionandguidedcompletionmightbeaffected.
NewSupportTechnologiesforEnterpriseIntegration3512.
2IntrospectiveModelingThisisoneareathatisnotwellsupportedbecauseofaccidentsinthehis-toryofhowenterprisemodelingdeveloped.
Processmodelswerefocusedonspecifictasks,usuallytasksdirectlyassociatedwiththebasicworkoftheenterprise.
Suchprocessesusuallyhave"secondorder"processesassociatedwiththem,processesassociatedwithmonitoringandadaptingthebasicprocessforexample.
Thesecanbelikewisemodeledandmanagedonacase-by-casebasis.
Butenterprisemodelscombineallthese.
Itisdesirabletotreatthemallthesamefromoneperspectivejustforconsistencyinaggregationandman-agement.
Buttheyareofadifferentnature:basicprocessesdotheworkoftheenterprise;whilesecondorderprocessesproducebetterfirstorderproc-esses.
Infact,essentiallyalloftheanalysesappliedthroughenterprisemod-elingareofthissecondordertype.
Theirexistenceisthejustificationofdo-ingenterprisemodelinginthefirstplace.
Anditisoftenthecasethat25%ormoreoftheprocessesintheenterpriseareofthissecondordertype.
Thisconditionwheretheyaretreatedthesamebutaredifferentrequiresthequalityof"introspection.
"Introspectivemodelshavesomewayof"un-derstanding"othermodelsandthereforethemselves.
Despitethebasicre-quirementforintrospection,virtuallynomajormodeltechniquessupportit.
Thisneedstoberemedied.
Introspectionisrequiredforsystemoptimizationthatismanagedjustlikethebasic(firstorder)processesare.
Itisrequiredforwell-behavedstateandconfigurationmanagement.
Itisprobablyrequiredforscalableapproachestotrusted,securesystem.
Itiscertainlyrequiredforanymeasureofautono-mousself-correction.
2.
3MultiworldModelingThemultiworldsproblemisalsonewtoenterprisemodeling,andalsoisoverlookedbecauseofitsorigininfocusesonlocaldomains.
Whenonemodelsashopfloor,onecanassumethatallthebehaviorconformstothesamefundamentallaws.
Forinstance,physicsapplies.
Thesamebasiccorpo-ratepoliciesandexternalregulationsapply.
Therewillstillbelotsof"soft"stuffmissing,butitwillallbemissingfromthesameworld.
Enterprisesarenotsowellbehaved.
Enterprisesarenotlogicalmachines;theyhavelargecomponentsfromdifferentdomainsthatinteractasasystem,andnotallthesedomainsexistinthesame"world.
"Someofthosethatdif-feractuallyhavedifferentcausalmechanics,different"physics"ifyouwill.
Forexample,inthelegalworld,truthoperatesbydifferentrules.
Inthatworld,many"truths"varyfromfact-basedrealityandmuchoftheactivityof352Goranson,HTeta/thosepeopleintheenterpriseistotolerablymaintainthatdistanceforthebenefitoftheenterprise.
{Theworkgroupdidnothavetimetoharmonizethediversedefinitionsoftheuseoftheword"truth,"soitisinquoteshere.
Thedifferentopinionsconcernedlong-standingargumentsoverwhethertruthisbasedinnatureorisafabricatedquality.
Thatthetruthabout"truth"wasunattainableissubmittedasanexampleoftheproblemphenomenon.
Forreadability,thetermshouldbeunderstoodasquotedbelow.
)Hereisanotherexample,theso-called"blackpot"defense.
Amanbor-rowsapotfromhisneighborandreturnsitbroken.
Hisneighborsues.
Thelawyerargues1)myclientneverborrowedthepot;2)itwasbrokenwhenmyclientgotit;3)itwasnotbrokenwhenmyclientreturnedit.
Inmanylegalsystems,theseassertionscanbemadeinanon-exclusivemanner.
(In-cidentally,thisisaprimaryreasonwhyexpertsystemshavenotbeenappliedinthelegaldomain,astheyhave,sayinmedicaldiagnosis.
)Inadditiontoparalleltruthsandtruthsthatarenon-fact-basedbutbasedonconstrainedproofs,onehasotherpeculiaritiesofthelegalworld.
Therearelotsofartificialconcepts,actorsandrelationshipsthathaveno"realworld"counterpart.
Theseincludeownership,intellectualproperty,rights,liabilitiesandvalue.
Oneneedstomodelnotonlythesenotions,butalsotheprocessesweusetoknowandusethesenotions.
Alloftheseareaccommodatedtoacrudeextentinordinaryprocessmodels,appearingasconstraintsandattributes.
Butthatisnothowtheyaremanagedinthelegalworld.
Andtomodeltheenterpriseinatrulyholisticway,thatworldmustbeincluded.
Andthelegalworldisjustthetipoftheotherworldsiceberg.
Everyonehasafavoritehorrorstoryofsomethingthatmadesenseinthefinancialcorneroftheenterprise,butwhen"integrated"intooperationsrequiredprocessestoactinawaythatwasobviouslyharm-ful,orcounterproductiveordownrightstupid.
Marketingandhumanresourceshavesomeoftheseotherworldsdistinc-tions.
Andintheircases,itiscomplicatedbyalsoinvolvingsoftelementsaswell.
Clearly,therearesomesolidoverlapsofworlds:humanresourcesandlegalworldssharethesamerichcausalnotionsofrights.
Legalandfinancialworldssharethesamerichcausalnotionsofownershipvalue.
Butthereareotherareasofmismatches.
Thesearenotamatterofdifferentmodelingmethods,norasimplematterofsemanticconfusion.
Theseare"discontinui-ties"inthebasiccausalstructureoftheworlds,meaningthattherecannotbea"smooth,"coherentsystemofmodelsthatcoversallspaces.
Obviously,thesediscontinuitiescanmakebigtroublewhenintrospectivecontrolisap-plied:forinstanceafinancialcollectionofprocesseschangingphysicalop-erationsandmissingsomeclear"reality"ofthoseprocessesresultinginthehorrorstoriesnotedabove.
NewSupportTechnologiesforEnterpriseIntegration353Mostofthegrouprecognizesthatthisisaveryhardproblem.
Theintentinraisingitisnottosuggestasolution.
Noreasonablyinexpensiveorpracti-calsolutionmayexist.
Rather,thegroupwouldliketosuggestalightweightmechanismthatraisesaflagwhensuchcausaldiscontinuitiesoccur.
Theflaggingofaproblemwouldthrowtheissuetointerventionbyahumanteam.
Theworkgroupwasnotunanimousinunderstandingtheproblemthiswayasalreadynoted.
Analternativeviewsuggestedthattherecanonlybeonetruth,onerealworld.
Theproblemisnotsomuchabouttheworld,butaboutwhatweknoworbelieveabouttheworld.
Engineeringactivitiesusingtraditionalphysicswillprobablyassessthevariouspossiblewaysinwhichthecurrentstateofaffairsmayhavecometopass.
Thereisonlyonewayinwhichtheworldis-sowemayhavetohedgeourbetsastowhichwaythatisandaccommodatearangeofpossibilities.
Inthisperspective,the"blackpot"exampletradesonthedifferencebe-tweenwhatthelaw"believes"andwhatthelawyerbelieves.
Thenotionofartificialconcepts,actorsandrelationshipsopensoneuptoquestionsabouthowwecometoknowandusethingsthatarenot'real'andsocannothaveanycausalrelationshipwiththerealworld.
Sothemultiworldsproblemisthattheyseemlikedifferentworldstothepeople'living'inthem.
Furtheranalysismayshowthatthereisnoconsistentwayoflookingatthemall.
Inanycase,therewasunanimousagreementthatdespitethedifferentdefinitions,theproblemeitherdoesexistorappeartoexistinawaythathasthesameeffect.
2.
4MultilevelModelingThereisanotherproblemintheenterprisethatsuperficiallyisrelatedtothemultiworldsproblem.
Itisbetterbehavedininternal"physics,"andinfactrelativelyeasytodescribe,butitseemstobeadeceptivelythornyprob-leminaggregatingmodelsatvariouslevelsfrombottomtotop.
Atpresent,enterprisemodelingassumesthattheworldeffectivelycon-sistsoftwolevels:thatofprocessesandthatofprocessassembliesthatatsomepointcanbeperceivedasanenterprise.
Strictlyspeaking,anenterprisecanbeseenasalargeprocess,whichcanbesuccessivelysubdividedintosubprocessesuntilanintuitivelyatomiclevelisreached.
Sofarastheotherway,anyreasonablecollectionofprocessescanbethoughtofasanenter-prise.
Andthesesub-enterprisescanbecombined.
Infact,mostapproachestovirtualenterprisesusethisassumption.
Alas,therealworlddisappointsagain.
Actually,enterprisesarenotatallhomogeneouslybehavedatalllevels.
Thewaythatmanufacturingcellsaremanagedandmeasuredisfundamentallydifferentthanthewayaplantis.
354Goranson,H.
T.
eta/Thewaythatadesignteamismanagedandmeasuredisadifferentthingaltogetherfromhowacompanyis.
Thisispartlyaneffectofscale,andpartlyaresultoftheforcesthatcometobear:forinstance,acompanyisdi-rectlyresponsibletofinanciers,whiletheeffectoffinanciermetricsatthedesignteamlevelissubstantiallyfilteredandtransformed.
Inreality,thereare"levels"atwhichthedynamicsandmetricsofthesystemchangeinsubstantialwaysfromthosebelowandabove.
Theselev-els,theirnumberandplacement,willvaryaccordingtosituation.
Often,thesedifferencesarehiddeninthetacitknowledgeofcompanypolicies,whichiswhywearediscoveringtheexistenceoftheselayersasvirtualen-terprisesbecomemorecommon.
Theaggregationmethodsusedincomposingenterprisemodelshavetobemorespecificthanmereagglomeration.
Thisisespeciallythecaseasvir-tualenterpriseparadigmsbecomemoreused,andasEVA(economicvalueadded)metricsareappliedatthelowerlevelsturningthemintoselfishprofitcenters.
2.
5Meta-toolsforModelingAlloftheaboveledtheworkgrouptoconsiderthegeneralclassofmeta-tools.
Thisisabroadcategory,createdonlyforthediscussion.
Itincludes:-Toolstomodeltheenterpriseofenterprisemodeling.
Theobservationwasmadethatifmodelingissogood,whydon'tweuseitourselves-Toolsthatwouldbecomesomeofthereusable"second-order"toolsnotedabove.
Goodcandidatesareautomaticsyntaxanalyzersandgenerators,toolstoexplorenewmodelparadigms,and"interlingua"toolsthatwouldmapbetweensyntaxandsemanticvariants.
-Formalismsandtheoriesthatarealreadymaturebutwhichhavenotbeenwellappliedtothemodelingproblemsdescribedabove.
Exam-plesaresituation,typeandcategorytheories.
Veryneartermactionswerefleshedoutfortheuseofanextendedwork-group.
Theseinvolvedexploringtoolswhichcanmodelothertools,andus-ingtheCIMOSAwebsitetoexposesomeoftherelevanttheoriestothemodelingcommunity.
3SUGGESTEDACTIONSANDPROJECTSThediscussionoftheworkgroupwascenteredonanumberofhighvalueproblems.
Thesewereconsideredinturn,andspecificactionsidentifiedfortests,researchorfurtherdiscussion.
Eachoftheseispresentedbelow,innoparticularorder.
NewSupportTechnologiesforEnterpriseIntegration355Ineachcase,practicalimplementationwasaguideline,withrevolution-arycapabilityasagoal.
Thegroupfeltthatratherthanreinventwhatexists,anewcapabilityshouldbeeasily"insertable"intoexistingtoolsandmeth-ods.
Manyoftheprojectsnotedbelowaddressratherdifficultproblems.
Thegeneralfeelingwasthatacompleteneartermsolutiontomanyoftheseproblemsisunachievable.
Butsometrackingandnotationofthein-adequacywouldbearealstepforward.
Forexample,the"manyworlds"problemofbelowiscertainlynotpracticallysolvable.
Itinvolvesharmoniz-ingincompatiblecausalmechanics.
Ifafacilitymerelyindicatedthattherewasaproblemwithsomeindicationofitsnature,whichwouldbeofsignifi-cantuse.
3.
1Project1:SoftModelingProblem:Thisprojectdevelopsaroadmapforgracefulimplementationofsoftmodelingwithinexistingandemergingmodelingtechniques.
Bothprocessandenterprisemodelswillbeaddressed.
Approach:Projects1through4areproposedwiththesameapproach,thoughtheyshouldberunindependently,inparallel.
Theprojectissuggestedasadirectfollow-ontotheICEIMT,usingthesameadministrativeprocesses.
Internationalexpertswillbeidentifiedforfocused,facilitatedworkshops.
Thesewillbemuchlonger(four-fivedays).
Theparticipantswillbepaidtoproducepositionpapersandattend.
Profes-sionalfacilitationwillbeprovided.
Inotherwords,fortheseprojects,nonewresearchisplanned,merelytheexploitationofpriorresearchinrelevantareasnotnormallyencounteredintheenterprisemodelingcommunity.
Disciplinesofinterestwillbelogicians,andtheoristsinthecategory,situation,typeandgraphtheories,Expertsinknowledgerepresentationinagentsystemswillbeincluded,aswillleadersintherelevantstandardsef-forts.
ExpectedBenefits:Majornationalresearchestablishmentsareawareoftheproblem,butareunsureoftheexpectedbenefitsandpromisingap-proaches.
Thebenefitofawell-structuredresearchroadmapshouldmobilizepentupinterestandproducesomeearly,leverageableresults.
3.
2Project2:IntrospectiveModelingProblem:Thisprojectaddressestheproblemofmodelswhichmonitor,measureandchangemodelsintheenterprisemodelingcontext.
Itfocusesondevelopingasetofprioritiesforimplementation,anorderedlistoftheex-356Goranson,H.
T.
eta/pecteddifficultiesandbenefits,andanindicationofleverageabletechnolo-gies.
Approach:(SeeProject1)ExpectedBenefits:Aswithproject1,thebenefitsofaresearchroadmapshouldbringlighttoacloudyarea.
Awell-definedroadmapisprobablytoomuchtoexpectfromthisproject,giventhediversityofcommunitiesin-volved.
Buttherearelikelymanyreadysolutionsthatmightbeadapted,oncetheproto-roadmapisproduced.
3.
3Project3:MultiworldModelingProblem:Thisprojectconcernsidentifyingthebasicsofthemultiworldproblemasdescribedabove.
Noimmediatesolutionisexpected,insteadathoroughelucidationoftheproblem.
Approach:(SeeProject1)ExpectedBenefits:Thebenefitsofthisprojectarepatternedafterthetwoabove.
Ofthethree,thiswillbeleastmatureintermsofawellorderedresearchplan.
Butimmensevalueisexpectedfromaclearunderstandingoftheproblemitself.
Aswithmodelingingeneral,amodeloftheproblemwillproduceimmediateinsightintopitfallsfortheentirecommunityofenter-prisemodelers,givingthemavirtual"placeholder"forproblemareasthatwillrequireinterventionbyhumanteams.
3.
4Project4:MultilevelModelingProblem:Thisprojectbeginsprogresstowardanorderedunderstandingofthemultilevelproblemdescribedabove.
Thegroupsupposesthatanun-derstandingofthisproblemisanecessarystepinapplyingenterprisemodel-ingininterestingvirtualenterprisecases.
Approach:(SeeProjectl)ExpectedBenefits:Theprojectisexpectedtoresultinademonstrationagentsystem(perhapsopensource)thatexhibitsmultilevelevolution.
Thissystemwillemulatelevelsseenincommontypesofvirtualenterprises.
Thisproblemisexpectedtoone,whichmanycleverenterprisemodelingcenterswilljumpononcefirstresultsaredemonstrated.
Quitepossibly,thesolu-tionsthatemergewillbeinreactiontotheresultofthisprojectinsteadofdirectlybuildingonit.
3.
5Project5:Additive"Abstraction"Problem:Thisprojectaddressesaneednotdescribedabove.
NewSupportTechnologiesforEnterpriseIntegration357Thebasicproblemisthatenterprisemodelingatpresentisatwo-stepprocess:processesaremodeled,andthentheyareaggregated.
Eachisalossystep,involvingabstraction.
Theworkgroupbelievesthatthesetwostepsproducearatherprofoundlossofrelevantbehaviorbecauseabstractingforeffectiveindividualproc-essesprobablyisinadifferent"direction"thanforenterpriseregistration.
Theprojectintendstoproducearoadmaptowardaprocesshumorouslycalled"additiveabstraction,"orcontextualization.
Inthisapproach,proc-essesaremodeleddirectlyintotheenterprisecontext,immediatelyinheritingtheinformationneededforthelocalprocessownertoseerelevantcharacter-isticsoftheenterprise.
Thiswillcertainlyresultinmorecomplexprocessmodelsthanthe"di-rect''processmodelsoftoday.
Buttheexpectationisthattheywillbemoreusefulintheenterprisecontext.
Approach:Oftheprojectsinthisreport,thisistheonlyonesuggestedasa"traditional"project,withadedicatedprojectteamcharteredwithproduc-ingaprototypesystem.
Internationalinvolvementissuggested,inparttotapthelargebutdis-tributedknowledgebase.
Buttheinternationalscopeisalsointendedtoby-passthelong,languidprocessusuallyresultingfromnormalnationalandEuropeanUnionchannels.
Academicswillbeinvolvedfortheirexpertise,buttheprimemoverissuggestedtobealeanandmeancommercialproto-typer.
Theopensystemsprocessissuggestedformaximumtechnologytransferandvisibility.
Theprojectwillexplorefourinter-relatedconcepts:speechacts,metaphor,narrativeand"explanation.
"Speechactsareaformalparsingandorderingoftheinformationex-changedbetweenactiveentities.
Thisisawell-studiedareawhentheinfor-mationiswell-behavedandtheentities"understand"eachotherwell,buttheapproachwillbeextendedforthefourcasesaddressedintheprojectsabove.
Theresultingspeechactperformativeswillguidetheexplorationintometaphor.
Theuseofmetaphorisexpectedtobeanaidincontextualizingtheinformation.
Thefocusoftheexplorationintometaphorwillbeonnota-tionsthatconveycertaincontextualpatternswhilemakingthatcontextcleartonon-specialists.
Boththeperformativesandmetaphoriccontextwillsuggestmeansforbuildingstructurednarrative.
Thisnarrativeisexpectedtoprovidethestruc-tureforanannotationstrategythatwillprovidetheexplanationsdesired.
Thetargetapplicationisasastructuredannotationlanguagetoaddtoex-istingmodelingenvironments.
Thenotionistoallowexistingmethodstosupportwhattheyalreadydo,andreverttoattachedannotationsforthoseareas(soft,multiworld.
.
.
)whichtheydonot.
Theprojectwillfocusonthis358Goranson,HT.
eta/annotationlanguagetoincrementallyimprovetheinformationandcontextcapturedinastructuredcomputableway.
Multimediadeliveryandpresentationmethodswillbeexplored.
ExpectedBenefits:Thedirectbenefitofthisprojectwillbetodefineanewapproachtomodeling:forenterprisemodelsratherthancomposableprocessmodels.
Thisisexpectedtohitasweetspotintheclevernessofthecommunityandproducearashofnewmodelingapproaches.
DirectinputtothedevelopingProcessSpecificationLanguageandUnifiedEnterpriseMod-elingLanguageprogramsisexpected.
3.
6Project6:DefinitionsProblem:Thisprojectworkstowarddefinitionsofenterpriseandbehav-iorintheexpandedcontextidentifiedbytheworkgroup.
Theimpetusfortheprojectgrewoutofanunexpecteddifficultythegrouphadinclosingonsuchdefinitions.
Thefundamentalproblemseemstobeamatterofthesimultane-ityofpurposes:enterpriseviewtooptimizethesystemthroughprocesses,andprocessviewofenterprisecharacteristicstoimprovetheenterprisethroughincrementalaction.
Approach:Thisisproposedasavirtual,web-basedproject.
Itwilllever-agethemanypeoplealreadyfundedtoworkinthisgeneralarea.
Theonlycostswillbeprofessionalfacilitation,whichneedbesubstantialbecauseofthecontinuingneedforfocusingandrefinement.
Draft,web-basedresultsandemail-baseddiscussionwillfeedtheproc-ess.
Totalconvergenceonconsensuspositionsisnotexpected.
Consensuswillbeencouragedwherepossible.
Otherwise,acrispdescriptionofcon-trarypositionsissought,togetherwithunderlyinganalysisofthephiloso-phiesbehindthedifferences.
ExpectedBenefits:Theresultofthisprojectwillbedirectlyfedtostan-dardsefforts,producingandexpectedgreaterunderstandingoftheuniquechallengesofcommonenterprisemodelrepresentation.