RoadmaptothesemanticwebTheclientofuseraneedsdatafromatravelagency.Howcanitgetit?Heneedsinstructionsthatenablehimtocontactatravelagency.Thisimpliesthatthereisastandardnamefortravelagencies.Thisnameisfournishedbythe“WebServicesOntology”.Sotheclientcouldsendsomethinglike:?whoweb_services:serviceweb_services:travel_agent.usingthestandardwebservicesontology.Thisqueryreturnsalisttravelerservicesstillinthewebservicesontology.Thisdatasentmightcontainarulesparte.g.fordisplayingamenutotheuser.Thefollowingstandardsarealreadysupposedhere:RDF,astandardruleslanguage,astandardontologylanguage=OWL+itsextensions(astandardwebservicesontology)and,lastbutnotleast,standardactions(like“displayonscreen”or“sendaquery”).Aqueryissentusingabasicontology(mightbe“theuniversalcore”)(anotherstandard).TheresponseusesalsoOWL,Universalcore,ruleslanguageandanOWLextensionspecificallydesignedfortravelagencies.Theresponsecontainsananswertothequerybutalsothepossiblerulesthatpermit(orinstruct)theagentoftheusertocontactfurtherserviceslikethedifferentrailwayentitiesorflightcompanies.Sotheclientoftheuserdoesnotneedtointerpretitselftheontologyofthetravelagency;thenecessaryrulesandtravel_agency_ontologyaresenttohim(orqueriesandanswersaresentbackandforthovertheinternetdependingonwhatismostefficientorpractical).OfthenecessarystandardsRDFandOWLexistnow.EfficiencyofinferenceimplementationWhenseekingtounifyagoalitispossibletotakeeveryclauseofthedatabaseandseewhetheramatchispossible.Thisclearlyisnotefficient.Betterismakeadictionaryofallatomsandthenumberoftheclauselthatcontainthoseatoms.Thenbeforeunificationalistofpossibleclausesisretrievedfromthedictionary(whocontainsalsotheaddressoftheclauses)andtheunificationisdone.Itisalsopossibletoputtheclausesinarelationaldatabase.IntroducingprogramminglanguagefeaturesinNotation3:Thecasestudyaboutthetravelagentshowsthattheinferenceenginehasacomplextasktoaccomplish:determiningpathsandschedulingitineraries.Though,nodoubt,alotcanbeaccomplishedusingfactsandrule