Ontology 26 enterprise architectureAll slides
Model Futures 2004-2007 Unless Otherwise StatedOntology & Enterprise ArchitectureIan issues with Enterprise ArchitectureExamine some of these issuesOntolo can be a difficult subjectgyjEspecially for those immersed in information > The IDEAS , Process Modelling, Information, Data, Apps, Different levels of abstractionTypes of people (posts), types of organization, types Actual systems, locations, etc.Need a way to
what exists, or what could existtie it all Enterprises have specialist groups and > ..and these can -2gytankmodel
futuresEA & The Real WorldVery easy to say that EA is just about modellingBut, the models tend to describe actual systems, processes, etc.StfEAlidtitdtibidttSome aspects of an EA are livedata e.g.site ttEven models have impact on the daily same real world things may also be represented in other model nts that are also part of the EAIn other words, stratification of model and data is too simplistic an approach for EA EAmodel
futuresOntolo & Information much hype about ontologyBut how much do people really understand about it ?There are two main types of ontology that iITappear Formal ontology usually based on AI concepts of reasoning a purpose
a machine to make decisions.Phill describes the things were interested inThifEAihlkiis the atter, as were seekingto have a common Ontological ConceptsygpUseful to know a few terms bfliia classcarvw golfbefore explaining morea sub-classan individualWhere would the service of serviceSpecialise from processnaturalintentional processserviceitit servicemodel
futuresmy service, performed 13/11/06TaxonomBut surely if we just need common understanding, all we dit?neeis a taxonomy True up to a point, but all we then have are names, with no idea traditional taxonomy e.g.IPSV or the UK Defence Taxonomy is a structure of broader and narrower terms:class (i.e.
type of thing)individual (i.e.specific thing)model
futuresTaxonom & Ontologyygyclasswhole-part relationshipAbove shows a typical taxonomy short-cutAn ontology would distinuish between all of these.In addition, a well-founded ontology can provide ood deal more GrouppConsisting of representatives from the defence dttfftideparts of four countriAustralia, Canada, UK, USA+SddNATOb+ Sen and
as obGoal is to develop a common model for approachSttiffitiiltdtditthtg f princlo ensure common undersge mostReaching down to country-specific definitions whose meaning may need to be understood by other nationsclasses, instances, propertieshigh-level patterns (upper ontology)commonly used relationships:whole-part, sequence, partipation, etc.common objects (agreed taxonomy)internationally accepted terms:person, organization, materiel, etc.national extensionnational extensionnational extensionterminology specific to nations thatwhich may be useful to other nations-e.g.
Bowman, Bradley FV, etc
futuresFoundationThe nations involved were using different modelling (inc.UML Meta-Models)jOntologyAll of these modelling approaches are based on formal as users tend to adopt a given styleThese differences were making it hard to establish a common gapproach between the nations there was too much scope for misunderstanding between afoundational layer (based on IEEE Candidate Upper model
futuresThe Languae of OntologygggyBefore we can begin the training, it is important that hdtdifthtldeveryone as a common undersng of s and language of the known as classes, categories, known as relationships or associations)TipAlways try to ground your thinking by taking it back to concrete examples esp.when dealing with types, work with example > Draw it as a Venn Diagram (notation to follow)model
futuresVenn Diagram NotationgThe typical Venn diagram is slightly enhanced to deal > TypeofTypeAn individual that isan instance of A Subtypeand therefore of A TypeA TypeASubtypeType
TypeType 1Type2Type3VW GolfMy VW GolfA instance of A Subtypebut is an instance of A TypeType2Individual1An individual that is notan instance of A TypeOne of JohPrescotts JagsJohn PrescottIndividual 1Individual BORO methodology for developing business modelsdflbhddlBORO provides a forma, step-by-step methodeveEnsures that there is no scope-creep due to modelling alone ( a way of de-conflicting several stakeholder views (cf > Results in a formal ontologyTends towards simplification establishes high-level patterns slide Crown Copyright 2007The BORO ProcessDoes it exist in space & time ?i.e.can you kick it (now, in the past or in the > can be no debate about semantics.If not, is it a te of Individual, or a type of type of ,yp,ypypindividual ?TYPEWhat is it a type of ? Always trace back down to individuals again, this de-conflicts semantics for slide Crown Copyright 2007GeneralisationThe basic BORO process establishes the ontic categoriesThis is the first essential stageHowever, this has not built us an i.etaxonomyand this is where the difference with BORO becomes , hierarchy that is built is clean.and strange things start to happen high level patterns start tt..some of them are quite surprisingThis slide Crown Copyright 2007What does BORO let us do ?Compare y) The dirtier the data the better works on the assumption that most data quality problems stem from the fact that systems do tdhtthtthhhithdtthnodo what e users want, so they shoe-tactually need.Comparison is guaranteed accurate provided the BORO method and criteria for identity are rigidly followed.Integrate the conversion between themBuild the next generation of systemsImplement the high-level patterns everything else is dataUniversal applicationsThis slide Crown Copyright 2007BORO Example -StandardspWhat is a standard ?A document ?The act of ratifying an agreement ? with each of these cases one by out the typesWork out the tuplesPattern starts to emergeAs more business data from other subject areas is analysed, it btthtStddijtilfbecomes apparent at ari a special ca