from a single ontologically sound conceptual model to multiple physical schema languages bruce t....

1

Separating Semantics and Implementation

From a Single Ontologically Sound Conceptual Model to Multiple Physical

Schema Languages

Bruce T. Bauman, U.S. DoD

2

AgendaMotivationChallengeOntoUMLCompiling

RDFSXSDSQL

SummaryFurther WorkConclusion

3

Model Types

Implementation

Perspective

System

Project

Enterprise

RepresentationNeutral

ProductSpecific

Conceptual

Logical

Optimized PhysicalPhysical

TechnologySpecific

Greater Detail

Fewer Models

Greater

Complexity

Reflects

Semantics

Greater Reuse

4

Model Types - Languages

Implementation

Perspective

System

Project

Enterprise

Neutral Product

Conceptual

Logical

Optimized PhysicalPhysical

Technology

OntoUML

OntoUML with limited encoding annotations

Technology specific tooling

Full use of encoding annotations

5

Why existing languages don’t cut itLanguages pitched at analysis and design

(e.g., UML, ERD)Optimized for design targeting specific

technologiesDon’t have a well defined semantic mapping

Implementation languages (e.g., XML Schema, DDL, OWL, RDFS)Clearly they have made implementation trades.

Natural Language Not precise enough.

6

Unified Foundational Ontology (UFO) Created by Giancarlo GuizzardiFor Conceptual / Analysis modelingOntologically BasedUML class diagram notation

7

Our subset of OntoUMLClasses

<<kind>> <<category>> <<role>> <<roleCategory>> <<dependent>> <<associative>> <<event>>

Associations <<non-dependency>> <<dependency>> <<composition>> <<aggregation>> (merelogical

relations)AttributesDatatypes

<<primitive>> <<domain>>3 <<enumeration>> <<structure>> <<union>>3

8

ClassesConcept Rigidity Identity Existential

DependenceExample

<<kind>> Rigid Unified Independent Person

<<role>> Anti-Rigid Unified Independent Spouse

<<dependent>>

Rigid Unified Dependent Mental State

<<associative>>

Rigid Unified Dependent Marriage

<<category>>

Rigid Dispersive Independent Party

<<roleCategory>>

Anti-Rigid Dispersive Independent Customer

9

The finished product

10

XSD – key assignment

11

Association Encoding - embedTarget

id : stringattrib : integer

Source


associationA B

1..* 1..*

<Source> <Id>Value</Id> <Attrib>Value</Attrib> <Target> <Id>Value</Id> <Attrib>Value</Attrib> </Target></Source>

Include Association: true false Global

AN/A

BNavigable Include Endpoint true false Encoding : embed

12

Association Encoding - reference

Target


Source


associationA B

1..* 1..*

<Source> <Target> <Id>FK Value</Id> </Target></Source>


AN/A

BNavigable Include Endpoint true false Encoding : reference

13

Association Encoding - linkTarget


Source


associationA B

1..* 1..*

<Source><Target href=“”/></Source>


AN/A

BNavigable Include Endpoint true false Encoding : link

14

Association Encoding - partsTarget


Source


associationA B

1..* 1..*

<Source> <Target>FK Value</Target></Source>


AN/A


15



Source


associationA B

1..* 1..*


AN/A


<Source><Association> <Target> <Id>FK Value</Id> </Target> </Association></Source>

16



Source


associationA B

1..* 1..*


AN/A


<Association> <Source>FK Value</Source> <Target>FK Value</Target></Association>

17

Successes / ObservationsIt actually does workUsed on many projectsModel reuse is occurringEncoding rules are sufficient however new

encoding patterns are still being discoveredProjects select very different encodings

18

ChallengesHard but not too hard

Physical Model Duality (encoding can seem like magic)

ToolsReally Hard

TalentCulture

Things to considerReturn on Investment

19

Future WorkBetter visualizationChange managementContinue to improve documentationPrototype other implementationsExplore semi-automatic translation

20

Questions

Bruce BaumanSenior System ArchitectU.S. Department of Defense

[email protected]

from a single ontologically sound conceptual model to multiple physical schema languages bruce t....

Documents