linked data techniques for mof compliant models

Post on 26-Jan-2015

112 Views

Category:

Education

1 Downloads

Preview:

Click to see full reader

DESCRIPTION

 

TRANSCRIPT

Web Science & Technologies

University of Koblenz ▪ Landau, Germany

An Ontology-based Approach for Enabling Linked Data Capabilities

to MOF Compliant Models

Fernando Silva Parreiras, Gerd Gröner, Tobias Walter

Gerd Grönergroener@uni-koblenz.de 2

WeST ICEIS 2012

Model Driven Engineering

Standards for• General purpose languages• Domain specific languages

Abstraction principle

Transformations

Specialization /Inheritance

➔ Engineering relies on a variety of languages (to describe models)

Gerd Grönergroener@uni-koblenz.de 3

WeST ICEIS 2012

… how to deal with the variety of languages?

ecosystem of variousmodeling languages→ islands

✔ different languages for different modeling purposes✔ different views on systems

BPMN UML AD

DSL1

GRL

DSL2

✗ information request across models of multiple languages✗ relationships between languages (e.g., BPMN and UML)

Gerd Grönergroener@uni-koblenz.de 4

WeST ICEIS 2012

What is missing so far?

BPMN

UML AD

DSL1 GRL

DSL2

Gerd Grönergroener@uni-koblenz.de 5

WeST ICEIS 2012

What is missing so far?

BPMN

UML AD

DSL1 GRL

DSL21. links / connections

2. understand links

Gerd Grönergroener@uni-koblenz.de 6

WeST ICEIS 2012

What is missing so far?

BPMN

UML AD

DSL1 GRL

DSL21. links / connections

2. understand links

3. add or extend languages

FD

Gerd Grönergroener@uni-koblenz.de 7

WeST ICEIS 2012

What is missing so far?

BPMN

UML AD

DSL1 GRL

DSL21. links / connections

2. understand links

3. add or extend languages

FD

4. views (or extractions) → querying

Gerd Grönergroener@uni-koblenz.de 8

WeST ICEIS 2012

Final Goal

➔ Obviously, it is hard and probably undesired to have a single unified modeling language

➔ Instead: build and enable a networked and federated development architecture

explore and query over models of

multiple languages

Perspective:

connect andlink

“Web of models”

engineer

Gerd Grönergroener@uni-koblenz.de 9

WeST ICEIS 2012

Idea

Representation:

Expressive languages with well defined semantics

Use existing Semantic Web technologies

Integration:techniques to identify correspondences

best practice (e.g., Linked Data):design, publish and consume data on the Web

➔ adopt technologies ➔ learn from best practice

Gerd Grönergroener@uni-koblenz.de 10

WeST ICEIS 2012

Outline

1) Introduction

2) Background

3) Integrating MOF and OWL

4) Conclusion

Gerd Grönergroener@uni-koblenz.de 11

WeST ICEIS 2012

Web Ontology Language (OWL)

OWL DL (or OWL 2 DL)➔ Compatible with RDF (and LOD principles)➔ Modeling “layers”:

TBox:Classes (atomic and complex), Properties (object and datatype properties)

ABox:Individuals

powerful for describing

domain concepts in networked

environments

➔ Approaches to establish mappings and correspondences between MOF (meta-) models and OWL models

Gerd Grönergroener@uni-koblenz.de 12

WeST ICEIS 2012

… integrate OWL to MOF models

➔ represent and understand metamodels, models and relationships

syntax→ additional modeling

constructs

solid and grounded semantics

➔ techniques to identify equivalence or similarity

syntax and semantics

➔ connect / relate elements of models and metamodels

languageconstructs

Gerd Grönergroener@uni-koblenz.de 13

WeST ICEIS 2012

Outline

1) Introduction

2) Background

3) Integrating MOF and OWL

4) Conclusion

linkingprinciples (with OWL constructs)

extendingMOF

mappingMOF → OWL

integrationmanagement

Gerd Grönergroener@uni-koblenz.de 14

WeST ICEIS 2012

Linking MOF Metamodels with OWL Constructs

EquivalentClass (uml:Activity bpmn:BpmnDiagram)

EquivalentClass (uml:ActivityNode bpmn:Vertex)

TransitiveObjectProperty(uml:general)

SubObjectPropertyOf( ObjectPropertyChain

(bpmn:outgoingEdge bpmn:target) uml:successorNodes)

EquivalentObjectProperties

(uml:successorNodes bpmn:successorActivities)

Gerd Grönergroener@uni-koblenz.de 15

WeST ICEIS 2012

Extending MOF

EMOF ➔ every Object is an EObject

(EClass, EReference, EAttribute, ...)➔ build on EAnnotations

Realization steps:

1. Use EMOF metametamodel to represent metamodels

2. Annotate (enrich) metamodels● EReference → ObjectProperty (OWL)● EClass → Class (OWL)

3. Define additional constructs➔ equivalent classes

Gerd Grönergroener@uni-koblenz.de 16

WeST ICEIS 2012

Example: UML and BPMN metamodel

linking

additionalOWL

constructs

→ OWL annotations

Gerd Grönergroener@uni-koblenz.de 17

WeST ICEIS 2012

Mapping MOF to OWL

Gerd Grönergroener@uni-koblenz.de 18

WeST ICEIS 2012

Mapping MOF to OWL – Example

Gerd Grönergroener@uni-koblenz.de 19

WeST ICEIS 2012

Integration Management

Problem:

reconcile resources that are expressed by different modeling languages

➔ techniques from ontology alignment

➔ identify correspondences

Linking

Align-ment

1. syntactic matchinge.g., uml:Activity ↔ bpmn:BpmnDiagram

Gerd Grönergroener@uni-koblenz.de 20

WeST ICEIS 2012

Integration Management

Problem:

reconcile resources that are expressed by different modeling languages

➔ techniques from ontology alignment

➔ identify correspondences

Linking

Align-ment

1. syntactic matching

2. relational structure comparison

Gerd Grönergroener@uni-koblenz.de 21

WeST ICEIS 2012

Integration Management

Problem:

reconcile resources that are expressed by different modeling languages

➔ techniques from ontology alignment

➔ identify correspondences

Linking

Align-ment

1. syntactic matching

2. relational structure comparison

3. extensional techniques

Gerd Grönergroener@uni-koblenz.de 22

WeST ICEIS 2012

Integration Management (2)

Querying MOF-models

➔ SPARQL: W3C standard query language (primarily for RDF)

➔ SPARQLAS: query language that allows OWL syntax

• more readable queries• translated to SPARQL

✔ Query both levels: model and metamodel

(RDF representation)

(more expressive OWL representation)

Gerd Grönergroener@uni-koblenz.de 23

WeST ICEIS 2012

SPARQLAS – Query Example

What Use Cases do I have to test If I update the component west.twouse.reasoner?

Namespace: uml = <http://www.eclipse.org/uml2/3.0.0/UML#>Namespace: srs = <http://west.uni-koblenz.de/SRS#>Namespace: mf = <http://west.uni-koblenz.de/EclipseManifest#>

Select ?nameWhere: ?component mf:name "west.twouse.reasoner"^^xsd:string ?component srs:requirement ?requirement ?requirement srs:useCase ?uc ?uc uml:name ?name

Gerd Grönergroener@uni-koblenz.de 24

WeST ICEIS 2012

Conclusion

➔ Semantic Web technologies for linking and querying software models

ecosystems of variousmodeling languages→ islands

Situation:

1. OWL constructs (e.g., sameAs, EquivalentWith) to connect/link concepts and instances of different languages

2. OWLizer for MOF-based modes

4. SPARQLASfor querying multiple software models and metamodels

3. find correspondences

Gerd Grönergroener@uni-koblenz.de 25

WeST ICEIS 2012

Conclusion (2)

Techniques and practices from the Semantic Web:

➔ Intended for linking and integrating data ➔ Efficient processing

➔ vision: “web of models”

Gerd Grönergroener@uni-koblenz.de 26

WeST ICEIS 2012

Conclusion (2)

Techniques and practices from the Semantic Web:

➔ Intended for linking and integrating data ➔ Efficient processing

➔ vision: “web of models”

top related