Combining ontology-enriched Domain-Specific Languages

Combining ontology-enriched Domain-Specific Languages Workshop on Transforming and Weaving Ontologies in Model Driven Engineering (TWOMDE) 04.10.2009, Denver, Colorado Tobias Walter Jürgen Ebert

Objectives
Motivation
Scenario
Foundations
Metamodels
Description Logics
Integration
enriching DSLs by ontology languages
combining ontology-enriched DSLs
Conclusion

Integration Scenarios BEDSL Metamodel OWL Metamodel transform BEDSL+OWLMetamodel + BEDSL+OWL Model BEDSL+PDDSL+OWL Model transform BEDSL+PDDSL+OWL Metamodel PDDSL Metamodel OWL Metamodel PDDSL+OWLMetamodel + transform + PDDSL+OWL Model + Scenario 1 (WCDSL Paper) Scenario 2 (here)

Scenario 1
Integration of DSLs and ontology languages allows:
Using the structural DSL to build domain models
Creating additional (local or global) annotations to define constraints and axioms to restrict the usage of concepts in the model and to allow reasoning
Comarch Domain Model:

Scenario 2
Integration of ontology-enriched DSLs
Combination of domain models
Domain models contain concepts with instances
Combination constraints and semantics
merge model elements merge model elements + constraints Comarch Domain Model BEDSL model PDDSL model

Metamodel (BEDSL)
BEDSL – Business Entity Domain Specific Language
BEDSL models (M1 layer) conform to metamodel (M2 layer)
M2 layer M1 layer BEDSL Metamodel BEDSL Model conformsTo conformsTo

Metamodel (PDDSL)
PDDSL – Physical Device Domain Specific Language
PDDSL models (M1 layer) conform to metamodel (M2 layer)
M2 layer M1 layer PDDSL Metamodel PDDSL Model conformsTo conformsTo

Description Logics
Description Logics (DLs) are logics designed to represent and reason on structured knowledge
The domain of interest is structured into ( TBox ):
concepts, which correspond to classes, and denote sets of individuals
roles, which correspond to associations, and denote binary relations on individuals
The knowledge is asserted through so-called assertions ( ABox )
DLs provide the foundations for standard ontology languages, like OWL2

Example (OWL2)
Class: Device
Class: Cisco7603
SubClassOf: Device,
hasConfiguration exactly 1 Configuration7603
Class: Configuration
Class: Configuration7603
SubClassOf: Configuration
ObjectProperty: hasConfiguration
Domain: Device
Range: Configuration
M1 layer

Metamodel of OWL2 (Excerpt) M2 layer OWL2 Metamodel

Integration (BEDSL + OWL)
Combine BEDSL Entity and OWL Class
M2 layer M1 layer conformsTo conformsTo

Integration (PDDSL + OWL)
Combine PDDSL Element and OWL Class
M2 layer M1 layer conformsTo conformsTo

Combination of ontology-enriched DSLs
Combination of BEDSL+OWL language and PDDSL+OWL language
Allows for multi-viewpoint modeling of business entities and device configurations
Integration of:
Metamodels at M2 layer
Conforming models at M1 layer
Semantics and constraints of domain models
Problem:
How and when to combine domain model elements?
How to combine constraints and semantics?

Transformation-based Integration
Transformation at M2 layer
Define which metamodel elements are combined
Corresponding Transformation at M1 layer
Define which conforming model elements are transformed
Pre-condition of applying the transformation on A and B
Extracted from additional intensional semantics of domain model elements
(e.g. A and B are equivalent domain concepts and can be merged)

Merge Transformation
Merge Transformation
Merge of:
metaconcepts MA and MB to MC (at M2 layer)
concepts A and B to C , if the following pre-condition holds:
Constraints of merged model elements are transformed with regard to the precondition
Class: A EquivalentWith: B

Specialization Transformation
Specialization Transformation
Create specialization relationship between metaconcepts MA and MB (at M2 layer)
Concepts A and B are transformed, if the following pre-condition holds:
Constraints of model element B are transformed with regard to the precondition
Class: A SubClassOf: B

Example (ok)
Merge of BEDSL Entity and PDDSL Element
Transformation at M2 layer:
Tranformation at M1 layer:
1) merge model elements 2) merge constraints

Example (not ok)
Merge of BEDSL Entity and PDDSL Element
Transformation at M2 layer:
Tranformation at M1 layer:
1) merge model elements 2) merge constraints unsatisfiable class

Conclusion
Combination of metamodels and domain models: Combining modeling languages with existing models and constraints
Integration at M2 layer: Transformations define how to combine metamodels
Integration at M1 layer:
Transformations of model elements to target model
Transformation and recalculation of constraints of model elements in the target model
Usage of reasoning technology to verify model correctness

Finally
Thanks for your attention
supported by
www.most-project.eu

Combining ontology-enriched Domain-Specific Languages

by Tobias Walter on Feb 06, 2011

Accessibility

Tags

Upload Details

Usage Rights

Statistics

Combining ontology-enriched Domain-Specific Languages — Presentation Transcript