Modelsruntime 2007 October 2 2007 Nashville Coherent Support

  • Slides: 11
Download presentation
Models@runtime 2007, October 2, 2007, Nashville Coherent Support for Models at Run-Time through Orthogonal

[email protected] 2007, October 2, 2007, Nashville Coherent Support for Models at Run-Time through Orthogonal Classification Colin Atkinson Matthias Gutheil [email protected] uni-mannheim. de [email protected] uni-mannheim. de Chair of Software Technology University of Mannheim, Germany http: //swt. informatik. uni-mannheim. de

Overview 1. Motivation 2. Current OMG Modeling Infrastructure 3. The Orthogonal Classification Architecture 4.

Overview 1. Motivation 2. Current OMG Modeling Infrastructure 3. The Orthogonal Classification Architecture 4. Pan Level Model Metamodel 5. Example 6. Conclusion 2

Motivation ■ In mainstream software development the word “model” traditionally refers to a description

Motivation ■ In mainstream software development the word “model” traditionally refers to a description of classes of objects and their properties ■ ■ the main motivation for deploying such models at run-time is to be able to check whether objects confirm to the constraints on their classes This means that instances, which have traditionally had only halfhearted support in mainstream modelling environments, need to be fully and cleanly supported ■ ■ ■ instances no longer treated as second class citizens extend the general usage of the term model to include objects as well as classes full notational support for objects ■ databases should be populatable through “models” of objects ■ ■ full run-time support for constraint checking … 3

Current OMG Modeling Infrastructure UML 2. x, MOF 2. x 4

Current OMG Modeling Infrastructure UML 2. x, MOF 2. x 4

Orthogonal Classification Architecture (OCA) Core language Pan Level Model L 2 Classifer O 0

Orthogonal Classification Architecture (OCA) Core language Pan Level Model L 2 Classifer O 0 O 2 O 1 Modeling language Instance Class L 1 Profession L 0 instance. Of Professor instance. Of Einstein Model System state or “real world” 5

Characteristics of the OCA ■ Unification of Class and Object Concepts ■ ■ Clabjects

Characteristics of the OCA ■ Unification of Class and Object Concepts ■ ■ Clabjects – unfied class/object Strict Metamodeling ■ Every Clabject is an instance of a Clabject from the level above ■ Exception: ■ Top ontological level ■ Level agnostic notation ■ Unifrom represenation of associations/links and attributes/slots ■ Dual facet of relationships ■ Exploded and imploded forms ■ Deep Instantiation ■ All elements with potency 6

Pan Level Model Metamodel (partial) 7

Pan Level Model Metamodel (partial) 7

Example Class Job. Relation potency = 0 potency = 2 Salary potency = 2

Example Class Job. Relation potency = 0 potency = 2 Salary potency = 2 type : double value = undefined Profession potency = 2 level 0 Salary potency = 1 type : double value = undefined Professor teaches Course potency = 1 level 1 P L M Salary potency = 0 type : double value = 4567 Einstein teaches Physics potency = 0 level 2 8

Imploded rendering of Connectors Class Job. Relation potency = 0 potency = 2 Salary

Imploded rendering of Connectors Class Job. Relation potency = 0 potency = 2 Salary potency = 2 type : double value = undefined Profession potency = 2 level 0 Salary potency = 1 type : double value = undefined Professor Course potency = 1 level 1 P L M Salary potency = 0 type : double value = 4567 Einstein Physics potency = 0 level 2 9

Conclusion ■ The OCA framework ■ supports multi-level modeling ■ strict metamodeling ■ deep

Conclusion ■ The OCA framework ■ supports multi-level modeling ■ strict metamodeling ■ deep instantiation ■ ■ can be used as a platform for a model-based runtime system ■ is the basis for the integration of ontology and modelling technologies Next challenges are ■ design the Metamodel of the TAQL (Transformation, Action and Query Language) ■ implementation ■ different concrete syntax for different model elements 10

Finally ■ Thank you for your attention! ■ Time for your questions. . .

Finally ■ Thank you for your attention! ■ Time for your questions. . . (in 5 minutes) 11