Janne Merilinna, Olli-Pekka Puolitaival,

1
Verification and Validation in the Context of DSM

• Janne Merilinna, Olli-Pekka Puolitaival,
• John Menke, Tihamer Levendovszky,
• Jonathan Sprinkle, Mika Karaila,
• Edgars Rencis, Hiroshi Kazato,
• Takashi Kopayashy

2
Table of Context

• Metamodel Testing
• Code Generator Testing
• Model Testing
• Conclusion

3
Metamodel Testing

• Causes of Errors
• Metamodel does not cover the problem domain
• Metamodel does not fit to the framework
• Error categories
• Missing rules
• Missing associations
• Missing constrains
• Missing cardinality
• Missing entities
• How to test the metamodels
• Verification
• We need to have a huge number of test cases, i.e.
  application models
• But we cannot verify the metamodel as such

4
Code Generator Testing

• Causes of errors
• Templates are wrong
• Logic of the code generators is wrong
• Testing
• Metamodel limits the input
• Application models are the input, i.e. the test
  cases for code generator
• We dont want to test only the code generator
• From research point of view this might be
  interesting
• Practice?

5
Model Testing

• Causes of errors
• Incorrect metamodel
• Formal methods
• Transforming application models to something else
  where we can apply formal methods
• Petri Net etc.
• Deadlock
• Structure semantics
• Etc.
• Does these kind of tests really say anything in
  practice
• You still have to generate code and you still
  have your framework on which you generate code
• BUT can you detect something earlier?
• Model-Based Testing?

6
Conclusion

• Sun was shining
• The beer was good
• Loud discussions
• Which was good
• However
• To be continued in the next SPLASH (OOPSLA) DSM10