Model Driven Software Development Project (Elite Graduate Program, Summer Term 2017). University of Augsburg / Technical University of Munich / Ludwig-Maximilian University.

Implementation of: Hennicker, Rolf, and Alexander Knapp. "Activity-driven synthesis of state machines." Fundamental Approaches to Software Engineering (2007): 87-101.

1. Install the latest version of Eclipse Modelling Tools: https://eclipse.org/modeling/
2. Open up Eclipse and click Help->Install Modelling Components and install the following components

• Acceleo
• Operational QVT
• Sirius
• XText
• Papyrus
• OCL Tools

3. Clone the repo and import it by right clicking on your Project Explorer -> Import -> General-> Existing Projects into Workspace. Browse the SSDToStateMachine Project and import the project. Make sure that the Copy projects into workspace is NOT checked.
4. If your IDE cannot resolve the custom defined URIs (in our case something like mm:...) then right click on the project (SSDToStateMachines) -> QVT Settings -> Metamodel Settings. Click Add and make sure the following three metamodels point to the correct .ecore files:
   • mm:sequequencediagram should point to SSDToStateMachine/metamodels/sequencediagram.ecore
   • mm:behaviour should point to SSDToStateMachine/metamodels/behaviour.ecore
   • mm:ioautomaton should point to SSDToStateMachine/metamodels/ioautomaton.ecore

In order to run the project you will need to create a run configuration.

1. After you've imported the project, click on the Run arrow symbol in Eclipse and select Run Confirgurations
2. On the left side search for Operational QVT Interpreter and double click on it to create a new run configuration
3. Set the transformation module to platform:/resource/SSDToStateMachine/transforms/0Transformation.qvto
4. In the Transformation parameters section set the IN Model to platform:/resource/SSDToStateMachine/models/atm.uml . This file contains the two scenarios that are described in the paper.
5. In the Transformation parameters set the OUT Model to platform:/resource/SSDToStateMachine/result/Transformation.uml The generated model will be written to this file.

• Jan Göbel (slashburn)
• Katja Ludwig (kaulquappe23)
• Jens Wöhrle (SweetyGott)
• Alexander Knapp (supervisor)

• State Machine design: In each state, there is only one method in the do activity. If n actions are done in the do activity (as described in the paper cf. pp. 98), a chain of states is generated. This is an implementation detail and simplifies recursive calls. This doesn't change the state machine's behavior.
• User message (consistent with paper): An arbitrary actor (user or system actor) is sending a message to a user (in our implementation the user could also send a message to himself)
• System message (consistent with paper): An arbitrary actor (user or system actor) is sending a message to a system actor (e.g. atm, bank, consortium)
• User and system communications are distinguished by their message type (asnc -> usermessage, sync -> system message)

Check out the following tutorials to learn more about QVTO:

• http://redpanda.nl/index.php?p=tutorial1
• Nolte, Siegfried. QVT-Relations Language. Springer Science & Business Media, 2009.