60 likes | 162 Views
What can BX make from Sense?. Arend Rensink, University of Twente Banff, December 2013. Context: EU-project SENSATION. models. SDF3 (Synchronous Data Flow). UPPAAL (Timed Automata). results. s df3 xml (w. xsd). u ppaal xml (w. xsd). Not well-structured, too large.
E N D
What can BX make from Sense? Arend Rensink, University of TwenteBanff, December 2013
Context: EU-project SENSATION models SDF3(Synchronous Data Flow) UPPAAL(Timed Automata) results sdf3xml(w. xsd) uppaal xml(w. xsd) Not well-structured, too large Lot of plain-text elements senseecore uppaal ecore What can BX made from Sense?
Dataflow graphs • Elements • Actors (nodes): name • Ports: name, direction (in, out), rates • Channels (edges): name, initial tokens, source/target port • Processors with execution time for subset of actors • Petri Net-like semantics • Actors = transitions, channels = places, rates = weights • Non-atomic transition firing (actors take time) What can BX made from Sense?
Challenge • Language differences SDF3Sense Many spurious elements Dedicated language Source/target ports Source/target actors References all by name References by identity Processor type implicit Processor type explicit • Bidirectionality • Sense-to-SDF3 should be (provably) left inverse to SDF3-to-Sense • Executability • Transformation definition must be executable • Underlying technology • XSD-backed XML versus Ecore • Simplified option: Ecore-euivalent on SDF3 side Is there a solution that allows a proof of bidirectionality, and is directly executable on concrete models? What can BX made from Sense?
SDF3 What can BX made from Sense?
Sense metamodel What can BX made from Sense?