$30 off During Our Annual Pro Sale. View Details »

The many meanings of UML 2 Sequence Diagrams: a...

The many meanings of UML 2 Sequence Diagrams: a survey

Z. Micskei, H. Waeselynck. "The many meanings of UML 2 Sequence Diagrams: a survey", SoSyM 10:4, 2011, Most Influential Paper award @ MODELS 2021

Scenario languages are widely used in software development. The complexity of the language (and the diversity of the goals Sequence Diagrams are used for) yields several possible choices in its semantics. This paper collects and categorizes the semantic choices in the language, surveys the formal semantics proposed for Sequence Diagrams, and presents how these approaches handle the various semantic choices.

Link: http://dx.doi.org/10.1007/s10270-010-0157-9

More Decks by Critical Systems Research Group

Other Decks in Research

Transcript

  1. The many meanings of UML 2 Sequence Diagrams: a survey

    Hélène Waeselynck MODELS 2021 SoSyM 10-years impact paper Zoltán Micskei mit.bme.hu/~micskeiz homepages.laas.fr/waeselyn/ Budapest University of Technology Budapest, Hungary LAAS-CNRS Toulouse, France micskeiz
  2. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper Working on

    testing mobile systems Birth & timeline 4 2006 Motivating example: a group membership protocol based on spatial proximity
  3. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper Working on

    testing mobile systems A new test language for test scenarios Birth & timeline 5 2006 2007 Being at communication range, But not at a safe distance (RangeNotSafe) Being at a safe distance (Safe) Config. C1 Config. C2 n4 n3 n2 n1 n2 n1 n4 n3 Being at communication range, But not at a safe distance (RangeNotSafe) Being at a safe distance (Safe) Config. C1 Config. C2 n4 n3 n2 n1 n2 n1 n4 n3 Being at a safe distance (Safe) Config. C1 Config. C2 n4 n3 n2 n1 n2 n1 n4 n3 hello from 2 n1 n2 n4 GetLeader LeaderAddress GroupInfo GroupChange GroupChange C1 C2 Config changes from C1 to C2 hello from 2 hello from 2 hello (Safe, RangeNotSafe) n3 GroupChange Failure hello from 2 n1 n2 n4 GetLeader LeaderAddress GroupInfo GroupChange GroupChange C1 C2 Config changes from C1 to C2 hello from 2 hello from 2 hello (Safe, RangeNotSafe) n3 GroupChange Failure Spatial view Event view A fail scenario for concurrent split & merge Does this fail scenario occur in the test trace? 1. Detect the occurrence of the spatial configurations → graph matching problem 2. Aanalyze the order of events in the identified configurations → UML SD Semantics Was expected to be a minor aspect of our work…
  4. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper Working on

    testing mobile systems A new test language for test scenarios Birth & timeline 6 2006 2007 Check existing semantics and select one (few weeks?) 2011 Semantics paper published in SoSyM 10:4
  5. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper Working on

    testing mobile systems A new test language for test scenarios Birth & timeline 7 2006 2007 Check existing semantics and select one (few weeks?) New elements in UML 2.0 Still many questions “…constructs open up a veritable pandora’s box” S. Pickin
  6. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper Working on

    testing mobile systems A new test language for test scenarios Birth & timeline 8 2006 2007 Check existing semantics and select one (few weeks?) 2008 Collecting 12 works: problems and detailed examples Tech report & submitting to SoSyM (July)
  7. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper Working on

    testing mobile systems A new test language for test scenarios Birth & timeline 9 2006 2007 Check existing semantics and select one (few weeks?) 2008 Collecting 12 works: problems and detailed examples Tech report & submitting to SoSyM (July) “I regret to inform you that…” (Sept) “We encourage you to significantly revise the paper and resubmit as a new manuscript to SoSyM.”
  8. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper (Many thanks

    to them!) Feedback from reviewers 10 I am not sure that I always agree with the authors when they call [...] "problems" Whether or not this is "counter intuitive" [...] is debatable in several of the formalizations […], this is not considered to be a problem. It seems to be assumed that interactions should [always] have a (direct) operational interpretation • The discussion of the semantics was too much biased by our initial objective (analysis of test traces) • The paper was a catalogue of problems. A catalogue of solutions would be more helpful!
  9. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper Working on

    testing mobile systems A new test language for test scenarios Birth & timeline 11 2006 2007 Check existing semantics and select one (few weeks?) 2008 Collecting 12 works: problems and detailed examples Tech report & submitting to SoSyM (July) “I regret to inform you that…” (Sept) 2009 Change perspective, rework paper & submit again Focus on choices and options instead of problems
  10. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper Working on

    testing mobile systems A new test language for test scenarios Birth & timeline 13 2006 2007 Check existing semantics and select one (few weeks?) 2008 Collecting 12 works: problems and detailed examples Tech report & submitting to SoSyM (July) “I regret to inform you that…” (Sept) 2009 Change perspective, rework paper & submit again Good news! Minor revision, revise paper
  11. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper Working on

    testing mobile systems A new test language for test scenarios Birth & timeline 14 2006 2007 Check existing semantics and select one (few weeks?) 2008 Collecting 12 works: problems and detailed examples Tech report & submitting to SoSyM (July) “I regret to inform you that…” (Sept) 2009 Change perspective, rework paper & submit again Good news! Minor revision, revise paper 2010 “It is a pleasure to accept your manuscript…” 2011 Paper published in SoSyM 10:4
  12. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper Working on

    testing mobile systems A new test language for test scenarios Birth & timeline 15 2006 2007 Check existing semantics and select one (few weeks?) 2008 Collecting 12 works: problems and detailed examples Tech report & submitting to SoSyM (July) “I regret to inform you that…” (Sept) 2009 Change perspective, rework paper & submit again Good news! Minor revisions, revise paper 2010 “It is a pleasure to accept your manuscript…” 2011 Paper published in SoSyM 10:4 2011 Paper on test language for mobile systems (TERMOS)
  13. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper 17 Sequence

    Diagrams in the OMG specification Overview of 13 proposed semantics Semantic choices in SDs Structure of the final paper
  14. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper 18 A

    sd ex1 B Sequence Diagrams in the OMG spec. m1 C m2 Weak sequencing alt m3 m4 [A.cond ] Combined- Fragment Operand
  15. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper 19 •

    Checking traces • Refining requirements Goals • Set of traces • Petri nets, automaton… Formalisms • Denotational vs. operational • Ideas from other languages (LSC) Approaches Diversity of proposed semantics (selection)
  16. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper 20 •

    OMG: valid, invalid, inconclusive • Variations: 2 classes (e.g., safety) Semantic choices: an example Categorizing traces Simple diagram illustrating the consequences of each option Visual summary of options (~feature model)
  17. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper 22 Semantic

    choices: Combining fragments To synch, or to not Message above or below a CF might be misleading!
  18. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper 24 Choices

    and predicates Who / when / what to choose • OMG specification is permissive • Non-deterministic choices • Variations: various restrictions UML: “guard should be placed on the lifeline where the first event occurrence will occur” “First” event might be a set of events!
  19. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper 26 •

    OMG: “…represents traces that are defined to be invalid” • Vastly different options based on where SDs are used (refinement, safety properties…) Example: A more puzzling choice Negative (neg) fragment
  20. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper 28 Composition

    of conformance operators Many meanings, but assigned meaning may be surprising Suggested options: • Double neg is identity, thus trace is valid • A negative trace cannot turn into positive, thus invalid • Trace is inconclusive QUIZ: trace in double negation is valid, invalid or inconclusive?
  21. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper 30 Summary:

    using the choices framework Example: What choices did we select for our test language? Complete or partial? Allow partial traces Combining fragments Synch on entering and exiting Choices and predicates Explicit global time point for the choice Assert/negate Instead neg as operator, global false predicate at the end of diagram Conformance operators Nesting is restricted
  22. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper 32 Using

    the results of the paper DOI Design choices in a test language for mobile computing systems Test scenarios for autonomous robots and vehicles Laser Guided Vehicle DOI ADAS, automated driving DOI Education & training Industrial consultation & collaboration
  23. Micskei, Waeselynck: MODELS 2021 SoSyM 10-years impact paper 33 Citations

    in research communities (examples) Modeling languages, semantics… Requirements engineering, systems engineering Communications and networks Security and privacy Sequence diagrams are used in many domains and communities!
  24. Micskei, Waeselynck: „The many meanings of UML 2 Sequence Diagrams:

    a survey”. SoSyM 10:4 (2011). MIP talk @ MODELS 2021 34 Hélène, Zoltán sd sosym paper Genesis Content Impact are worth to take Listen to the editor and reviewers Many meanings can be useful Focus on options and not problems Be persistent Impact is a long- term matter