Abstract
Live Sequence Charts (LSCs) extend standard Message Sequence Charts (MSCs) by introducing a formal distinction between mandatory and provisional behaviors. By incorporating "temperature" (hot vs. cold) and "modes" (universal vs. existential), LSCs enable the specification of liveness and safety properties, bridging the gap between informal scenario-based modeling and rigorous formal verification.
Introduction
In the late 1990s, Message Sequence Charts (MSCs) were widely used to document system use-cases. However, they suffered from semantic weakness. A standard MSC describes a partial ordering of events but fails to clarify the intent: does the chart represent a possible execution or a required one? This "grand dichotomy" between inter-object scenario descriptions and intra-object state machine logic left a gap in expressing liveness—the guarantee that an event will eventually occur. Standard MSCs could not distinguish between a sample execution and a universal requirement, nor could they specify "anti-scenarios" (forbidden behaviors).
Methods
The authors propose LSCs as a "smooth extension" of the ITU standard for MSCs. The fundamental innovation is the introduction of temperature and modes to differentiate between what might happen and what must happen.
As shown in the table above, LSCs categorize elements by their "temperature":
- Hot elements (mandatory): Represented by solid lines, these must be satisfied. If a system reaches a hot location, it must progress. A false hot condition results in a system abort.
- Cold elements (provisional): Represented by dashed lines, these are optional. If a cold condition is not met, the system gracefully exits the current chart.
Furthermore, LSCs operate in two primary modes:
- Universal Charts: These must be satisfied by all system runs. They often use "activation conditions" to trigger mandatory behavior.
- Existential Charts: These must be satisfied by at least one run, serving as sample scenarios.
The formal notation for message exchange is also refined to support both synchronous and asynchronous communication between instances and the environment.
The formal semantics are grounded in a "skeleton automaton" that tracks the "cut" (progress) of the chart, ensuring that "promises" (hot events) are eventually fulfilled.
Results and Key Insights
LSCs significantly increase the expressiveness of scenario-based modeling. They can explicitly model safety properties (ensuring "bad" things never happen) and liveness properties (ensuring "good" things eventually happen). By making conditions "first-class citizens," LSCs allow for the specification of forbidden scenarios and complex control flow (if-then-else, loops) within a visual framework. This provides a formal path to relate inter-object scenarios to intra-object statecharts.
Discussion and Future Work
While powerful, LSCs introduce increased semantic complexity. Early versions relied on variables rather than explicit states and omitted detailed real-time constraints. A major ongoing research direction is the automatic synthesis of statecharts from LSC specifications, a challenging problem in formal methods. Despite these hurdles, LSCs remain a cornerstone for bridging the gap between requirements engineering and formal system design.
Summary updated