Thread–Based Analysis of Sequence Diagrams

Abstract

Abstract. Sequence Diagrams (SDs) offer an intuitive and visual way of describing expected behaviour of Object Oriented (OO) software. They focus on modelling the method calls among participants of a software system at runtime. This is an essential difference from its ancestor, basic Message Sequence Charts (bMSCs), which are mainly used to model the exchange of asynchronous messages. Since method calls are regarded as synchronous messages in the Unified Modelling Language (UML) Version 2.0, synchronous messages play a significantly more important role in SDs than in bMSCs. However, the effect of this difference has not been fully explored in previous work on the semantics of SDs. One important aim of this paper is to identify the differences between SDs and bMSCs. We observe that using traditional semantics to interpret SDs may not interpret SDs correct under certain circumstances. Consequently, we propose a new method to interpret SDs which uses thread tags to deal with identified problems.

Controllability Problems in MSC-Based Testing

“…It would be worth extending current work to IDs. However, it has been shown that there are significant differences between IDs and MSCs [53,54]. Finally, an MSC-based testing tool which implements algorithms and coordination services proposed in this paper would have been a nice contribution.…”

A Thread-tag Based Semantics for Sequence Diagrams

“…A Thread-tag based SD and its corresponding informal semantics are then proposed to resolve these issues. These two parts are based on results from [6]. In addition to presenting the two parts in detail, [6] also provides an analysis of the primary differences between SDs and basic MSCs and argue that meta-classes of UML 2.0 can not be used to resolve the issues within traditional semantics.…”

“…These two parts are based on results from [6]. In addition to presenting the two parts in detail, [6] also provides an analysis of the primary differences between SDs and basic MSCs and argue that meta-classes of UML 2.0 can not be used to resolve the issues within traditional semantics. Finally, in this section, informal semantics for plain SDs are extended to more complex SDs in which three new meta-classes: CombinedFragment, InteractionOperand and InteractionOperator are considered.…”

“…Since there appears to be no accurate mapping from OccurrenceSpecifications to threads with UML 2.0 metaclasses [6], we propose a new approach that extends the notation of UML 2.0. The extension should have two functions: firstly, to group all OccurrenceSpecifications in one SD to different threads; secondly, to maintain the temporal order of the OccurrenceSpecifications belonging to one thread.…”

“…The informal thread-tag based semantics for SDs was first introduced in previous work by the authors [6]. There in, only serval core meta-classes of UML 2.0 were considered.…”