Search

Event-B provides a promising feature of refinement to gradually construct a comprehensive specification of a complex system including various aspects. It has unique difficulties to design complexity mitigation, while obeying Event-B consistency rules, among the potentially large possibilities of refinement plans. However, despite of the difficulties, existing studies on specific examples or high-level and intuitive guidelines are missing clear rationales, as well as principles, guidelines or methods supported by the rationales. In response to this problem, this paper presents a method for refinement planning from an informal/semi-formal specification. By defining primitive rationales, the method can eliminate undesirable plans such as the ones failing to mitigate the complexity. In a case study on a popular example from a book, we derived an enough small number of valid plans only by using the general and essential rationales while explaining the one presented in the book.

Software projects often fail because stakeholders are omitted. Existing stakeholder analysis methods rely on practitioners to manually identify and prioritise stakeholders, which is time consuming, especially in large projects with many stakeholders. This chapter investigates the use of Web 2.0 technologies, such as crowdsourcing and social networking, to identify and prioritise stakeholders. The investigation is based on the application of StakeSource in practice. StakeSource is a Web 2.0 tool that uses social networking and crowdsourcing techniques to identify and prioritise stakeholders. This chapter describes our experiences of and lessons learnt from applying StakeSource in ten real-world projects from six organisations in UK, Japan, Australia, and Canada, involving more than 600 stakeholders. We find that StakeSource can yield significant benefits, but its effectiveness depends on the stakeholders’ incentives to share information. In some projects, StakeSource elicited valuable stakeholder information; in other projects, the stakeholder responses were insufficient to add value. We conclude with a description of factors that influence stakeholder engagement via the use of Web 2.0 tools such as StakeSource. If collaborative tools such as StakeSource were to find a place in requirements engineering, we would need to understand what motivates stakeholders to contribute.

The Service-Oriented Computing (SOC) paradigm envisions the composition of loosely coupled services to build complex applications. Most current selection algorithms assume that all services assigned to a certain task provide exactly the same functionality. However, in realistic settings larger groups of services exist that share the same purpose, yet provide a slightly different interface. Incorporating these services increases the number of potential solutions, but also includes functional invalid configurations, resulting in a sparse solution space. As a consequence, applying naive heuristic algorithms leads to poor results by reason of the increased probability of local optima. For that purpose, we propose a functionality clustering in order to leverage background knowledge on the compatibility of the services. This enables heuristic algorithms to discover valid workflow configurations in shorter time. We integrate our approach into a genetic algorithm by performing repair operations on invalid genomes. In the evaluation we compare our approach with related heuristic algorithms that use the same guided target function but pick services in a random manner.

Many developers have implemented native smartphone applications (NSAs) that have the same functionalities as those of existing web applications (WAs). They need to redesign web pages as views of NSAs owing to their constraints, such as compact displays. However, it can produce a NSA with low global navigability. We propose a framework that can support developers in designing view transitions of NSAs on the basis of WAs. We focus on web templates to leverage well-designed web page transitions. Our framework 1) extracts a page transition model from a WA to create candidate view transitions of a NSA, 2) provides an interface where developers design these views to solve the constraints, and 3) suggests design modifications to increase global navigability calculated by proposed criteria of navigation costs for users. After examining case studies, we concluded that our framework could support developers to design easy-to-navigate NSAs.

Active efforts on Service-Oriented Computing have involved a variety of proposals on service selection and replacement methods to achieve quality assurance with adaptability in service composition. However, each method has specific expectations on when it is activated and how it affects service composition. Blind use of such methods can thus lead to inconsistency. This paper proposes a framework to integrate selection and replacement methods. The framework supports to clarify and analyze expectations in service composition and partner management life-cycle as well as to construct a consistent implementation accordingly. The proposed framework facilitates to test, introduce and replace service selection and replacement methods according to the environment and its change, for a variety of application domains.

Service-Oriented Architecture enables the composition of loosely coupled services provided with varying Quality of Service (QoS) levels. Given a composition, finding the set of services that optimizes some QoS attributes under given QoS constraints has been shown to be NP-hard. Until now the problem has been considered only for a single execution, choosing a single service for each workflow element. This contrasts with reality where services often are executed hundreds and thousands of times. Therefore, we modify the problem to consider repeated executions of services in the long-term. We also allow to choose multiple services for the same workflow element according to a probabilistic selection policy. We model this modified problem with Linear Programming, allowing us to solve it optimally in polynomial time. We discuss and evaluate the different applications of our approach, show in which cases it yields the biggest utility gains, and compare it to the original problem.

Quality of Services (QoS) plays an essential role in realizing user tasks by service composition. Most QoS-aware service composition approaches have ignored the fact that QoS values can depend on the time of execution. Common QoS attributes such as response time may depend for instance on daytime, due to access tendency or conditional Service of Level Agreements. Application-specific QoS attributes often have tight relationships with the current state of resources, such as availability of hotel rooms. In response to these problems, this paper proposes an integrated multi-objective approach to QoS-aware service composition and selection.

In order to make practical use of formal methods, it is not sufficient for engineers to obtain general, fundamental knowledge of the methods and tools. Actually, it is also necessary for them to carefully consider their own contexts and determine adequate approaches to their own problems. Specifically, engineers need to choose adequate methods and tools, determine their usage strategies, and even customize or extend them for their effective and efficient use. Regarding the point, this paper reports and discusses experiences on education of formal methods in the Top SE program targeting software engineers in the industry. The program involves education of a variety of scientific methods and tools with group exercises on practical problems, allowing students to compare different approaches while understanding common principles. In addition, the program involves graduation studies where each student identifies and tackles their own problems. Statistics on problem settings in the graduation studies provide interesting insights into what top-level engineers tackles after learning formal methods.

Service selection is a central challenge in the context of a Service Oriented Architecture. Once functionally sufficient services have been selected, a further selection based on non-functional properties (NFPs) becomes essential in meeting the user's requirements and preferences. However, current descriptions of NFPs and approaches to NFP-aware selection lack the ability to handle the variability of NFPs, that stems from the complex nature of real-world business scenarios. Therefore, we propose a probabilistic approach to service selection as follows: First, to address the inherent variability in the actual values of NFPs at runtime, we treat them as probability distributions. Then, on top of that, we tackle the variability needed in describing NFPs, by providing conditional contracts. Finally, from usage patterns, we compute user-specific expectations for such NFPs. Further, we depict a typical scenario, which serves both as a motivation for our approach, and as a basis for its evaluation.

Organization of mobile agents into a group has appeared as a new paradigm for dynamic deployment of composite services. However, it has not been discussed how multiple mobile agents cooperate with each other, handling conflicts in their requirements. In response to this problem, this study proposes a model for cooperative mobility based on the notion of agreements. Agent behavior defined in the proposed model involves agreement establishment and enforcement for cooperative mobility. Such behavior can be customized only by specifying requirements/constraints of each agent, eliminating the necessity to write down the whole behavior to handle agreements. The model is described in a formal way, using Event Calculus, and it is proved the model leads to no occurrence of defined inconsistency. The model has been implemented on an existing agent framework, Freedia, combined with its dynamic partner management mechanism.