Evaluation of Novel Approaches to Software Engineering

Valuable information can be obtained from the relationships that hold between the elements involved in any Model-Driven Engineering (MDE) process. This information can be then used to support impact change analysis, validation of requirements, etc. However, dealing with traceability is a complex and error-prone task if no tool support is provided to that end. The adoption of MDE can definitely alleviate such complexity. For instance, MDE techniques such as models transformations, matching or weaving, can be used to automate the production and management of traceability, without requiring an extra effort from any of the stakeholders involved in the project. In this line, this work presents the different visualization mechanisms for traceability information supported by iTrace, a framework for the management of traceability in the context of MDE. They provide insights into how the elements of a given project relate to each other, offering simple and intuitive representations of such relationships with different granularity levels. These visualizations help to reduce the inherent complexity of dealing working with model transformations, making it possible for instance to understand the typology of the elements processed by a particular mapping rule without mastering the entire transformation language or even the transformation under study.

Software processes have evolved significantly since the first formal appearance of software engineering. The academia and the industry have introduced, embraced or rejected various methodologies that are more or less efficient in theory and in practice. A current popular trend can be found in Agile methodologies widely adopted in the last decade. Since software processes are constantly evolving, it is vital to see how they evolve over time. This work presents the current state of the adoption of Agile methodologies with an emphasis on Scrum development method. Study results from 44 different countries were collected during the months of March and April 2012. The results are enlightening in order to understand how Agile development and Scrum are viewed today, to see where their success factors lie, discover if they offer benefits in comparison to heavyweight approaches and discuss their future evolution.

Enterprise architectures (EA) try to develop an alignment between an enterprise’s technology infrastructures with its business objectives and are often facilitated by an EA framework (EAF). EAFs provide the processes to create and govern an EA and have been used to understand both strategy and business architecture to synthesize a supporting information system infrastructure. However, existing EAFs do not provide lightweight, systematic process for eliciting the needed inputs to develop an EA. The contribution of this work is a lightweight, systematic approach for eliciting the enterprise vision, mission and objective requirements necessary as input to an EAF. We make two basic claims for this idea. First, the utilization of the Vision-Mission-Objectives-Strategy-Tactics (VMOST) queries provides a lightweight approach for eliciting required EA knowledge from stakeholders. Second, the use of the Grounded Theory Method, a qualitative analysis technique, provides a structured, systematic approach for analyzing and documenting elicited EA requirements. To illustrate these claims, we apply our lightweight, EA elicitation approach to a real world enterprise using the case study approach as a research methodology.

Model-driven development (MDD) emphasizes platform-independent models. Approaches such as the Object Management Group’s Model Driven Architecture (MDA) are built on a foundation of standards and specifications, but require significant effort to encode and interpret the models during the transformation to the final application. A second approach to MDD uses domain-specific languages (DSLs) as a means of simplifying the models and transformations for applications within that domain. In this paper we look at AXIOM, a DSL for mobile application development, and how it allows for platform-independent models to be used to generate native code in a lightweight manner.

Software testing is known to be expensive, time consuming and challenging. Although previous research has investigated relationships between several software properties and software testability the focus has been on static software properties. In this work we present the results of an empirical investigation into the possible relationship between runtime properties (dynamic coupling and key classes) and class testability. We measure both properties using dynamic metrics and argue that data gathered using dynamic metrics are both broader and more precise than data gathered using static metrics. Based on statistical analysis, we find that dynamic coupling and key classes are significantly correlated with class testability. We therefore suggest that these properties could be used as useful indicators of class testability.

Online testing approaches are becoming crucial in today’s complex systems. By that way, testing a protocol at run-time has to be performed during a normal use of the system without disturbing the process. The traces are observed and analyzed on-the-fly to provide test verdicts and no trace sets should be studied as a posteriori to the testing process. In this process, it is a challenging work to keep the same preciseness in conformance testing and the same efficiency in performance testing. In this paper, aiming to find a solution, we present a novel online passive testing approach based on Horn-Logic. In order to evaluate and assess our approach, we also developed a prototype and experimented it with a set of Session Initiation Protocol properties in a real IP Multimedia Subsystem environment. Finally, the preliminary results and discussions are provided.

Many processes in Software Engineering, and specifically in the Unified Software Development Process, require the participation of several actors who may play different roles. Collaborative software (groupware) can solve the problems that arise when trying to deal with such processes. Within this scope, we have developed a domain independent synchronous collaborative tool that can be specialized to work with several types of diagrammatical domains. Among those domains, the diagrams used in the Unified Process can be found. In this paper we describe how we have instantiated this model-based tool to work with some diagrams in the Unified Process. Also, in the paper we explain how we have carried out some studies with this tool to obtain conclusions regarding several issues, including the analysis of the communication and coordination among users, and the relationship between them and the quality of the work.

Inherent properties of games, such as rules, goals and interaction, have made them popular to address challenges and sort obstacles in a wide variety of contexts. Within Software Engineering, a challenging activity of the software development process is the Inception phase, in which stakeholders’ needs, required functionalities, objectives, risks and constraints of a software product are established. An alternative to optimize stakeholders’ participation during the Inception phase and make more efficient its outcomes is to do it through games.

Taking into account the uncommonness of games in Software Engineering development process and the lack of complex methods that include games, this paper presents ActiveAction, a method that combines classical requirements specification techniques with games. ActiveAction resulted in a successful game-based strategy that has improved the Inception phase of the software development projects based on the fact that stakeholders express their ideas freely in an unstressed environment; real-life scenarios are simulated to identify exceptions; requirements and risks are determined in a collaborative manner. It is concluded that the inclusion of games in such a challenging activity as software projects inception, is feasible and reported promising results that benefit both stakeholders and software developer organizations.

The quality of the delivered software relies on rigorous testing performed and testing is as good as the test-cases. However, designing good test-cases is a challenging task. The challenges are multi-fold and designing test-cases is often delayed towards the end of implementation phase during software development. In this paper, we propose an approach to automatically generate the test-cases from logical form of the requirements specifications during early phases of software development. Our approach is based on courteous logic representation of requirements. The knowledge stored in the courteous logic representation of requirements is used to automatically generate trivial as well as functional test-cases. We present an evaluation of the effectiveness of our generated test-cases through the case-studies conducted. We further report our observations for the empirical study conducted with subjects from different backgrounds.

We present our framework for visualization, simulation and validation of cyber-physical systems in industrial automation during development, operation and maintenance. System models may represent an existing physical part – for example an existing robot installation – and a software simulated part – for example a possible future extension of the physical industrial automation setup. We call such systems cyber-virtual systems. Here, we present our VxLab infrastructure for visualization using combined large screens and its applications in industrial automation. The methodology for simulation and validation motivated in this paper is based on this infrastructure. We are targeting scenarios, where industrial sites which may be in remote locations are modeled, simulated and visualized. Modeling, simulation and the visualization can be done from different locations anywhere in the world. Here, we are also concentrating on software modeling challenges related to cyber-virtual systems and simulation, testing, validation and verification techniques applied to them. Software models of industrial sites require behavioral models of both human and machine oriented aspects such as workflows and the components of the industrial sites such as models for tools, robots, workpieces and other machinery as well as communication and sensor facilities. Furthermore, facilitating collaboration between sites and stakeholders, experts and operators is an important application of our work. This paper is an extension of our previously published work [1].

When addressing mobile applications, it is a technological landscape that is emerging with new requirements and restrictions requires a reassessment of current knowledge about the processes of development of these types of systems. These new systems have different features, ranging from planning to completion of the design, and therefore a particular area that is being addressed differently when it comes to estimating software. The estimation processes in general are based on characteristics of the systems to attempt to quantify the complexity of the implementation. For this reason, it is important to analyze the main models currently proposed for estimating software projects and consider whether it is suitable for mobile computing. Thus, the main objective of this paper is to present an estimation method for mobile applications still in the design phase, giving basis for all the features addressed in this scenario.