Complexity cohesion and coupling have been recognized as prominent indicators for software quality. One characterization of software complexity is the existence of dependency relationship. Moreover, degree of dependency reflects the cohesion and coupling between software elements. Dependencies on design and implementation phase have been proven as important predictors for software bugs. We empirically investigated how requirements dependencies correlate with and predict software integration bugs, which can provide early estimate regarding software quality, therefore facilitate decision making early in the software lifecycle. We conducted network analysis on requirements dependency networks of two commercial software projects. We then performed correlation analysis between network measures (e.g., degree, closeness) and number of bugs. Afterwards, bug prediction models were built using these network measures. Significant correlation is observed between most of our network measures and number of bugs. These network measures can predict the number of bugs with high accuracy and sensitivity. We further identified the significant predictors for bug prediction. Besides, the indication effect of network measures on bug number varies among different types of requirements dependency. These observations show that requirements dependency network can be used as an early indicator of software Integration bugs.

Background: Change propagation analysis helps predict the parts of the software that may be affected if a change is made. Existing research on change propagation focuses on design and code level changes. However, as a software evolves, the requirements that drive these changes also have intricate dependencies. Understanding the effect of these requirement dependencies on change prorogation is useful but not trivial. More than twenty requirements dependency types have been identified in the literature, however there still lacks an evaluation of the applicability of these dependency types in requirements and change propagation analysis. Aim: We aim to investigate whether these dependency types are useful for change propagation analysis. Method: We conducted a case study in a real-world industry project. This case study evaluates two representative dependency models covering twenty five types of dependencies. Results: Our initial evaluation has found that five dependency types are particularly useful in change propagation analysis and practitioners with different backgrounds have various viewpoints on change propagation. Thus change impact analysis should involve a wide range of stakeholders including project managers, requirements engineers, designers and developers. Conclusions: Our case study provides insights into requirements dependencies and their effects on change propagation analysis for both research and practice.

Background: Software systems must evolve in order to adapt in a timely fashion to the rapid changes of stakeholder needs, technologies, business environment and society regulations. Numerous studies have shown that cost, schedule or defect density of a software project may escalate as the requirements evolve. Requirements evolution management has become one important topic in requirements engineering research. Aim: To depict a holistic state-of-the-art of requirement evolution management. Method: We undertook a systematic review on requirements evolution management. Results: 125 relevant studies were identified and reviewed. This paper reports the preliminary results from this review: (1) the terminology and definition of requirements evolution; (2) fourteen key activities in requirements evolution management; (3) twenty-eight metrics of requirements evolution for three measurement goals. Conclusions: Requirements evolution is a process of continuous change of requirements in a certain direction. Most existing studies focus on how to deal with evolution after it happens. In the future, more research attention on exploring the evolution laws and predicting evolution is encouraged.

Background: Systematic Literature Reviews (SLRs) have been gaining significant attention from Software En-gineering (SE) researchers since 2004. Several researches have also working on improving the scientific and techno-logical infrastructure available to support SLRs in SE. Objective: The study reported in this paper aims to vali-date the QGS-based search process for SLR, i.e. whether a more effective and/or productive search can be achieved by following such a systematic process. Method: We used a dual-case study, in which each case includes two observations of SE literature search for the same SLR but using and not using the QGS-based approach. Results: The overall sensitivity and precision of each observation were calculated for the search cases that im-plemented different search design. Conclusions: A systematic search process (QGS-based search in this paper) appears to gain higher sensitivity and precision in the two cases of SLR. Such a search process may help capture more relevant studies as well as save re-viewers ’ time spent in literature search activities. Our ob-servations also proof that an integrated search strategy is recommended for SLRs in SE to avoid the possible limita-tions of applying single manual or automated search. 1

More and more software applications are developed within a software ecosystem (SECO), such as the Face book ecosystem and the iPhone AppStore. A core asset of a software ecosystem is its users, and the behavior of the users strongly affects the decisions of software vendors. The number of active users reflects user satisfaction and quality of the applications in a SECO. However, we can hardly find any literature about the number of active software users. Because software users are one of the most important assets of a software business, this information is very sensitive. In this paper, we analyzed the traces of software application users within a large scale software ecosystem with millions of active users. We identified useful patterns of user behavior, and proposed models that help to understand the number of active application users. The model we proposed better predicts the number of active users than just looking at the traditional retention rate. It also provides a fast way to monitor user satisfaction of online software applications. We have therefore provided an alternative way for SECO platform vendors to identify rising or falling applications, and for third party application vendors to identify risks and opportunity of their products.

Requirements evolution is one of critical problems influencing software engineering activities. Despite there is much research on requirements evolution, there still lacks quantitative understanding of requirements evolution. In this paper, we quantitatively analyze requirements evolution across multiple versions of an industrial software product. Based on data of requirements evolution and defects, we analyze the relationship between requirements evolution and requirements as well as between defects and requirements evolution. We also analyze the evolution characteristics about requirements modification. Our findings include estimation of the number of defects using evolved requirements may increase accuracy of defect estimation and business rule is the most volatile part in requirements. These findings deepen our understanding of requirements evolution and can help software organizations manage requirements evolution.

Resource allocation in software development is important and many methods have been proposed. Related empirical research is yet scarce and evidence is required to validate the theoretical methods. This paper introduces the staffing pattern as a metric of resource distribution among project phases, and verifies its effect on software quality and productivity using real project data. The main findings are: (1) there exist different staffing patterns in reality; (2) the staffing pattern has significant effect on software quality (post-release defect density); (3) the staffing pattern has no significant effect on productivity; (4) the effort invested on test, document or code inspection possibly explains the effect of staffing pattern on software quality; (5) the effort consumed by rework perhaps counteracts the effect of other potential factors on productivity. Preliminary heuristics are suggested to resource allocation practices.

In software projects it is crucial to control the actual project schedule, cost and product quality, against the project plan. Variations often occur during the running project, and big ones might have significant impacts on the whole project. Thus, it is necessary to identify causes quickly, and take the right corrective actions in time. This paper provides a threedimensional software development model, called AHA (Activity-Human-Artifact). It combines three key factors, activity, human and artifact, corresponding to project schedule, cost, and product quality, respectively. We also define some relevant elements, attributes and relationships among different dimensions in the AHA model, and their impacts with respect to schedule, cost and quality. On the basis of our model, a project tracking process is developed to guide analysts to locate potential causes for different kinds of variations, and effectively correct them in real projects. We also take a software project as the application scenario to illustrate our model and the tracking process identified.

XP (eXtreme Programming) has become popular for IID (Iteration and Increment Development). It is suitable for small teams, lightweight projects and vague/volatile requirements. However, some challenges are left to developers when they desire to practise XP. A critical one of them is constructing the release plan and negotiating it with customers. In this paper, we propose a risk-driven method for XP release planning. It has been applied in a case study and the results show the method is feasible and effective. XP practicers can follow it to decide a suitable release plan and control the development process.