Context: Software projects are common inputs in Empirical Software Engineering (ESE) studies, although they are often selected with ad-hoc strategies that reduce the generalizability of the results. An alternative is the usage of available datasets of software projects, which should be current and follow explicit rules for ensuring their validity over time. Goal: In this context, it is important to assess the general state of software datasets in terms of purpose, last update, project characterization, source code metrics, and tools to extract source-code-related artifacts. Method: We conducted a systematic mapping study retrieving software datasets used in ESE studies published from January 2013 to December 2021. Results: We selected 74 datasets created mainly for software defects, software estimation, and software maintainability studies. The majority of these datasets (64%) explicitly stated the characteristics to select the projects, and the most common programming languages were Java and C. Conclusions: Our study identified scarce efforts to keep datasets updated over time and also provides recommendations to support their construction and consumption for ESE studies.

Continuous Delivery is a practice where high-quality software is built in a way that it can be released into production at any time. However, systematic literature reviews and surveys performed as part of this Doctoral Research report that both the literature and the industry are still facing problems related to testing using practices like Continuous Delivery or Continuous Deployment. Thus, we propose Continuous Testing Improvement Model (CTIM) as a solution to the testing problems in continuous software development environments. It brings together proposals and approaches from different authors which are presented as good practices grouped by type of tests and divided into four levels. These levels indicate an improvement hierarchy and an evolutionary path in the implementation of Continuous Testing. Also, an application called EvalCTIM was developed to support the appraisal of a testing process using the proposed model. Finally, to validate the model, an action-research methodology was employed through an interpretive theoretical evaluation followed by case studies conducted in real software development projects. After several improvements made as part of the validation outcomes, the results demonstrate that the model can be used as a solution for implementing Continuous Testing gradually at companies using Continuous Deployment or Continuous Delivery and measuring its progress.

It is straightforward to find ourselves avoiding tasks or situations every day, instead of making an effort to deal with them. This behavior is known as “procrastination”, and it is very common in industrial societies. In the educational context, one of the most likely procrastination situations is the development of a bachelor degree thesis or Master's thesis where the student is supposed to work on an original written work. This paper presents the findings of a survey on typical barriers of the thesis's development, carried out with 51 Computer Science Master's students. The survey measured participants' procrastination using the General Procrastination Scale proposed by Lay and their current situation in the thesis's development. The findings of the survey show that the main barriers to begin the final thesis work are lack of time and motivation. Furthermore, general and descriptive information about the students' current situation was obtained.