Secondary and primary studies have considered
Secondary and primary studies have considered the decision to select the provider and to select components after the component origin is chosen. Examples of secondary and primary studies outside our scope, but still related to decision-making are listed in Tables 1 and 2. The secondary and primary studies related to the provider and component selection decision levels are presented in Table 1. Primary studies in Table 1 are a subset of existing studies related to the provider and component selection. While, Table 2 provides a list of secondary studies related to the different topics with respect to the individual component origins. The primary studies that compare the component origins are included in the result of the review study and are mentioned in the result section.
As seen from Table 2 no secondary studies on COTS and outsourcing were conducted. Using the selection of relevant keywords of each comparison origin in Scopus, in combination with the research method of systematic literature review and systematic mapping studies, COTS and outsourcing of components did not yield any results. As the focus of this SAR302503 study is on decision-making in choosing the component origin, no deeper analysis of the studies is provided.
Results The answers to the research questions are provided in this section. The contribution of the primary studies with respect to the comparison categories is shown in Table 8.
Discussion From the high-level themes discussed in 4.2.1 Project metrics factors, 4.2.2 External factors and 4.2.3, we can see that OSS has more positive influence on the decision when project metrics and external factors are considered. In-house development and OSS have more positive influence when software development process factors are considered. However, when the individual factors/themes are considered each component origin has some advantage over some other component origins. The results indicate that some trade-offs between factors and dependencies on the context need to be considered which is discussed in this section along with the advantage and disadvantages of the component origins. The advantages are discussed in Table 24. The advantages of outsourcing are not discussed as none of the primary studies have mentioned them. The information presented in Table 24 can be used as an initial set of factors to be considered by practitioners in the decision-making. In addition, there are some trade-offs Replication fork need to be considered along with the advantages. Figs. 4–6 illustrate the trade-offs between the different factors. The trafe-offs that should be considered along with advantages of COTS over OSS are depicted in Fig. 4 and discussed in Section 5.1. The trafe-offs that should be considered along with advantages of OSS over COTS are depicted in Fig. 5 and discussed in Section 5.2. Whereas, Fig. 6 discusses trade-offs that are related to the advantages of in-house development over COTS and OSS, which is discussed in Section 5.3. Section 5.4 discusses the advantages of COTS and OSS over in-house development, no trade-offs were mentioned for the advantages discussed in Section 5.4 in the primary studies. The trade-offs and dependencies depicted in Figs. 4–6 are aggregated from different primary studies. The “↔” arrow represents the trade-off between different factors and “ → ” arrow represents the dependencies on the context that affects the factors. Depending on the context, the factors have positive or negative impact.
Conclusion RQ1: What are the research types, methods and quality of the contributions? All the research types classified by Wieringa Wieringa et al. (2006) were identified. Hence, the primary studies consisted of a good mix of empirical and non-empirical results. This distribution of the research types allowed to identify conflicting results based on context dependencies. The most common research type is evaluation research indicating that the research is empirical and rooted in industrial practice, which is a positive result. The most common research method is survey indicating that the results are not specific to a particular context. However the context is not described sufficiently in the primary studies. The results of quality assessment indicate the lack of rigor in the primary studies in terms of reporting the context, validity threats and research design.