br These component origins have been chosen as they are
These component origins have been chosen as they jw products are well-defined options related to developing with components (see for example Morisio and Torchiano (2002); Carney and Long (2000) and Gacek and Arief (2004)). The individual component origins (COTS, OSS, in-house developed components and outsourced components) have already been well explored in a number of existing literature reviews. Hauge et al. Hauge et al. (2010), for example, conducted a systematic literature review on OSS, and Maras et al. (Maras et al., 2012) conducted a review on developing software with components such as COTS. The selection of actual components may take place on multiple decision levels, namely choosing the component origin, choosing the provider, and choosing the actual component. Secondary studies on component provider selection (Khan et al., 2011) and component selection (Morandini et al., 2014) exist. No systematic literature review addressing decision-making with regard to component origins was identified, highlighting the need for such a review of the literature. Hence, the inclusion and exclusion criteria focus on the component origin, the decisions on different levels are not independent as is shown in the Fig. 1. As there is no synthesis on the component origin decision level, this study focuses on component origin selection. The decisions on how to select and integrate component (after the component origin is chosen) are not considered within the scope of this study. Systematic review studies are a means to aggregate evidence through a scientific and repeatable process. A systematic literature review was used as the research method to find the factors and solutions in the literature using the guidelines in Wohlin (2014) and Kitchenham et al. (2009). The guidelines in Ivarsson and Gorschek (2011); Cruzes and Dybå (2011); Wieringa et al. (2006) and Petersen and Gencel (2013) are used for quality assessment, data analysis, classification of research types and validation of the study respectively. Systematic review studies are a means to aggregate evidence through a scientific and repeatable process. The specific contributions achieved through the systematic review are the identification of:
The remainder of this paper is structured as follows: Section 2 presents related work on existing literature reviews closely related to our study. Section 3 presents the research method. The results are presented in Section 4 and discussed in Section 5. Section 6 concludes the paper.
Related work Three different levels of decisions are recognized as shown in Fig. 1. The light grey block represents the decision level. The white block represents the decision space where the different options (dark gray) are traded-off. The arrows represent the decision control flow. As seen in Fig. 1 the decision is not taken on the same level, the control shifts across different levels. For example, once the component origin is chosen, the provider is selected followed by the actual component selection. However, while evaluating components, the decision-maker might want to consider alternative provider or even an alternative component origin. This indicates that once a decision is taken at a decision level, thermiogenesis can be revised by the activities in the next decision level. The decision on each decision level is a research study in itself. Secondary studies on decision level for provider (Khan et al., 2011) and component selection (Morandini et al., 2014) have been conducted. No systematic literature review on the topic of component origin selection while using a systematic approach for study identification (snowball sampling (Jalali and Wohlin, 2012) and database search (Zhang et al., 2011)) was identified. Primary studies to select component origins exist, although the research problem to select component origin is not among the frequently researched topics (Vale et al., 2016). In the study (Vale et al., 2016) different research topics on component-based software engineering from its inception were reviewed. The results in Vale et al. (2016) indicate that none of the most researched topics is based on decision-making to select the component origin. However, industrial research partners have corroborated that it is an area where improvements can be made. Based on the industry interest and the limited number of studies in the area, we conclude that it is an area worthy of further research.