An Approach for Change Impact Analysis in Web systems - A Industrial Case study
International Journal of Software Engineering and Its Applications
There is a growing body of research on change impact analysis (CIA) approaches that specifically addresses changes and their impacts on architecture design. However, there is little research focus on approaches that particularly support the identification of impacts on architecture design resulting from business process changes i.e. early identification of change impacts in Web systems. To address this problem, we have proposed a systematic, structured and rigorous approach called as process
... el of CIA (PMCIA). PMCIA consists of a set of defined steps/activities, inputs, outputs, and employs architecture design information. In this paper we have presented the results of PMCIA validation in industrial setting through a detailed case study. The case study was carried out across two releases of a selected Web system project in an organization. The case study results indicate that the proposed approach, indeed, supports for early identification of change impacts in Web systems and provides Web developers the necessary components for systematically performing CIA. architecture design knowledge during CIA. To illustrate that PMCIA practically supports for early identification of Web systems in industrial setting, we poses two research questions as: RQ1: Does the proposed process model of CIA support Web developers/architects in early identification of change impacts in Web systems? RQ2: Does the proposed process model for CIA address the specific characteristics of Web systems? To answer these research questions we choose to execute of PMCIA in an industrial case study. Case study represents an appropriate choice for our research goal as they enable us to validate our proposed solution in a real project and within a complex organisational environment. Furthermore, a number of similar examples are reported in research literature where case study as a research methodology has been successfully used for validation of Web engineering approaches , for Web application design process,  and for model-driven process in Web engineering  . In this paper, we provide the execution results of PMCIA performed across two releases of a selected Web system project in a software development organisation. The case study results provide the necessary evidences both toward the instantiation and preliminary validation of PMCIA. Essentially, the results indicates that PMCIA leads to correct and adequate identification of impacts, and support for early identification of change impacts in Web systems as compared to the situation when PMCIA is not being used. This paper begins by describing the objectives and settings of case study in section 2. In section 3, we briefly describe selected Web system, present the case study execution results and analysis results. Additionally in section 4, we present discussion detailing the important findings whilst taking into consideration the research questions RQ1 and RQ2. We finish this paper by providing conclusions and future work in section 5. Case Study Objectives and Settings The objective of the case study is to address the research questions RQ1 and RQ2 with regard to the validation of PMCIA. Specifically, we intended to validate whether PMCIA supports Web developers/architects in early identification of change impacts in their Web systems project? Further, our intention was to look at whether process model of CIA addresses the characteristics of Web systems? Keeping in view the above research questions, we have defined three criteria for case study site selection as (i) Case study site should be organization that engaged in Web systems development, (ii) the organization should committed to adopt new process for the purpose of improving their Web development practices, and (iii) the size of the company should be medium to large, and should employ current web systems development technologies. For case study investigation we have selected two different releases of same Web system project and compare the findings of exiting CIA approach (in first release) and proposed PMCIA as a new approach (in second release) respectively. Data Collection and Analysis We have used both qualitative and quantitative data analysis approaches. For quantitative data such as responses from closed-ended questions, we tabulated in a spreadsheet and analysed with basic statistical method as frequency analysis. In order to analyse responses from open-ended questions and project documents, content analysis was used as a data analysis method. Content analysis was applied to examine written or recorded communication and helped us to determine the presence of certain words or concepts within texts  . Additionally, interpretive analysis was used as a data analysis method. During interpretive analysis, researcher synthesized the data by involving both induction and interpretation and therefore avoided specifying concepts in advance of the synthesis. For a single case study, like ours, triangulation is also regarded as a means of looking at a single complex phenomenon under investigation in order to better understand the context in its multifaceted form. We used both triangulation of data and methods (see Section 6.3). We made linkage between qualitative and quantitative methods  with the emphasis on qualitative methods. At few occasion, in our case study qualitative data helped us to interpret, clarify, and illustrate quantitative findings.