Information-Centric Assessment of Software Metrics Practices

Combining Perceptions and Prescriptions in Requirements Engineering Process Assessment: An Industrial Case Study

“…Each inquiry cycle, whether perceptiondriven or prescription-driven, involves engaging stakeholders, collecting data, analyzing data, and debating findings, see Table 1. Planning each cycle in detail involves selection of stakeholders to engage, selection of projects to analyze, and selection of sources of data (documents, observation, interviews, and workshops) [7], [10].…”

“…Perception-driven inquiry captures data about individual beliefs and experiences in the specific company context; the detailed planning of these cycles can find inspiration in a number of inductive assessment methods: systematically selecting projects and people and triangulating between interview data and project documentation [7], engaging stakeholders in structured debates of different system models of current practices [10], systematically prioritizing perceptions across stakeholder groups [11], using survey data to rank projects according to specific success variables [12], and identifying and prioritizing perceived problems based on interviews with key project managers [17], [7], [10], [11], [12]. Prescription-driven inquiry captures data on how current practices benchmark against predefined processes, best practice, and predefined questions; there are a number of RE-focused prescriptive methods available for this purpose: analyzing RE related data based on SW-CMM mini assessments [24], using the RE version of SW-CMM, R-CMM [23], [24], [25], [26].…”

“…Nielsen and Nørbjerg [17] report research in which stakeholders in low-maturity companies were actively engaged in identifying key problems and implementing what was perceived to be sensible improvements; Wohlin and Andrews [12] present a method for analyzing project success based on subjective data; Liu et al [11] provide a method to prioritize process assessments from multiple stakeholders and aggregate them into requirements for improvement; Frederiksen and Mathiassen [10] present an informationcentric method based on stakeholder participation for assessing software metrics programs; and Pettersson et al [7] have developed a lightweight method for software process assessment and improvement planning based on involvement of key stakeholders. While these methods document the feasibility of basing assessments on stakeholder perceptions, none of them focuses specifically on RE, and none of them provides guidance on how to effectively combine perceptions with prescriptions.…”

“…In contrast, prescriptive methods center improvement on the best practices within industry. While the dominant approaches to software process assessment and improvement are prescriptive (e.g., [8], [9]), there are a number of inductive approaches available (e.g., [5], [7], [10], [11], [12]) and stakeholder perceptions and subjective judgments have more generally been adopted as an important approach to inform other software disciplines (e.g., [13], [14], [15], [16]). …”

“…Assessment specificity: How well does the assessment uncover improvement problems and opportunities within the company? This challenge is typically addressed through inductive methods (e.g., [7], [10], [11], [12], [17]). Besides providing insight into the specific improvement context, these methods help achieve buy-in from stakeholders by involving them in the process.…”