A measurement based software quality framework
Quality can be assured from multiple viewpoints and various processes can be put into focus. Debugging, testing, peer reviews, process audits and assessments, measurement and risk based quality improvement are common examples of assuring software quality.
In this report we focus on measurement, which can help strengthening the quality processes or products.
One common job of quality assurance experts to assess the quality the customer releases of a software project. In many cases they have the right to decide the go/no go of a customer release. This decision is especially important in automotive, medical, aerospace and safety critical projects where errors can cause loss of human lives.
Without having a systematic approach in these scenarios the decisions made by quality assurance experts are based only on their experience and insight. Consequently, experience-based quality assurance may not provide the same results if it is performed by different personnel and/or in different projects. This is because human decisions may differ depending on the background, situation and point of view on quality.
Moreover, even if the decision is supported by measurement, different metrics, different data collection and visualisation methods can affect and make the decision of quality assurance experts and the comparison of projects difficult.
In order to eliminate the problem of experience-only based quality assurance (and to unify the way of assuring quality among quality experts) we propose an objective and repeatable approach to quality assurance which is based on measurement, proposes unified metrics and goals through projects and decreases the dependency on the experience of the quality assurance experts. We call our proposed solution the Measurement based Software Quality Assurance Framework (MSQF).
The framework helps the quality assurance experts making objective and comparable decisions in software projects by defining and assessing measurable quality goals and thresholds, directly relating these to an escalation mechanism.
We have found the application of MSQC helps achieve a transparent, organization-wide, objective measurement program based on shared and unified metrics, measurement data collection measurement and calculation.
To read this external content in full, download the complete paper from Arxiv.org.