The requirements engineering activities within a software project are known to be critical to the successful production of a correctly functioning system. This is particularly so when considering the varying views of multiple stakeholders. One promising approach for improving the outcome is to introduce formal negotiation. Clearly, making the set of requirements more closely represent all the stakeholders' perspectives and perceptions, underpins a sound basis for project estimation, improved system quality and a reduction of the resources necessary. However, such benefits are mere speculation if it is not possible to provide empirical evidence. The purpose of this research was to set up a metric framework to measure a number of key aspects of negotiation within requirements elicitation and validation. A number of experiments and trials were designed and implemented to quantify the expected outcomes of requirements negotiation. First, negotiation is shown to be beneficial to identify and to resolve conflicts between stakeholders. In particular, it is valuable in the allocation of priorities for a particular requirement and in the decision on its inclusion or exclusion. This is particularly useful in evolutionary and incremental paradigms such as the Agile methods. Also, it is necessary when time and resources are constrained - a condition in most software projects. Whereas this first outcome is hardly surprising, it does validate the fundamental purpose of negotiation. Second, having demonstrated that negotiation leads to an agreement, there is a need to check whether that consensus is heading in the right direction. The concept of a perfect iii requirements specification (here termed the "gold standard") is introduced and measurements of progress towards such an ideal are made.
|Qualification||Doctor of Philosophy|
|Publication status||Unpublished - 2012|