Development of an expansible taxonomy of software requirements nomenclature: Viewing the elicitation process as communicative discourse

This research was undertaken to achieve an understanding of the mental conceptualizations of systems analysts for requirements nomenclature (expressions) used in research and practice. New approaches for requirements elicitation are necessary due to ambiguous requirements cited as the cause for software failures. The primary objective of this research is to develop an expansible taxonomy of requirements nomenclature using personal conceptualizations of the meaning of requirements nomenclature as a guide in the development of a requirements taxonomy, which can be expanded as nomenclature and nature of a particular domain changes. The techniques used to gather data included a combination of questionnaire to obtain background information from participants, discourse analysis of an email list used by practitioners and literature to obtain requirements expressions, repertory grid to obtain mental conceptualizations of the requirements expressions and term sorting which revealed an overall conceptual understanding of the relational information about language in use during requirements elicitation.; It was postulated for this research that common nomenclature is used during requirements elicitation and specification discourse, but that the mental conceptualizations of nomenclature meaning, although similar among the participants, will differ in usage and reflect variations affected by education and work experience. These results, including noted differences, provided the guidelines for the development of an unnumbered, three-level taxonomy of common nomenclature.; The taxonomy developed into a hierarchical-faceted conceptual structure representing emergent themes discovered from the repertory grid content analysis. The emergent categorical themes (Level A labeled "Category") were developed during content analysis of each participant's constructs. The constructs represent the way in which the individual construes the expressions through the repertory grid process. Level B, which is labeled "Facets", represents the actual constructs given by the participants. The constructs, in turn, were elicited during the repertory grid interviews using the common requirements expressions used in discourse. These expressions represent Level C and are labeled "Basic-level expressions". Although relatively few Level-A categories emerged, it was found that the same basic-level expressions that comprise the every-day discourse were used in different categorical areas by the same individual as well as by different individuals. That is, for some individuals, an expression, such as "quality requirement" would be used in the context of testing and measurement, point-of-view or when defining user needs. This shows a subtlety of understanding by an individual, which could hinder the requirements development process.; With this knowledge and the use of a structured interviewing technique like repertory grid to develop a taxonomic conceptual aid that can be changed and expanded for the contextual needs of the work domain, the true conceptualizations of the stakeholders and systems analysts can be revealed to afford a deeper understanding of the domain language of the domain in which the systems development is taking place.