| Requirements envisaging is the process of transforming vague and informal requirements into precise descriptions. Envisaging evokes ideas, project criteria, and the explanation of alternative solutions which are refined or discarded. At the envisioning stage of system development, complex systems are typically described in a fragmentary and highly contextual manner. This conflicts with the abstract and decontextualized formal language used by software experts. As a consequence, requirements envisaging, which seeks to bridge this gap, is a challenging phase of system development in which to provide automated support.; In requirements envisaging, domain experts will often convey partial descriptions of system and environment behavior arising in restricted situations, namely, scenarios. Scenarios play an important role in envisaging by mediating communication and by describing alternative situations and rationale explored during design. Despite this importance, scenarios are not, in general, formally captured as part of requirements documentation.; This dissertation is a step toward automated support for envisaging with scenarios. For this task, the representation used to capture scenarios must support human-tool collaboration. The tool described herein supports capturing scenarios in a formal manner despite their fragmentary and contextual nature. The goal is to let people who are not necessarily computer experts create scenarios easily and allow other people to readily understand the concepts conveyed in these scenarios.; The main accomplishments reported in this thesis are: an observational study of domain and software experts utilizing scenarios, the development of a formal representation for scenarios, an automated tool that allows people to create scenarios in that representation, and evaluation of the representation and tool in a real world domain outside those studied during development.; These accomplishments are a step toward: bridging the communication gap between domain experts and system design experts, moving some of the burden of work from people to machines, the documentation of domain knowledge and rationale, and the traceability between requirements and implementation by providing a formal means to capture scenarios. |
