Engineering interaction requirements of mobile applications that support people-driven collaborative processes

Abstract

Participants involved in a people-driven collaborative process (PDCP) usually require a mobile collaborative application to support their activities. Given that the workflow of a PDCP cannot be established at design time, conceiving these mobile systems is far from simple, in part due to the diverse interaction needs of the participants and the scarce knowledge to help designers address the coordination and collaboration services at early stages of the system development. It means that developers and stakeholders have to deal with ad hoc products in bespoke projects. In the design of such products, the developers mainly specify both the interaction scenarios to be supported with the mobile application and the interaction services to be used by the participants to carry out the process. Such specifications (typically visual models) should be analyzed, validated, and refined by developers and stakeholders in order to reach an early agreement about the scope of the mobile application being developed. A study conducted in this thesis work shows that existing languages and notations, used to create visual models of the interaction scenarios, are limited to address the challenges involved in this modeling activity. Therefore, this dissertation proposes a language named CIMoL (Computer-Mediated Interactions Modeling Language) for modeling computer-mediated interaction scenarios involved in PDCPs. This language, together with an automatic prototyping tool (named CIMoL Modeler), allows software engineers and stakeholders to identify and validate the interaction and awareness services required by the participants in a PDCP. Two instruments were used to evaluate CIMoL: an expert study considering a set of design aspects reported in the literature and a case study where the participants conceived mobile collaborative systems to support a particular PDCP. The results were positive, indicating that CIMoL effectively can help developers and stakeholders to co-conceive, validate and refine the interaction scenarios, and then reach a shared understanding about the scope of the system being developed. As future lines of research, we propose a further evaluation of CIMoL involving a larger group of participants and PDCPs in additional business domains, to identify both opportunities for improvement and the level of maturity for adoption in the industry.