Using the crowds to satisfy unbounded requirements

The Internet is a social space that is shaped by humans through the development of websites, the release of web services, the collaborative creation of encyclopedias and forums, the exchange of information through social networks, the provision of work through crowdsourcing platforms, etc. This landscape offers novel possibilities for software systems to satisfy their requirements, e.g., by retrieving and aggregating the information from Internet websites as well as by crowdsourcing the execution of certain functions. In this paper, we present a special type of functional requirements (called unbounded) that is not fully satisfiable and whose satisfaction is increased by gathering evidence from multiple sources. In addition to charac- terizing unbounded requirements, we explain how to maximize their satisfaction by asking and by combining opinions of mul- tiple sources: people, services, information, and algorithms. We provide evidence of the existence of these requirements through examples by studying a modern Web application (Spotify) and from a traditional system (Microsoft Word).