Section: New Results

Computer-mediated Social Communication Interoperability

Participants : Rafael Angarita, Nikolaos Georgantas, Valerie Issarny, Cristhian Parra Trepowski, Christelle Rohaut.

People increasingly rely on computer-mediated communication for their social interactions. This is a direct consequence of the global reach of the Internet combined with the massive adoption of social media and mobile technologies that make it easy for people to view, create and share information within their communities almost anywhere, anytime. The success of social media has further led – and is still leading – to the introduction of a large diversity of social communication services (e.g., Skype, Facebook, Google Plus, Telegram, Instagram, WhatsApp, Twitter, Slack, ...). These services differ according to the types of communities and interactions they primarily aim at supporting. However, existing services are not orthogonal and users ultimately adopt one service rather than another based on their personal experience. As a result, users who share similar interests from a social perspective may not be able to interact in a computer-mediated social sphere because they adopt different technologies. This is particularly exacerbated by the fact that the latest social media are proprietary services that offer an increasingly rich set of functionalities, and the function of one service does not easily translate -both socially and technically- into the function of another. As an illustration, compare the early and primitive social media that is the Email with the richer social network technology. Protocols associated with the former are rather simple and email communication between any two individuals is now trivial, independent of the mail servers used at both ends. On the other hand, protocols associated with today's social networks involve complex interaction processes, which prevent communication across social networks.

The above issue is no different than the long-standing issue of interoperability in distributed computing systems, which require to mediate (or translate) the protocols run by the interacting parties for them to be able to exchange meaningful messages and coordinate. And, while interoperability in the early days of distributed systems was essentially relying on the definition of standards, the increasing complexity and diversity of networked systems has led to the introduction of various interoperability solutions, among which the (Enterprise) Service Bus paradigm.

In the above context, we have specifically introduced the "social communication bus" paradigm so as to allow interoperability across computer-mediated social communication protocols. Our work is motivated by our research effort within the AppCivist project. AppCivist provides a software platform for participatory democracy that leverages the reach of the Internet and the powers of computation to enhance the experience and efficacy of civic participation. Its first instance, AppCivist-PB, targets participatory budgeting, an exemplary process of participatory democracy that let citizens prepare and select projects to be implemented with public funds by their cities [17]. For city-wide engagement, AppCivist-PB must enable citizens to participate with the Internet-based communication services they are the most confortable with. The need for interoperability in this context is indeed paramount since the idea is to include people in the participatory processes without leaving anyone behind. This has led us to revisit the service bus paradigm for the sake of social communication across communities, so as to gather together the many communities of our cities.

Our contributions span:

• Social communication paradigm: Based on the survey of the various forms of computer-mediated social communication supported by today's software services and tools, we have derived how the approaches to middleware interoperability may apply to social communication interoperability.

• Social Communication Bus architecture: We leverage the VSB bus (see § 6.2) that supports interoperability across interaction paradigms as opposed to interoperability across heterogeneous middleware protocols implementing the same paradigm. The proposed bus architecture features the traditional concepts of bus protocols and binding components, but those are customized for the sake of social interaction whose coupling differs along the social and presence dimensions.

• Social Communication Bus instance for participatory democracy: We have refined our bus architecture, introducing the Social-MQ implementation that leverages the RabbitMQ message broker. The resulting implementation has been integrated within the AppCivist-PB platform for evaluation.

In order to inform the further study of the "Social Communication Bus" paradigm, we have analyzed existing practices and supporting technologies promoting citizen collaboration. In relation with our work on the AppCivist-PB platform, our study has concentrated on Participatory Budgeting (PB) campaigns, with a special focus on US-related initiatives, as a mean to understand the current and future design space of ICT for participatory democracy. We then derived new design opportunities for ICT to facilitate citizen collaboration in the PB process, and by extension, to reflect on how these technologies could better foster deliberative decision-making at a scale that is both small and large.

This research is carried out in collaboration with the Social Apps Lab at CITRIS at UC Berkeley in the context of CityLab@Inria and Inria@SiliconValley.