Regresar

Towards conversational syntax for domain-specific languages using chatbots

Abstract:

Traditionally, users have interacted with computers through graphical or command line interfaces. However, these may still be too technical for certain users, or cumbersome to use in some scenarios (e.g., in mobility). To tackle this issue, recent advances in natural language (NL) processing have boosted the proliferation of chatbots: programs whose user interface is NL and are frequently integrated within social networks. In this paper, we explore the usage of NL as concrete syntax for domainspecific modelling languages, and propose an approach to automate the creation of modelling chatbots that converse with users to assist them in building domain-specific models. As chatbots are deployed on social networks, modelling becomes collaborative. We provide an implementation of our approach on top of Google's DialogFlow, and illustrate its usefulness on the basis of a case study to build and deploy streaming data applications using a conversational interface.