Discovering Speech Acts in Online Discussions: A Tool-supported method

The increasing participation of users of software applications in online discussions is attracting the attention of researchers in requirements elicitation to look at this channel of ommunication as potential source of requirements knowledge. Taking the perspective of software engineers who analyse online discussions, the task of identifying bugs and new features by reading huge threads of e-mails can become effort demanding and error prone. Recognising discussants’ speech acts in an automated manner is important to reveal intentions, such as suggesting, complaining, which can provide indicators for bug isolation and requirements. This paper presents a tool-supported method for identifying speech acts, which may provide hints to software engineers to speed up the analysis of online discussions. It builds on speech act theory and on an adaptation of the GATE framework, which implements computational linguistic echniques.