Um estudo empírico sobre máquinas de tradução em tempo real para equipes distribuídas de desenvolvimento de software

Abstract

Distributed Software Development is increasingly present into the culture of information technology companies. The number of companies that spread its teams trying to reduce costs, improve products quality and improve productivity increases every year. This scenario, however, demands a huge cooperation between people that, in many cases, do not master the same language. A Speech Translation System is an alternative to this scenario, simultaneously translating from a language to another. This master thesis presents an empirical study, which consists of the historical review of the rise of recognition tools, translation and speech synthesis to its current state, as well as addressing technical characteristics of the same. The empirical research base has two experiments conducted in partnership with the University Aldo Moro of Bari, in Italy, using part of the tools available in the market and in development of two prototypes that make the integration of speech recognition, machine translation and speech synthesis to facilitate communication between distributed teams of software projects. The research contributes in order to show that the currently available technologies for communication between distributed teams that don't dominate the same language are close to be really effective and if they can be used in daily activities in software development teams. In addition to compatibility between tools, this research tries to point which the way forward to integrate voice Recognizers, Machine Translation and Speech Synthesis.