A carregar...
Publicação
Modelo de interação flexível num Editor de Modelos operado por voz e gestos
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 17.63 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
A Engenharia Orientada a Modelos (MDE) surgiu como uma metodologia, fundada em teoria e ferramentas, para desenhar e desenvolver sistemas de software. A MDE recorre a linguagens de modelação, desde propósito geral, como o UML standard, a linguagens de domínio específico. Estas são suportadas por editores de modelação, que tipicamente oferecem um paradigma gráfico e textual, concentrado na visão, contudo ignorando outras abordagens como voz e som que poderiam ser utilizadas num contexto industrial ou mesmo até por uma questão de usabilidade.Esta dissertação tinha a intenção de refazer o modelo de interação fixo de uma plataforma já existente, de nome Model-By-Voice, cujo propósito é construir editores de modelos que trabalham com linguagens de modelação utilizando apenas voz e som, trazendo flexibilidade na interação oferecida aos seus utilizadores quando estes fazem operações na mesma (criar elementos, ler, atualizar e apagar) e navegam nos seus modelos.Model-By-Voice utiliza as ferramentas Sphinx-4 e Google Cloud Speech para reconhecimento de voz e a ferramenta FreeTTS para sintetizar voz. O modelo de interação antigo,embora funcional, apresentava limitações relevantes ao nível da utilização prática da plataforma, uma vez que o utilizador tinha de treinar para utilizar a ferramenta. Este era um problema ao nível da produtividade e curva de aprendizagem uma vez que a adaptação a este tipo de plataformas/IDEs exige um tempo considerável.Desta forma, este trabalho reformulou o modelo de interação da plataforma Model-By-Voice, fazendo a plataforma adaptar-se ao utilizador e não o contrário.Como resultado deste trabalho, um protótipo de prova de conceito foi desenvolvido e foi realizada uma experiência piloto com utilizadores reais para averiguar a sua usabilidade.
Model-Driven Engineering (MDE) has emerged as a methodology, grounded in theory and tooling, to design and develop software systems. MDE makes use of modelling languages, ranging from general purpose like the standard UML, to dedicated modelling languages like domain-Specific Modelling Lamguages. Those are supported by modelling editors, that typically offer graphical and textual notations, concentrated on vision, but ignoring other approaches like voice and sound that could be used in industrial settiings or even for a matter of accessibility.This dissertation had the intent of remaking the rigid interaction model of an exist-ing platform named Model-By-Voice, which purpose is to build model editors that deal with modelling languages just using voice and sound, to bring felixibility in the offered interaction to its users when operating (create elements, read, update, and delete) and navigating through their models.Model-By-Voice utilizes the tools Sphinx-4 e Google Cloud Speech for voice recognition and the tool FreeTTS to synthetize voice. The previous interaction model, although functional, had relevant limitations related to the pratical use of the platform since the user had to train himself to use the tool. This was a problem in terms of produtivity and learning curve since it takes a considerable amount of time for the users to adapt to this kind of platforms/IDEs.Therefore, this work reformulated the interaction model of the platform Model-By-Voice to make the platform itself to adapt to the user and not otherwise.As a result of this work, a proof of concept prototype was designed and we performed an experimental pilot study with real users to learn of its usability.
Model-Driven Engineering (MDE) has emerged as a methodology, grounded in theory and tooling, to design and develop software systems. MDE makes use of modelling languages, ranging from general purpose like the standard UML, to dedicated modelling languages like domain-Specific Modelling Lamguages. Those are supported by modelling editors, that typically offer graphical and textual notations, concentrated on vision, but ignoring other approaches like voice and sound that could be used in industrial settiings or even for a matter of accessibility.This dissertation had the intent of remaking the rigid interaction model of an exist-ing platform named Model-By-Voice, which purpose is to build model editors that deal with modelling languages just using voice and sound, to bring felixibility in the offered interaction to its users when operating (create elements, read, update, and delete) and navigating through their models.Model-By-Voice utilizes the tools Sphinx-4 e Google Cloud Speech for voice recognition and the tool FreeTTS to synthetize voice. The previous interaction model, although functional, had relevant limitations related to the pratical use of the platform since the user had to train himself to use the tool. This was a problem in terms of produtivity and learning curve since it takes a considerable amount of time for the users to adapt to this kind of platforms/IDEs.Therefore, this work reformulated the interaction model of the platform Model-By-Voice to make the platform itself to adapt to the user and not otherwise.As a result of this work, a proof of concept prototype was designed and we performed an experimental pilot study with real users to learn of its usability.
Descrição
Palavras-chave
Model-By-Voice Interação Homem-Máquina Modelo de Interação Engenharia de Software Orientada a Modelos Modelling Workbench