NameYadav, Harsh (author), Marsic, Ivan (chair), Gruteser, Marco (outside member), Parashar, Manish (outside member), Rutgers University, Graduate School - New Brunswick,
SubjectElectrical and Computer Engineering,
Multimodal user interfaces (Computer systems),
DescriptionA User Interface (UI) is generally tightly coupled to the business logic of the application, and a great deal of modification is required to model it to separate the application logic. Manifold is an answer to such an issue. It offers a generic user interface framework where different application designers could develop separate application logic using the same Manifold front-end. Manifold employs the Model-View-Controller (MVC) design pattern to achieve its generic nature by providing separation of concerns between the UI and the application domain. The event frame concept incorporated on Manifold allows translation of user events to operations on application data via a central Controller object. The Controller object connects the Manifold front-end to the back-end application domain. This approach highly expedites the development of new features on the user interface, which remains decoupled from the rest of the application. It also allows building different applications on different servers and visualizing them on the Manifold UI. This thesis focuses on the enhancement of the Manifold framework both from the UI- and application-development perspective. New features based on industry standards are incorporated and the shortcomings present in the previous versions are eliminated to enhance the user-experience with Manifold. The flexibility of the Manifold framework allows development of applications with separate application logic and attaches them directly with the Manifold UI. The thesis developed applications and algorithms to visualize Manifold objects as fields in a relational database. The relational database approach allows saving the UI objects into a hierarchical data structure, rather than saving them as such on the file system. This provides added benefits of persistent information about objects, allowing data-analysis to be performed on them at various steps. This approach allowed extending the use of Manifold to areas such as text-, pattern- and speech recognition. Text recognition allows user to generate graphical objects (“glyphs”) on Manifold by typing in plain text data. Pattern recognition allows providing users with visual feedback if a collection of graphics in Manifold could be recognized as a particular pattern class. Speech Recognition allows the user to issue commands to the Manifold UI to perform specific tasks, which otherwise require manual manipulation. The newly developed features are analyzed for their structural and code complexity with direct repercussions to their re-use in future designs and performance optimization obtained by employing the best software engineering practices. The new features are evaluated with respect to their performance in the Manifold framework.
NoteIncludes bibliographical references
Noteby Harsh Yadav
CollectionGraduate School - New Brunswick Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.