Generating Interactive Prototypes from Query Annotated Discourse Models

Abstract

Model Based User Interface Development offers the possibility to design User Interfaces without being concerned about the underlying implementation. This is achieved by devising models at a high level of abstraction, thus creating the potential for involving users or domain experts to achieve a user-centered design process. Obtaining a running interactive application from such models usually requires several model transformations. One of the current problems is that while a user interface is generated after these transformations, other parts of the interactive system such as the application logic need to pre-exist or they must be written manually before the interface can be tested in a realistic scenario. This leaves the domain experts dependent on programmers and increases the time between iterations. In this paper we work with Query Annotations, which were previously used only for modeling at low levels and for generating fully functional interfaces, and we aim to generalize them for the high-level modeling approach called Discourse Modeling. The direct expected benefit of this generalization is the possibility to generate complete, readily testable interactive prototypes, rather than just their user interfaces. In addition, Query Annotations can serve as the mapping between the various levels of abstraction and bring to the domain experts a better understanding of the transformation process, as well as the possibility to modify the interfaces and models directly.

Declarative Interaction Towards Evolutionary User Interface Prototyping

“…Declarative interaction was inspired to the author through field observations and own experience [2,3]. Early versions of declarative controllers were demonstrated in [3] and refined in [9,11,12,13]. We have also shown a combination with MBUID approaches in [10].…”

“…The first generation of Makumba technologies, based on Java Server Pages (JSP) and pre-Web 2.0 user interaction, is by now outdated. Kis [11,12,13] has explored the possibilities of combining data from a multitude of APIs into one interactive application, rather than data from a single relational database, thus using data as a "prototyping material". This has resulted in Endev [12], a Javascript framework that is, however, not production-ready.…”

Rule-enhanced task models for increased expressiveness and compactness

“…A connection can be achieved either by creating a static link or with a query mechanism that provides loose coupling and a higher flexibility. Advantages of latter approach are discussed in [12], where database queries are used to connect elements of discourse models with domain model elements. In regard to semantic domain models, specialized query languages like SPARQL [10] are available to create a connection.…”