Catálogo de publicaciones - libros

In this article, we propose the concept of meta-User Interface (meta-UI) as the set of functions (along with their user interfaces) that are necessary and sufficient to control and evaluate the state of interactive ambient spaces. This set is -, since it serves as an umbrella beyond the domain-dependent services that support human activities in an ambient interactive space. They are -oriented since their role is to help users to control and evaluate the state of this space. We present a dimension space to classify, compare, and contrast disparate research efforts in the area of meta-UI’s. We then exploit the generative power of our design space to suggest directions for future research.

The success of model-based approaches to user interface design depends on the ability to solve the mapping problem as well as on the availability of tools able to reduce the effort of establishing and maintaining of links between models throughout the development life cycle. In this paper a tool supporting a small set of mapping rules is presented. The tool enables the designer to produce task model fragments at operational level based on the patterns of mapping between task and domain models. The task model fragments are generated in XML format and can be further loaded in task modeling tools like CTTE or Teresa.

In this paper we discuss techniques on how task models can enhance visualization of the usability test log. Evaluation of the usability tests is a traditional method in user centered design. It is a part of the methodology for design of usable products. We developed a tool for visualization of usability logs that uses the hierarchical structure of the task models to group and visualize observers’ annotations. This way of visualization is very natural and allows investigation of the test dynamics and comparison of participant’s behavior in various parts of the test. We also describe methods of visualization of multiple logs that allow for comparison of the test results between participants. For that purpose we present a new visualization method based on alignment of the visual representations of tasks, where binding between the task model and the log visualization is used. Finally, we present an evaluation of our tool on two usability tests, which were conducted in our laboratory and we discuss observed findings.

Today a large variety of mobile interaction devices such as PDAs and mobile phones enforce the development of a wide range of user interfaces for each platform. The complexity even grows, when multiple interaction devices are used to perform the same task and when different modalities have to be supported. We introduce a new dialog model for the abstraction of concrete user interfaces with a separate advanced control layer for the integration of different modalities. In this context, we present the Dialog and Interface Specification Language (DISL), which comes with a proof-of-concept implementation.

Model-based approaches are a suitable alternative to cope with the increasing complexity of eServices made available in the last years by eGovernment applications. However, up to now, only a few studies have investigated which are the requirements for notations and tools devoted to support eService modeling. The main goal of this paper is to make a state of knowledge on the specification of user activity and processes in eGovernment eServices. Our results advocated for a hybrid approach for modeling combining task models and process models.

Software engineering and human-computer interaction communities use methods, techniques and tools that are not easily integrated. Our work argues that the development process could be improved by providing the designer with models, languages and tools that provides a seamless integration of software engineering and human-computer interaction approaches. In order to achieve this goal we have developed a language the Interactive Message Modeling Language (IMML) to support the development of interactive systems. This paper presents and discusses the concepts and models that are the foundation of IMML. We also compare our process with traditional task-based perspectives.

Model-based approaches describe the process of creating UI models and transforming them to build a concrete UI. Developers specify interactive systems on a more abstract and conceptual level instead of dealing with low level implementation. However, specifying the various models is a complex and time consuming task. Pattern-based approaches encapsulate frequently used solutions in form of building blocks that developers may combine to create a user interface model. Thus they enforce reuse and readability and reduce complexity. In this paper we present a comprehensive framework that unites model-based and pattern-driven approaches. We introduce the “Patterns In Modelling” (PIM) tool, that implements this framework. We will demonstrate the functioning of the tool by using an illustrative example. We primarily focus on the creation of the task model and give a brief outlook how patterns will be applied to the other levels within the framework.

In current practice, user interface development is often based on a vague and undocumented design process, relying solely on the designer’s experience. This paper defines a pattern-based design process, which adds rigor to user interface design. The process is based on the notion of user variables to capture user requirements in a formal manner – based on discrete values that are amenable for tool support and automated analysis. Other context of use information is captured as context variables. Using these values as input, design patterns are selected to leverage best design practices directly into user interface development. Pattern-Oriented Design is then employed to derive a conceptual design, or early prototype, of the user interface. A case study with a Bioinformatics information site exemplifies the feasibility and applicability of this process.

We propose a model-based approach to integrate human error analysis with task modelling, introducing the concept of . Error Patterns are prototypical deviations from abstract task models, expressed in a formal way by a model transformation. A collection of typical errors taken from the literature on human errors is described within our framework. The intent is that the human factors specialist will produce the task models taking an error-free perspective, producing small and useful task models. The specialist will then choose from the collection of error patterns, and selectively apply these patterns to parts of the original task model, thus producing a transformed model exhibiting erroneous user behaviour. This transformed task model can be used at various stages of the design process, to investigate the system’s reaction to erroneous behaviour or to generate test sequences.

The benefits of using software design patterns have been widely reported. However, in order for user interface design patterns to achieve the same degree of success as software design patterns, it would be useful to document design pattern solutions using a representation language that can be readily transported into the definition or specification of the interactive solution. Moreover, patterns are fragmented, which may hinder the designers’ global comprehension about their design decisions. In this paper, we present a small study that illustrates the use of an interaction modeling language called MoLIC as a glue language, which binds together the design pattern solutions and novel design constructs (for which there are no patterns defined) into forming a whole interactive solution.