6TH HCIV WORKSHOP 2010
IN COOPERATION WITH
From a human-centered perspective, the most important dimension of representation design is "fit" to the physiological, cognitive, and social aspects of interaction with people. Well-designed visualizations leverage a knowledge of human physiology by exploiting peripheral and ambient channels of sight rather than solely depending on foveal vision. From a cognitive point of view, they implement semantic primitives that are directly isomorphic to pertinent aspects of prediction and control, rather than relying on a human's mental calculations. Software agents can be used to focus the operator's attention, and in turn the operator further redirects the attention of the agent by policy-based constraints. Our experience to date in revolutionary cockpit design and high-tempo large scale real time network watch applications supports the claim that, by putting the human squarely in the center of representation design, incredible gains in performance and scalability become quite possible. In this talk, we will explain how we are applying these insights to help provide more effective maintenance of situation awareness and more efficient response in studies of large-scale cyber security problems.
RAIMUND DACHSELT, MATHIAS FRISCH:
Creating and interacting with visual representations of node-link diagrams (e.g., UML diagrams) is one key aspect of software engineering. Though there is a multitude of software modeling tools available, the handling of such graph structures is still difficult. For example, workspace is usually insufficient when software models become huge and complex, and interaction techniques are limited to traditional point and click input. Since structural editors are often conceived as constrictive and inflexible, ad-hoc free hand sketching on paper or whiteboards is often preferred. As a consequence, drawn diagrams often have to be remodeled in digital tools, which is time consuming. In order to overcome these problems, the talk discusses the usage of novel visualization and interaction techniques for the software modeling process. This includes the application of interactive displays - in particular tabletops - , the interaction through Zoomable User Interfaces and the integration of sketched diagrams by means of digital pen and paper technologies. As one example of our research, the promising combination of pen and multi-touch gestures for diagram editing on interactive displays will be explained. Based on a collection of user elicited pen and hand gestures, a gesture set was developed and implemented, which supports two prevalent mental models: structural editing and sketching.
When teaching programming basics to our freshman students, we see that many of them are facing difficulties that arise from the complexity of the software tool that is used as programming environment. In order to get deeper knowledge about the influence of the complexity of the programming environment on the students' learning success, we conducted a user study among our students. The objective of the study was to find out whether the complex user interface of most of today's programming tools rather inhibits or rather supports the process of learning a programming language. For that purpose we selected five programming tools of different complexities that were used to solve simple programming tasks. The talk presents the results of this user study.
Brainstorming and Discussion.
In this talk, I will give a brief overview of the young history of software visualization, starting from traditional tools to animate algorithms, over the change to focus on complex problems in software engineering to visual analytics of software. Hereby, human-centered issues played an important role in the past, but they are even more relevant for current and future directions in software development and maintenance. These issues are exemplified on the basis of our own visualization approaches and external visualization tools.
This talk presents an expert evaluation of a collaborative system in which knowledge can be created and shared globally. An analytical evaluation was conducted as a Participatory Group Evaluation, which involved the objectives and rationale behind the development of a prototype. An empirical evaluation consisted of two experiments: A bottom up evaluation of the usability of the interface, and a qualitative Cognitive Walkthrough approach followed by a User and Work Centred Evaluation of collaborative knowledge sharing in a common real life work task. This evaluation resulted in a large number of user requirements.
There are several aproaches in cognitive psychology dealing with cognitive load. On the one hand, there are approaches which aim to reduce the load on cognitive resources (especially memory) by modifying the nature of the tasks like cognitive load theory. On the other hand, the theory of distributed cognition supports offloading of cognitive processes onto external resources (like computers). The main aim of this theory is to design such resources in a way that they can assist human reasoning efficiently.
Visual Software Engineering tools manage large multidimensional data sets. In such a case, pure Information Visualization techniques reveals their limits in terms of expressiveness and scalability. In this talk we deal with the critical issue of visual clutter reduction, presenting a general framework for analyzing and reducing clutter in Infovis techniques, dealing with both large data sets and high cardinality data sets. We detail our proposal using two running examples, i.e., 2D scatterplots and parallel coordinates.
GERRIT VAN DER VEER, INGO WASSINK :
Interaction with complex visualization generally requires a complex design process. Sketches and mock-ups of interaction are itself a cheap way to represent the design ideas at an early stage. We will show a way to assess people's understanding from early envisioning, using the concept of mental models, and the technique of teach-back. We will illustrate the approach and show how this way of early user involvement not only indicates potential problems and needs for improvements, but, additionally, provides creative input from the professional users even if these are not interaction designers .
KAI WILLENBORG :
In a competitive world, software that sells must provide a good user experience in addition to being easy to use. However, sales people and user interface designers sometimes differ in what they mean by good user experience. In the area of dashboards sales people often think of visually appealing, simple designs that follow recommendations for MS PowerPoint slides whereas UI designers aim to provide standardized, informative overviews that guide the user. With some examples of dashboards for incident management, we will illustrate how these differing goals have influenced dashboard design at SAP, and examine how the dashboards have evolved from the first mockups to their final versions. Here we will not only focus on the evolving designs but also on the reasoning and the persuasion process behind them.