1998 Conference Attendees

I. Animation-Embedded Hypermedia Visualizations of Abstract Processes

In this project supported by NSF under the Collaborative Research on Learning Technologies program, we are investigating the design of novel visualizations of fundamental computer algorithms as self-learning aids.

II. Modeling

Cartoonist is a programming environment that allows children to create graphical simulations with a purely visual language. What sets Cartoonist apart from other systems is that its language, based on a combination of rules and temporal constraints, allows easy explorations of a wide variety of models by simply refining or generalizing complex behaviors by adding and removing constraint rules, respectively. Furthermore, compared to similar systems, a wide range of parallel behaviors can be described and behavior can be specified by describing what should not happen. The latter is made possible by the declarative nature of the programming language.

This is particularly important when teaching about dynamic systems because the causal behaviors are not generally available to visual perception. By using a visual modeling environments that allows these systems to be "designed", learners will have the opportunity to view what is normally invisible.

III. Design

Learning by Design (LBD) is a framework guiding the development of curriculum, teaching methodology and software. Having been strongly involved in LBD's development at the EduTech Institute, we consider it a valuable framework for integrating software support into the classroom. In collaboration with EduTech, we we are developing an integrated software support system for LBD that emphasizes scaffolding of cognitive and metacognitive activities.

Projects in Planning

I. Visualizations of Abstract Processes: Theory and Further Analyses

We plan to extend this research in three directions:

(a) A detailed micro-study of interaction and problem solving while a student is working with a hypermedia visualization: we plan to collect and analyze a very rich set of data (time-stamped key-strokes, mouse actions, spatio-temporal gaze patterns, pupil size measurements, verbal protocols and gestures) on a small number of subjects to investigate the ecology of interaction with such systems and sources of learning benefits.

(b) Refining and extending our theory of multimodal comprehension from visualizations, and the corresponding visualization design approach, by working in a different domain. For this, we are exploring a collaboration with Cindy Hmelo, an educational psychologist who graduated from the Learning Technology Center at Vanderbilt and is currently at the Learning Research and Development Center, University of Pittsburgh. We are looking into the domain of biological systems.

II Modeling and visualization tools

Cartoonist will be further developed in two different directions.

(1) The view of creating simulations by programming them is dangerous and the activity of programming must be replaced by the activity of modeling, which can be viewed as a design process. Thus, Cartoonist will be extended to provide a language that (a) can describe a wider range of graphical simulations, and (b) supports modeling by scaffolding it as a design process.

(2) Constraint rules introduced in Cartoonist can be used to describe processes. We want to apply these descriptions rules to the design and implementation of dynamic data structures for introductory programming classes. This will allow animated visualizations of data structures to be created by the students, thereby moving from interactive visualizations to constructive visualizations for learning.

III Educational Technology for Introducing Freshmen to Computers

In the year 2000, Auburn University will transition to a semester system, and a common first year is being planned for engineering freshmen. This curriculum will contain a course called "Introduction to computing for engineers". We want to see how our research and development experience can be applied to build an interactive knowledge and visualization environment available over the internet for this course.

Our designs depart from earlier work on algorithm animations in three significant ways:

(a) Animations are chunked according to the semantics of process to be visualized.

(b) Animations are embedded in a context and knowledge providing hypermedia information presentation system.

(c) An architecture that conveys process information using three distinct kinds of presentations - bridging analogies, magnified visualizations and bird's eye visualizations - within a framework that promotes self-explanations.

Preliminary experiments indicate a statistically significant improvement in algorithm understanding over textual presentations in two different groups of students at different levels of expertise (freshmen/juniors taking their first data structures and algorithms course and sophomores/seniors taking a more advanced course). We are poised to conduct more studies in 1998. This work shows the benefits of visualizations, conceived as animations immersed in contextual knowledge over simple animations, and begins to provide a methodology for the design of such visualizations to effectively communicate abstract process knowledge to students.

II. Modeling

Research based on related work with similar systems (e.g., Cocoa) and visualization as desctibed above.

III. Design

Research based on systems developed at EduTech including McBAGEL, WebCaMILE, SMILE and WebSMILE.