Pitt Researchers Aim to Illustrate the Invisible

PITTSBURGH—Two University of Pittsburgh professors have been awarded a grant to improve visualization tools for teaching computer programming. Their results, they say, will create better programmers and, in turn, better programs, with fewer loopholes for viruses and worms.

"We wouldn't let a substandard plumber work on our heating system, but people buy software all the time that is written by substandard programmers," said the project's coprincipal investigator Michael Spring, associate professor of information science at Pitt.

"The weaknesses in programs that allow users' machines to be violated can almost always be traced back to programming errors that were made by less than competent programmers," said Spring. "This research is designed at improving our ability to teach programming in a way that will insure that we turn out programmers who have A-level, rather than C-level, skills."

Spring and the principal investigator of the project—Peter Brusilovsky, assistant professor of information science at Pitt—were awarded a three-year, $220,000 grant from the National Science Foundation. The focus of the grant is to extend the power of the educational tool visualization, which allows the student to see inside an otherwise invisible process.

In computer science, what is typically seen on the screen is just what the program "prints." To see how programs actually work, it's necessary to create a visualization. "Good visualizing techniques allow you to see what's going on inside your computer," said Brusilovsky.

Brusilovsky and Spring are investigating how to make visualization more explanatory and more sensitive to students' differing levels of knowledge.

"Normal visualization is passive—you just look, and you're not really learning by looking, even if it's a nice visual," said Brusilovsky. "Students can see what's happening, but they can't relate it to their basic knowledge. Without narration, they don't understand what's going on. That's why we want to get explanatory visualization."

The researchers also want to make their tools more adaptive. Different students have different levels of knowledge and skill: Some need concepts to be explained in less detail, some more. The technology will compile information on a student's knowledge level from quizzes and build a profile of the student's skills. Then, whatever knowledge is new to that particular student will be explained in more detail.

In the last phase of the project, faculty members of local colleges and universities will teach courses at Pitt using the technology and then take the tools back with them. Brusilovsky and Spring also will hold workshops to demonstrate the usefulness of the tools.

The researchers say the same concepts and techniques could be used in many learning situations where the most important things people need to understand are invisible.

For more information, visit www.sis.pitt.edu/~peterb.

###

11/1/04/tmw