Although many visualization techniques have been shown to be pedagogically effective, they are still not widely adopted. The reasons include: lack of suitable methods of automatic-generation of visualizations, lack of integration among visualizations, and lack of integration with basic integrated development environment (IDE) support. To effectively use visualizations when developing code, it is useful to automatically generate multiple synchronized views without leaving the IDE. The jGRASP IDE provides object viewers that automatically generate dynamic, state-based visualizations of objects and primitive variables in Java. Such seamless integration of a lightweight IDE with a set of pedagogically effective software visualizations is unique and is currently unavailable in any other environment. Multiple instructors have reported positive anecdotal evidence of their usefulness. We conducted formal, repeatable experiments to investigate the effect of these viewers for singly linked lists on student performance and we found a statistically significant improvement over traditional methods of visual debugging that use break-points.

The jGRASP lightweight IDE has been extended to provide object viewers that automatically generate dynamic, state-based visualizations of data structures in Java. These viewers provide multiple synchronized visualizations of data structures as the user steps through the source code in either debug or workbench mode. This tight integration in a lightweight IDE provides a unique and promising environment for learning data structures. Initial classroom use has demonstrated the object viewers' potential as an aid to students who are learning to write and modify classes representing data structures.Recently completed controlled experiments with CS2 students indicate that these viewers can have a significant positive impact on student performance.

In this paper, a method for extracting class dependency information from Java class files is described. Advantages and disadvantages of using this method are discussed.The problems of virtual dependence and synthetic methods are explained, and solutions offered.A convenient user interface is presented for making use of the dependency information in the form of an interactive UML class diagram, which is automatically generated from Java class files. This interface is a component of the jGRASP integrated development environment.

A framework for producing dynamic data structure visualizations within the context of a lightweight IDE is described. Multiple synchronized visualizations of a data structure can be created with minimal coding through the use of an external viewer model. The framework supplies a customizable viewer template as well as high-level APIs to a graph drawing library and the Java Debugger Interface. Initial classroom use has demonstrated the framework's ease of use as well as its potential to as an aid to student learning.

No abstract

Many studies have been conducted into how programmers comprehend software. Models based on these studies have pointed to top-down, bottom-up, and mixed models for describing the comprehension process. In most of these studies, however, a key element of the experiment is largely ignored: the underlying nature of the software being examined by the programmers. This paper discusses software measurements and metrics that must be considered as factors when conducting comprehension studies.

A framework for producing dynamic data structure visualizations within the context of a lightweight IDE is described. Multiple synchronized visualizations of a data structure can be created with minimal coding through the use of an external viewer model. The framework supplies a customizable viewer template as well as high-level APIs to a graph drawing library and the Java Debugger Interface. Initial classroom use has demonstrated the framework's ease of use as well as its potential to as an aid to student learning.

jGRASP has three integrated approaches for interacting with its dynamic viewers for data structures: debugger, workbench, and text-based interactions that allow individual Java statements and expressions to be executed/evaluated. These approaches can be used together to provide a complementary set of interactions with the dynamic viewers. Data structure identification and rendering were tested by examining examples from 20 data structure textbooks. Controlled experiments with CS2 students indicate that the viewers can have a significant positive impact on student performance. The overall result is a flexible environment for interacting with effective dynamic data structure visualizations generated by a robust structure identifier.