VARScope

1
VARScope
Personalized Guidance for Example Selection in
an Explanatory Visualization System
Dr. Peter Brusilovsky Gayathri
Krishnamoorthy School of Information
Sciences University of Pittsburgh, Pittsburgh,
PA October 15, 2006
2
Introduction

• Program Visualization key in computer science
  education
• Explanatory Visualization is one step ahead,
  augmenting visualization with natural language
  explanations.
• Explanations on the fly
• Pre-authored explanations (most common, easy to
  create)
• Growing tremendously every year

Problem to students Which is the best example
for me?
3
Focus of the Paper

• Solution for the example selection problem
• Personalized guidance using adaptive link
  annotation in the form of visual cues
• Way to enhance an existing explanatory
  visualization system with personalized guidance
  functionality
• Using an available student model server (eg.,
  CUMULATE)
• Providing an adaptive panel

4
VARScope - Features

• Variable Scope is one of the critical topics in
  introductory programming courses (Instructors
  survey)
• VARScope guides students in learning the concept
  of variable scope in C programming language by
  offering
• Personalized guidance in choosing the best
  example
• Visualizing the progress made on various concepts
  using progress bars
• Explanatory visualization of pre-authored
  examples

5
VARScope Interface
6
VARScope - Interface
Adaptive Panel
1
1. Annotated Links 2. Concept Grid 3. Progress
Bars
2
3
7
Concepts and Indexing

• 6 Domain concepts were identified
• Global, Local, Static, Block Scope, Parameter
  Extern
• Each example line was indexed with these concepts

8
Implementation
VAR Scope
Initiator Servlet
1. Html Link (user, group)
2. Initial Progress?
3. User, group, progress
6. Problem Suggestion , progress
USER MODEL
Applet
ADAPT2 Protocol
4. User events
5. User events
Relay Servlet
6. Progress, problems explored
9
Personalized Example Suggestion

• Example suggestion is influenced by
• Difficulty level of the example
• Proportional to the number of domain concepts
• Level 1, 2, 3 or 4, with level 1 on the easier
  side suitable for first-time exploration
• Users current knowledge of the domain concepts
• Comfort level of the user
• Calculated based on number of examples explored
  under each difficulty level
• Corresponds to one of the difficulty levels of
  the example

10
Personalized Example Suggestion

• Appropriate examples
• Examples unexplored under users current comfort
  level
• If all examples under current comfort level is
  explored, suggest examples from subsequent higher
  difficulty level
• In-appropriate examples
• Examples whose difficulty level is much higher
  than the users comfort level

11
Evaluation

• Questionnaire
• 6 users evaluated VARScope, shared their
  experience by filling out our questionnaire which
  captured users perception on
• Easy navigability
• Learning goal achievement
• Interface Design
• Ability to visualize progress
• Problem Suggestion
• Amount of Help provided
• Think-aloud
• 3 think-aloud studies
• were performed