An Overview of Algorithm Animation

1
An Overview of Algorithm Animation

• Xiaocong Zhou
• Nov. , 2007

2
Outline

• Basic Concepts
• Definition and goals of algorithm animation
• History and Some Important Systems
• Some features of the system
• Some snapshots of animations
• Features of Software Visualization Tools
• A famous taxonomy are given for showing main
  ideas in the area
• Further Reading and Resources

3
Basic Concepts

• What is visualization?
• The goals of visualization
• What is software visualization?
• The goals of software visualization
• What is algorithm animation?
• The goals of algorithm animation
• General Model of Software Visualization
• Structure and process

4
What is Visualization?

• Visualization is the process of transforming
  information into a visual form1
• It enables the scientist or engineer to perceive
  visually the features of information
• Scientific visualization vs. Information
  visualization
• Physic data vs. Abstract data
• Visualization plays a major role in the use of
  computer to support human reasoning
• Intelligence Amplification vs. Artificial
  Intelligence

1 S. Diehl. Software Visualization. Springer,
2007, ISBN 978-3-540-46504-1
5
What is Software Visualization?

• The visualization of artifacts related to
  software and its development1
• Visualizing the structure, behavior, and
  evolution of software
• Software visualization vs. Information
  visualization
• Static dynamic visualization vs. Static
  visualization
• The goal of software visualization
• To help to comprehend software systems
• To help to improve the productivity of the
  software development process

1 S. Diehl. Software Visualization. Springer,
2007, ISBN 978-3-540-46504-1
6
To Visualize Full Life-cycle of Software

• Software analysis design
• Translating textual document to graphic model
• Software Implementation
• Static data code visualization
• Pretty printing, visualizing the results of
  static analysis
• Software debugging testing
• Visual debugging testing
• Visualizing the test path
• Software maintenance
• Data code animation
• Visualizing software changes
• Software visualization and reverse engineering

7
A Narrow Definition of SV

• Visualization of algorithms and programs2
• Algorithm visualization
• Static algorithm visualization
• Algorithm animation
• Program visualization
• Static data visualization code visualization
• Data animation code animation

2 J. Stasko. Software Visualization. Slides of
a course named information visualization, 2006
8
What is Algorithm Animation?

• Algorithm visualization vs. Program
  visuali-zation3
• Algorithm Visualization is the visualization of a
  high-level description of a piece of software
• Often applied to education
• Program visualization is the visualization of
  actual implemented codes
• Often applied to software engineering
• Static algorithm visualization vs. Algorithm
  animation
• Algorithm animation is dynamic algorithm
  visualization
• Dynamic vs. static run or not run the program to
  be visualized

3 B. A. Price, R. Baecker, and I. Small. A
Principled Taxonomy of Software Visualization.
Journal of Visual Languages and Computing,
4(3)211-266, 1993
9
Sub-areas of Software Visualization
Abstract level
Software Visualization
Static algorithm Visualization
Algorithm Animation
Algorithm Visualization
Static code data Visualization
Code data Animation
Program Visualization
Static
Dynamic
10
Why Do People Animate Algorithm?

• Different people have different motivations for
  animating algorithms1
• Understanding and teaching
• Teachers visualize algorithms to better explain
  them to their students.
• Design and communication
• Developers visualize algorithms to better
  communicate the ideas to other experts.
• Optimization
• Developers visualize algorithms to better
  understand how they work and find possibilities
  to enhance them.
• Debugging
• Programmers use visualizations to find faults in
  their algorithm implementation

1 S. Diehl. Software Visualization. Springer,
2007, ISBN 978-3-540-46504-1
11
Why Do We Focus on Algorithm Animation?

• It is relatively easy to study and implement
• The structure and behavior of algorithm are clear
• The scale of algorithm implementations are small
• Algorithm animation is a pilot study of software
  visualization
• High-level visualization is easier to understand
• Software visualization is ultimately required to
  be high-level, since it is used to analysis and
  understand software
• Many methods and techniques can be extended to
  program visualization
• The talk focus on algorithm animation but is not
  limited to algorithm animation

12
General Model of Software Visualization
Programmer
Visualizer
Implement
Specify
Interact with
Software to be Visualized
User
Visualization
Mapping
Specify Support
Interaction Support
Rendering Support
Software Visualization (Aid) System
Develop
Developer
13
General Model of Software Visualization

• Three modules
• Software to be visualized target software
• Visualization of the target software
• Software visualization system helps to develop
  the visualization of the target software
• A key task
• Mapping the information of software to visual
  representation, and then generate the
  visualization of the software
• Four roles
• Programmer who develops the target software
• Developer who develops the software visualization
  system
• Visualizer who specifies how to visualize the
  target software
• User who interacts with the visualization of the
  target software

14
A Model of Algorithm Animation

• The specification of mapping forms a visual model
• Do not give the roles and algorithm animation
  system

1 S. Diehl. Software Visualization. Springer,
2007, ISBN 978-3-540-46504-1
15
Visual Model vs. Formal Model

• Software visualization gives a visual model of
  software
• Intuitive, easy to understand
• To help user analyze and understand the software
• Formal semantics gives a mathematical model of
  software
• Formal, accurate
• To help user analysis and verify the software
• How to combine these two models?
• Visual verification?
• Verified visualization?

16
Software Visualization Pipeline
1 S. Diehl. Software Visualization. Springer,
2007, ISBN 978-3-540-46504-1
17
The Process of Visualizing

• Data acquisition
• Extract and gather relevant data from various
  sources of information about a software system
• Its source code, its design, user documentation,
  state changes during its execution, test result,
  etc.
• Analysis
• Various kinds of analysis can be used to reduce
  the amount of data and to focus on the important
  parts
• Filtering, static program analysis, or
  statistical methods, etc.
• Visualization
• The resulting data is mapped onto a visual model
• That is, transformed into geometrical and
  graphical information
• Then is rendered onto the screen or some other
  kind of medium as a single image or series of
  image.

1 S. Diehl. Software Visualization. Springer,
2007, ISBN 978-3-540-46504-1
18
History and Some Important Systems
Sorting out Sorting
1980s
BALSA BALSA-II Zeus Zeus3D CAT JCAT
TANGO X-TANGO POLKA POLKA3D
1990s
JAWAA Jeliot JHAVE DsCats ANIMAL PAVANE LEONARDO
2000s
19
Sorting out Sorting

• 30 minute video produced by Ron Baecker at
  Toronto in 1981
• Seminal work in area
• It is often considered as the beginning of the
  research on algorithm animation (and SV)
• Illustrates and compares nine sorting algorithms
• Stick view is used to explain the algorithms
• Insertion sort Linear inserting, binary
  insertion, shell sort
• Exchange sort bubble sort, shaker sort, and
  quick sort
• Selection sort straight selection, tree
  selection, heap sort
• Dot view shows a race of all nine algorithms

20
SoS Views
2 J. Stasko. Software Visualization. Slides of
a course named information visualization, 2006
21
BALSA and Its Offspring

• BALSA (Brown Algorithm Simulator and Animator)
• The first well-known computerized system
• Produced by Marc Brown at Brown University
• Introduces the concept of Interesting events
• The relevant actions performed by the algorithm
  that are interesting for visualization purpose4
• Supports multiple dynamic view
• Its Offspring
• BALSA-II (extended with step and break points and
  a number of other features)
• Zeus, Zeus3D
• CAT (web-based), JCAT (Java based system)

4 C. Demetrescu, Irene Finocchi, and John T.
Stasko. Specifying Algorithm Visualization
Interesting Events or State Mapping. In S. Diehl
(ed.) Software Visualization, LNCS 2269, pp16-30,
2002
22
Example Animation of BALSA
2 J. Stasko. Software Visualization. Slides of
a course named information visualization, 2006
23
TANGO and Its Offspring

• TANGO (Transition-based Animation Generation)
• Produced by J. Stasko at Brown University
• Actually an interpreter for animation commands
• The algorithm code can produce commands defined
  in a language called SAMBA.
• These commands was then fed post-mortem into the
  TANGO interpreter to produce the animation
• Introduces the path-transition paradigm and
  supported smooth animation
• URL http//www-static.cc.gatech.edu/gvu/softviz/
  
• POLKA (Parallel program-focused Object-Oriented
  Low Key Animation )
• Supports to animate parallel algorithm

24
Example Animation of TANGO(1)
2 J. Stasko. Software Visualization. Slides of
a course named information visualization, 2006
25
Example Animation of TANGO(2)
1 S. Diehl. Software Visualization. Springer,
2007, ISBN 978-3-540-46504-1
26
Example Animation of TANGO(3)
1 S. Diehl. Software Visualization. Springer,
2007, ISBN 978-3-540-46504-1
27
Example Animation of TANGO(4)
1 S. Diehl. Software Visualization. Springer,
2007, ISBN 978-3-540-46504-1
28
Pavane Leonardo

• Pavane Developed at Washington University in
  19923
• Designed to provide three-dimensional
  visualizations of concurrent programs
• Visualization is specified in a declarative style
• The visualizer defines a transformation between
  the state of the algorithm and the final images
• the properties of a program which are necessary
  to verify are also properties which should be
  visualized
• Leonardo http//www.dis.uniroma1.it/demetres/Leo
  nardo/
• An integrated environment for developing,
  executing, and visualizing C programs
• It supports full reversible execution of C
  programs
• The latest version will support to run on Windows
• Visualization is specified by attaching graphical
  representations to key variables in algorithm

3 B. A. Price, R. Baecker, and I. Small. A
Principled Taxonomy of Software Visualization.
Journal of Visual Languages and Computing,
4(3)211-266, 1993
29
Example Animation of Leonardo(1)
http//www.dis.uniroma1.it/demetres/Leonardo/
30
Example Animation of Leonardo(2)
http//www.dis.uniroma1.it/demetres/Leonardo/
31
Other Systems Examples

• JAWAA(2003)
• http//www.cs.duke.edu/csed/jawaa2/
• DsCats(2002)
• http//www.cs.arizona.edu/projects/dscats/dscats.h
  tml
• ALVIS(2002)
• http//eecs.wsu.edu/veupl/soft/alvis/index.htm
• MatrixPro(2004)
• http//www.cs.hut.fi/Research/MatrixPro/
• Jeliot 3(2004)
• http//cs.joensuu.fi/jeliot/
• JHAVE(2000)
• http//jhave.org/
• Animal(2002)
• http//www.animal.ahrgr.de/index.php3
• The following URLs collect more example
  animations
• http//www.cs.hope.edu/alganim/ccaa/
• http//www2.hig.no/algmet/animate.html
• http//www.animal.ahrgr.de/Anims/animations.php3

The year is not refer to the system development
date but the important publications date
32
Features of Software Visualization Tools

• A taxonomy is defined to classify SV tools
• To define a structure of characteristics and
  functionalities of SV systems in a systematic way
• To show main ideas of SV systems
• To help visualizer to choose a proper SV tool for
  his own purpose
• The most well-known taxonomy is proposed by Price
  et al. in 19933
• I will give the taxonomy in detail, but I do not
  follow it faithfully
• it is already somewhat outdated and should be
  extended and updated

3 B. A. Price, R. Baecker, and I. Small. A
Principled Taxonomy of Software Visualization.
Journal of Visual Languages and Computing,
4(3)211-266, 1993
33
Overview of the Taxonomy
5 V. Karavirta. XAAL - Extensible Algorithm
Animation Language. Master's thesis, Department
of Computer Science and Engineering, Helsinki
University of Technology, 2005
34
Six Main Categories

• Scope
• What is the range of programs that the SV system
  may take as input for visualization?
• Content
• What subset of information about the software is
  really visualized by the SV system?
• Method
• How is the visualization specified?
• Form
• What are the characteristics of the output of the
  visualization?
• Interaction
• How dose the user interact with and control the
  visualization?
• Effectiveness
• How well does the visualization communicate
  information to the user?

3 B. A. Price, R. Baecker, and I. Small. A
Principled Taxonomy of Software Visualization.
Journal of Visual Languages and Computing,
4(3)211-266, 1993
35
Category Scope

• Generality
• Platform hardware operating system
• The early systems often run on a specific
  platform, but today a system should be platform
  independent
• Language paradigm language
• Imperative, object-oriented, logic (TPM Prolog),
  functional (KIEL Standard ML)
• Most popular languages Java, C
• Application Domain
• Concurrent distributed (Polka, PAVANE, VADE),
  computational geometry (GeoWin), network protocol
  (Polka-Rc)
• Scalability
• Program
• What is the largest program it can handle?
• Data Set
• What is the largest input data set it can handle?

3 B. A. Price, R. Baecker, and I. Small. A
Principled Taxonomy of Software Visualization.
Journal of Visual Languages and Computing,
4(3)211-266, 1993
36
Category Content

• Software life-cycle
• Analysis design, implementation, deployment
• Information type
• Software documents, codes, and data
• Abstraction level
• Concrete implementation, software model, and
  software property (e.g. invariant)
• Information focus
• Static structure focus, run time behavior focus,
  and evolution focus

37
Category Method

• Event driven and data driven
• Interesting events and state mapping
• To Apply to program visualization
• In particular, to utilize in code and data
  animation
• How to specify the mapping?
• Instrument code directly
• Hard coding, using special editor, and using
  preprocessor
• Build a pre-defined model for specifying the
  mapping
• A language is designed for specifying the mapping
• Automatically choose action or state to be
  animated
• UPWI (using expert system), Jeliot
  (semi-automated)

38
Category Method (cont.)

• When to generate the visualization?
• Live or post-mortem
• Evaluation of the mapping
• Intelligence
• To what degree can the visualization generate
  automatically?
• Customization
• To what degree can the user customize the
  visualization?
• Code Ignorance Allowance
• how much knowledge of the program code is
  required for generating the visualization?
• System-Code Coupling
• How tightly is the visualization system coupled
  with the code?

39
Category Form

• Medium What is the primary target medium?
• Paper, film, videotape, or display
• Currently, the focus is on web deployment
• Presentation Style What is the general
  appearance of the visualization?
• Graphic vocabulary
• Graphical primitives (pixel, line, rectangle,
  etc.)
• Color, fill style, font
• Dimension 2D, 3D
• Animation
• Sound
• Zeus and LogoMedia have made effective use of
  sound

3 B. A. Price, R. Baecker, and I. Small. A
Principled Taxonomy of Software Visualization.
Journal of Visual Languages and Computing,
4(3)211-266, 1993
40
Category Form (cont.)

• Granularity
• To what degree does the system present
  coarse-granularity details?
• Coarse-granularity details mean the visualization
  is abstract
• Elision
• To what degree does the system provide facilities
  for eliding information?
• That is, the user can move between coarse and
  fine granularity
• Multiple Views
• For example, stick view and dot view of sorting
  algorithm are showed simultaneously
• Program Synchronization
• Comparing the execution speeds of several
  programs by running a race

3 B. A. Price, R. Baecker, and I. Small. A
Principled Taxonomy of Software Visualization.
Journal of Visual Languages and Computing,
4(3)211-266, 1993
41
Category Interaction

• Style
• What method does the user employ to give
  instructions to the system?
• on-screen buttons, menus, command line
  statements, or scripted programs
• Navigation
• Elision Control
• Elide information or suppress detail
• Temporal Control
• Direction forward, backward
• Speed start, stop, speed control, replay
• Scripting Facilities
• Recording and playing back of interactions with
  the visualization

3 B. A. Price, R. Baecker, and I. Small. A
Principled Taxonomy of Software Visualization.
Journal of Visual Languages and Computing,
4(3)211-266, 1993
42
Category Effectiveness

• Purpose
• For what purpose is the system suited?
• Software engineering, education
• Appropriateness and Clarity
• How rapidly do the visual metaphors inspire
  understanding?
• Empirical Evaluation
• To what degree has the system been subjected to a
  good experimental evaluation?

3 B. A. Price, R. Baecker, and I. Small. A
Principled Taxonomy of Software Visualization.
Journal of Visual Languages and Computing,
4(3)211-266, 1993
43
Further Reading and Resources

• Anthology textbook
• J. T. Stasko, J. Domingue, M. H. Brown, and B. A.
  Price. Software Visualization Programming as a
  Multimedia Experience. MIT Press, 1998
• S. Diehl, editor. Software Visualization, volume
  2269 of LNCS State-of-the-art Survey. Springer
  Verlag, 2002
• S. Diehl. Software Visualization. Springer, 2007

44
Further Reading and Resources

• Related conference
• ACM Symposium on Software Visualization (SOFTVIS)
  
• ACM Annual Conference on Innovation and
  Technology in Computer Science Education (ITiCSE)
• Program Visualization Workshop (PVW)
• IEEE Symposium on Visual Language and
  Human-Centric Computing (VL/HCC)
• IEEE International Workshop on Visualizing
  Software for Understanding and Analysis (VISSOFT)

45
Further Reading and Resources

• Related Journal
• IEEE Transactions on Visualization and Computer
  Graphics (TVCG)
• Journal of Visual Languages and Computing (JVLC)
• Related People
• J. Stasko http//www-static.cc.gatech.edu/gvu/sof
  tviz/
• S. Diehl http//www.st.uni-trier.de/diehl/
• C. Demetrescu http//www.dis.uniroma1.it/demetre
  s/
• A. Kerren http//www-hagen.informatik.uni-kl.de/
  kerren/
• I. Finocchi http//www.dsi.uniroma1.it/finocchi/
  
• R. Sedgewick http//www.cs.princeton.edu/rs/
• V. Karavirta http//www.cs.hut.fi/vkaravir/

46
Reference

• S. Diehl. Software Visualization. Springer, 2007
• J. Stasko. Software Visualization. Slides of a
  course named information visualization, 2006
• B. A. Price, R. Baecker, and I. Small. A
  Principled Taxonomy of Software Visualization.
  Journal of Visual Languages and Computing,
  4(3)211-266, 1993
• C. Demetrescu, Irene Finocchi, and John T.
  Stasko. Specifying Algorithm Visualization
  Interesting Events or State Mapping. In S. Diehl
  (ed.) Software Visualization, LNCS 2269, pp16-30,
  2002
• V. Karavirta. XAAL - Extensible Algorithm
  Animation Language. Master's thesis, Department
  of Computer Science and Engineering, Helsinki
  University of Technology, 2005
• Andreas Kerren and John T. Stasko, Algorithm
  Animation Introduction, In S. Diehl Ed.,
  Software Visualization, LNCS 2269, pp.1-15,
  Springer-Verlag, 2002

47
Comments Suggestions