presentation slides - The School of Electrical Engineering and

Download Report

Transcript presentation slides - The School of Electrical Engineering and

Laboratory for Interactive Learning Technologies
Department of Information and Computer Sciences
University of Hawai’i
Communicative Dimensions of
End-User Environments
Christopher Hundhausen
LILT Lab
University of Hawai’i
Honolulu, HI USA
[email protected]
Sarah Douglas
HCI Lab
University of Oregon
Eugene, OR USA
[email protected]
Introduction
Worthy goal: Design effective end-user
environments (EUE’s)
But what does “effective” mean?
Traditional view: Effective = enhanced human
performance (Green & Petre, 1996)
 Fewer
“programming games” and “hard mental
operations”
 Faster learning
 Better program comprehension (“role
expressiveness”)
 Low viscosity (easy to make changes)
Christopher Hundhausen ● Laboratory for Interactive Learning Technologies ● University of Hawai’i ● [email protected]
Introduction
(cont.)
Alternative View: Effective = Enhanced Human
Communication (Collaborative EUE’s)
 Education: facilitate construction of shared
knowledge (Roschelle, 1990; Suthers, 1999)
 Software/Web Design: help build design consensus
(Landay & Myers, 1995; Lin et al., 2000; Damm et al., 2000)
 Our work: Algorithm Visualization EUE’s
 Students construct and present their own visualizations
Research Questions
 How might EUE’s impact human communication?
 What specific design features make them well suited
to facilitating human communication?
 How might communication-supporting EUE’s differ
from conventional EUE’s?
Christopher Hundhausen ● Laboratory for Interactive Learning Technologies ● University of Hawai’i ● [email protected]
Talk Outline
1. Empirical Foundation: Ethnographic
Studies
2. Provisional Framework of
“Communicative Dimensions”
3. Design Implications
4. Conclusions
5. Future Work
Christopher Hundhausen ● Laboratory for Interactive Learning Technologies ● University of Hawai’i ● [email protected]
Ethnographic Studies
Studied Two Semesters of Undergraduate
Algorithms Course with Visualization
Construction and Presentation Assignments
Study I
 Used SAMBA (Stasko, 1997) to construct input-
general visualizations
 Required 33.2 hours on average (n = 20)
 Discussed programming toil in presentations
Study II
 Used simple art supplies (pen, paper, scissors, etc.)
to construct “one-shot” visualizations
 Required 6.2 hours on average (n = 20)
 Discussed underlying algorithm in presentations
Christopher Hundhausen ● Laboratory for Interactive Learning Technologies ● University of Hawai’i ● [email protected]
Communicative Dimensions
1.
Programming Salience
Extent to which programming act focuses on
relevant domain concepts
 Low-level (graphics) programming generates
discussion about programming details
 Domain-specific programming generates discussion
about domain concepts

2. Typeset Fidelity
Extent to which program resembles textbook figure
Sketched appearance (low fidelity) invites
discussion, whereas polished (high fidelity)
appearance discourages it
 See also Schumann et al.’s (1996) architecture study


Christopher Hundhausen ● Laboratory for Interactive Learning Technologies ● University of Hawai’i ● [email protected]
Communicative Dimensions
(cont.)
3. Story Content
 Extent to which program portrays domain in terms
of story
 Programs with story content stimulate livelier
discussion that purely geometric/textual programs
4. Modifiability
Extent to which program can be dynamically altered
 Highly modifiable programs are better able to
mediate discussions about domain concepts than
static programs; they facilitate audience
participation

Christopher Hundhausen ● Laboratory for Interactive Learning Technologies ● University of Hawai’i ● [email protected]
Communicative Dimensions
(cont.)
5. Controllability
 Flexibility with which a presenter can control a
program’s execution


Forwards and backwards execution
Jump to arbitrary point
 High controllability enables presenter to
dynamically respond to audience’s questions and
comments
6. Referencability
 Ease with which conversational participants can
refer to elements of program
 Highly referencabable program facilitates
communication by serving as resource for
disambiguating contextual references (e.g., “that”)
Christopher Hundhausen ● Laboratory for Interactive Learning Technologies ● University of Hawai’i ● [email protected]
Design Implications
Dimension
Inhibit
Communication
Support
Communication
Programming Salience Low-level prog.
Domain-spec. prog.
Typeset Fidelity
Polished graphics
Sketched graphics
Story Content
Prohibit storyline
Support storyline
Modifiability
recompilation
Dynamic modification
Controllability
Forwards only
Forwards + backwards
Referencability
Mouse pointer only
Dynamic mark-up
Christopher Hundhausen ● Laboratory for Interactive Learning Technologies ● University of Hawai’i ● [email protected]
Example EUE’s on Each End
of Continuum
Conventional: SAMBA (Stasko, 1997)
1. Annotate
extern MyAnimator football_bsort;
main(int argc, char *argv[])
{
int n,i,j,temp,a[50],count;
char str[100];
. . .
football_bsort.Ready(n);
for (j=n-2; j>=0; --j) {
football_bsort.StartOuter();
for (i=0; i<=j; ++i) {
football_bsort.Compare(i,i+1);
. . .
}
football_bsort.InPlace(j+1);
}
. . .
}
football_bsort::Compare(int i,
int j) {
//write SAMBA commands to
//script file
outfile << "{\n“ << "moveto"
<< player[i] << " "
<< step_up_loc[i] << "\n"
<< "moveto " << ball
<< " " <<
step_up_ball_loc[i]
<< "\n"
<< "}\n"
<< "delay 7\n";
}
2. Compile
3. Execute
Christopher Hundhausen ● Laboratory for Interactive Learning Technologies ● University of Hawai’i ● [email protected]
Example EUE’s on Each End
of Continuum
(cont.)
Communication-Supporting: ALVIS
(Hundhausen, 1999)
Christopher Hundhausen ● Laboratory for Interactive Learning Technologies ● University of Hawai’i ● [email protected]
Summary
End-User Environments Can Play
Important Role in Facilitating Human
Communication
Communicative Dimensions Provide
Important Extension to Green & Petre’s
Cognitive Dimensions
Limitations
Plainly preliminary!
Apply mainly to collaborative EUE’s (e.g.
learning and design environments)
Christopher Hundhausen ● Laboratory for Interactive Learning Technologies ● University of Hawai’i ● [email protected]
Future Directions
Empirically Establish Cause-Effect
Relationships Between Design Features
and Communicative Activity
Ground Firmly in Underlying Theory of
Communication
Expand by Consolidating with Other Lines
of Empirical Work [e.g., Suthers’s (1999)
“Representational Guidance” Hypothesis]
Christopher Hundhausen ● Laboratory for Interactive Learning Technologies ● University of Hawai’i ● [email protected]
Thank you
Christopher Hundhausen
[email protected]
http://lilt.ics.hawaii.edu/~hundhaus
Information and Computer Sciences
University of Hawai’i
Human-Computer Interaction
Software Engineering
Networking
http://www.ics.hawaii.edu
Educational Technologies
Artificial Intelligence
Multimedia