CHI 2002 Poster Home Research Activities SAD Photos Misc

A Framework for Locomotional Design:
Toward a Generative Design Theory

Susanne Jul
University of Michigan
SJul@acm.org

A poster presented at CHI 2002. See also
Jul, S. (2002). A Framework for Locomotional Design: Toward a Generative Design Theory. ACM Conference on Human-Factors in Computing Systems, CHI 2002 Extended Abstracts, 862-863.

Poster Overview

Locomotion -- purposeful directed movement -- can be seen as fundamental to navigation. This yields

2. A conceptual framework of the rôle of locomotional design in navigational cognition, and of
3. The elements of locomotional design that affect navigational cognition, and
4. A set of principles to guide design,

leading, eventually, to a generative design theory (1) of navigational design.

1. Generative Design Theory

Generative design theory in the usability engineering cycle

The purpose of a generative design theory is to guide designers in generating viable design alternatives, much as music theory guides music composition.

Generative design theory is prescriptive of design rather than descriptive of behavior.

A generative design theory of human-computer interaction should

• Describe a set of behaviors and/or tasks
• Link performance on those behaviors/tasks to specific design elements
• Describe how to manipulate properties of those elements to affect performance

2. Locomotion: A Navigational Task

Locomotion -- purposeful directed movement
is in service of navigation -- finding and getting to places
which is in service of the user's task.

Locomotional design -- the support an environment offers for moving and directing movement
dictates problems of wayfinding -- spatial problem-solving and decision-making necessary to navigation
and determines (in part)

• what wayfinding problems must be solved,

• how complex they are,

• when they must be solved,

• what information is available to solve them,

• when information is available

• how much time is available to

  • perceive information

  • solve wayfinding problems

• what physical and cognitive resources are required to move and direct movement.

3. Locomotional Design Elements

Locomotional Structure

The network of locations at which movement can be stopped and routes along which movement is possible, (e.g., a freeway system and the towns to which it leads):

• Branch points -- Places where routes intersect, (e.g., exits on a freeway)

• Information points -- Places where information about locomotional structure becomes perceptible, (e.g., a hilltop from which a town and its freeway exits can be seen)

• Decision points -- Places where decisions can or must be made: A branch or information point

Locomotional Mechanism

The means of actually moving and of directing movement, (e.g., automobiles):

• Nature of movement -- Whether continuous or discrete, (e.g., driving vs. teleporting)

• Type of movement -- Whether ballistic or interruptible, (e.g., falling vs. flying)

• Timing of movement -- Speed, acceleration, etc.

4. Using Locomotional Design to Increase Wayfinding Performance

Derive basic locomotional structure and mechanism from navigational needs and the user's task:

• Locations and routes in the locomotional structure should correspond to locations and routes of the user's task, (e.g., hyperlinks to pieces of information within web pages)

• Locations and routes in the locomotional structure should be relevant to the user's task, (e.g., removing broken hyperlinks from a web page)

• The number of branch points in the locomotional structure should be minimized, (e.g., reducing the number of pages in a web site)

• The number of possible choices at branch points should be minimized, (e.g., reducing the number of options in a menu)

• "Natural" information points that allow locomotional structure to be perceived directly are preferred over artificial information points (e.g., created by signage), (e.g., "Go There" rather than "Click on arrow to go There")

• The number of decision points between an information point and the branch point about which it contains information should be minimized, e.g., context-sensitive help buttons)

• The speed at which a branch point is passed should be proportional to its number of options, (e.g., slowing the speed of a flyover as it is closer to the ground and more targets appear)

• The speed at which an information point is passed should be sufficient to perceive relevant information, (e.g., setting scroll rate so elements remain recognizable)

• The time from entry of an information point to departure from its associated branch point must be sufficient to process new information, but not allow information decay, (e.g., setting scroll rate so user doesn't overshoot target)

• The physical and cognitive resources required to follow routes and select branch point options should be minimized, (e.g., considering label area part of radio button)

5. Example: Inter-Object Navigation in Jazz

Jazz is a 2D multiscale environment with a zooming user interface

1. Objects are laid out on an infinite 2D surface

2. The surface can be viewed at an infinite range of magnifications

3. Objects can change their appearance depending on the magnification (scale) at which they are being displayed

View zoomed out Note photographs beginning to appear	View zoomed in Note word "Photograph" has disappeared and names have appeared	View zoomed in further

Conventional Locomotion

Conventional locomotion in Jazz is defined relative to the space itself

• Zoom is centered on the coordinate under the mouse
• Pan allows the view to be moved without changing its magnification

Task-Based Locomotion

Assumptions

• The user's task is to move from one object to another
• Locations are views of objects
• Routes are "straight" paths through space-scale
• Locomotional mechanism is zooming

Implication

• All movement should lead to views of objects

Design (Predictive Targeted Movement)

• On zoom-in, move towards the object nearest to the mouse
• On zoom-out, move towards the most magnified view that contains all objects (the Top of the World)

Zoom-in	Zoom-in	Top of the World

Experimental Results

Conventional and task-based locomotion was compared in an experiment using a within-subject design with repeated measures (24 subjects)

Differences in time usage (task-based locomotion relative to conventional locomotion)