From: Eric N. Wiebe
Based on personal research notes
Dates at end of citation indicate date abstracted
Selection based on keywords:
- visualization
- perception
- CAD
Exported from EndNotes file:
PUR-RefLib1-12/20
1. Augustine, M., & Coovert, M. (1991). Simulation and information order as influences in the development of mental
models. SIGCHI Bulletin, 23(1), 33-35.
This study examined two characteristics thought to influence the development of effective mental models. The order of
presentation of a conceptual model and the use of simulation (animation). The study found that simulation had a positive
effect on the development of the mental model whereas ordering had no appreciable effect. This study examined two
characteristics thought to influence the development of effective mental models. The order of presentation of a
conceptual model and the use of simulation (animation). The study found that simulation had a positive effect on the
development of the mental model whereas ordering had no appreciable effect.
8/91.
2. Baecker, R. M. (1987a). Cognition and information processing theories. In R. M. Baecker & W. A. S. Buxton
(Eds.), Readings in human-computer interaction: a multidisciplinary approach (pp. 299-307). Los Altos, CA: Morgan
Kaufmann.
Chapter 8, 299-307
It is often said that we are in the middle of an information revolution....This is a myth. Instead, what the new technologies
have brought us is a data explosion. Data does not become information until it informs." The visual channel is a main
delivery vehicle for this data/info. There are four types of complementary sources for information on visual perception:
- psychology
- human factors
- art appreciation/interpretation
- graphic design
Some of the authors represented later in the chapter argue that perception is an active process. For instance, Haber and
Wilkinson say we attempt to perceive structure in the images we see.
The role that the graphic designer can play in HCI has been forcefully articulated by Aaron Marcus. The art side is well
represented by Arnheim who talks about how images can function as pictures, symbols, and signs. These are different
types of representations that an image can take and the choice of what level of meaning to give to an image is an
important consideration. Designers and artists also have some important things to say about color and spatial and
temporal arrangement.
Another important aspect visual displays is how evolving technology is always changing the possibilities of what is
displayed.
"Given the difficulty of the [interface design] problem, good progress will probably be achieved through the
multi-disciplinary collaboration of the technologist telling us what is possible, the psychologist telling us what not to do,
and the designer suggesting what to do."
Extensive bibliography. 8/91.
3. Baecker, R. M. (1987b). The visual channel. In R. M. Baecker & W. A. S. Buxton (Eds.), Readings in
human-computer interaction: a multidisciplinary approach (pp. 207-218). Los Altos, CA: Morgan Kaufmann.
Chapter 7, pp 207-218
This chapter introduces a few key cognitive issues and terms that play a role in user interface design. Because it is a
convenient metaphor, the discussion treats human cognition from the perspective of the human being as an information
processor.
- The issue of limited resources
Critical resources are resources that are required for completing a task. Limited resources are resources that are in short
supply relative to their need.
- Resource utilization of tasks
Tasks can be discussed in terms of the resources that they consume: both in terms of their type and quantity. There is a
tradeoff between resources consumed and the quality of the data. You can minimize resource consumption by increasing
the quality of the information. Improved display quality and graphic design pushes back signal data limits. Training and
appropriate mental models pushes back memory data limits.
- Cognitive load
The measure of the number of resource consumed is known as cognitive load. It is an important issue in both isolated
tasks and large systems and is affected by stress, learning time, inability to 'timeshare'.
- Interference
One way to reduce interference is to reduce competition due to high cognitive load. Wickens states that different sensory
modalities utilize different resources. By redirecting some of the data to a different sensory mode, interference can be
reduced.
- Problem Solving
Problem solving requires attention and a large number of resources. Keep in mind that if a computer is adopted to help
perform a particular task, the chances are that the task has already taxed a human's problem solving ability.
-Skill transfer and consistency
A clear goal is to accelerate the process whereby novices begin to perform like experts. This can be done in part by skill
transfer.
-Mental models, analogy, and metaphor
"Learners make use of analogy between the system being learned and the previously experienced. The interpretation and
use of analogy is based upon the metaphor with which the system is conceptualized." Keep metaphors simple, direct,
and most congruent with he way the system actually works.
-Compatibility
Having good spatial congruence between the interface layout and its functionality is important.
-Summary
Every system can be characterized by the skills necessary to master it. This is the prescriptive model. Users, in turn can
be described by the skills they bring to the task (the descriptive model). The gap between these two models can be
closed through good design and training. These are the two main components of cognitive engineering.
Extensive bibliography. 8/91.
4. Banchoff, T. F., & Strauss, C. M. (1978). Real-time computer graphics analysis of figures in four-space. In D. W.
Brisson (Eds.), Hypergraphics: visualizing complex relationships in art, science, and technology Westview Press, Inc.
The book is a collection of articles given at a symposium on hypergraphics: the visualization of n-dimensional space. The
book opens by talking about how geometry is one of the purest links between art and science.
The chapter is the compendium to a set of two films of computer animation shown at some of the Brisson-organized
exhibitions. The films are titled "The Hypercube: Projections and Slicing" and "Complex Function Graphs: Squaring and
Exponential Functions" The first film has the most immediate interest to me. In the first movie a cube is used to represent
a number of concepts about 4-space. Consider the following properties of "cubes" in n-space:
Dimension 2D 3D 4D
Coord. range (±1,±1) (±1,±1,±1) (±1,±1,±1,±1)
# of verts 4 8 16
# of edges/vert 2 3 4
# of edges 4 16 32
They point out that rotation of a 3D wire-frame can help (or hinder) perception of 3D form by giving depth cues and that
some axes of rotation give more information than others. No matter what dimension the object is the rotated forms will
have to be projected onto a 2D screen. They subjectively came to the conclusion that perspective rather than parallel
projection of rotating 4D forms gave the most information.
The figures must be seen to fully appreciate what a hypercube looks like. But they point out that they are still visually
intuitive. The hypercube is two separate 3-cubes (one red and one green) with 8 edges connecting their corresponding
verts. They do not discuss what if any importance there is the the relative placement of the 3-cubes to each other. I
assume the the connecting edges must be the same length as the sides of the 3-cubes.
Another technique that they explore is called "slicing". Where a plane of n-1 dimensions cuts through the n-cube.
Depending on the angle of attack, different forms will result. With a 3-cube, the resulting 2-plane can be squares,
rectangles, parallelograms or a 6 sided form. A hexagon is formed from a plane halfway down and normal to the long
axis of the cube. A 4-cube, in turn, will have a 3-plane (3-D solid) as its slice. Again, depending on the angle of attack,
differing forms will result. Animation of the successive slices in any dimension are very effective.
The second film explores using slicing and projection on complex (imaginary numbers) function graphs.
5/91.
5. Barfield, W., Lim, R., & Rosenberg, C. (1990). Visual enhancements and geometric field of view as factors in the
design of a three-dimensional perspective display. In Proceeding of the Human Factors Society - 34th Annual Meeting,
(pp. 1470-1473). It is the authors contention that 3-D perspective vs planview displays facilitate superior performance
in the judgement of spatial relationships among objects. In particular, when the person needs to judge both horizontal
and vertical information. With only a planview display, the person has to mentally reconstruct the third dimension. There
are also ways of increasing performance of 3-D perspective displays. The article explores two of these:
- The ability to rotate the scene in real time
- Use shaded rather than wireframe models
Their display used in this study is modeled after the one used by McGreevy and Ellis but incorporates a number of visual
enhancements.
Four main conditions were studied:
- static wireframe
- static shaded
- rotational wireframe
- rotational shaded
For the rotational scenes, the subject used a mouse to control the rotation of the scene.
No improvement was found in using shaded images over wireframe. The ability to rotate the scene only seems to have
helped in making elevational judgements. In general, the pattern of performance they found was similar to McGreevy and
Ellis.
6/91.
6. Booth, K. S., & et al (1987). On the parameters of human visual performance: an investigation on the benefits of
antialiasing. IEEE CG&A(Sept.), 34-41.
This article is part of an ongoing effort to identify parameters of real-time human performance for graphic workstations.
For real-time graphics there is often the trade-off of more realistic rendering vs. speed. This article looks at one
particular rendering technique, antialiasing, and attempts to set a threshold of human performance. Basically, why waste
computing power if the user performance is not going up. This is a distinctly separate issue from whether the user
perceives the image to be of higher quality or more realistic.
The authors come to the conclusion that the real-time requirements of cognitive psychology experiments provide a good
measurement of the "envelope" within which real-time systems must operate if they are to be successfully used by
humans. The method used in this experiment is inspired by the Shepard and Metzler mental rotation cubes. They are
shown at different resolutions and levels of antialiasing and ask the users to:
- Give a subjective grade to the image quality
- Say whether the object has nine or ten cubes in it (response time measured).
The authors found that users preferred the higher degree of rendering but there was a threshold past which a higher
degree of rendering did not help performance. Future directions are to look more into the interaction of spatial, temporal,
and chromatic artifacts introduced by real-time display. They give as an example the artifacts generated on an aircraft
controller's screen.
6/91.
7. Cunningham, S., & al, e. (1990). Visualization in science and engineering education. In G. M. Nielson & et al (Eds.),
Visualization in scientific computing (pp. 48-57). Washington: IEEE Computer Society Press.
Currently graphics are being used in the sciences and engineering primarily is display data from finished work. The use of
graphics in the earlier stages of research projects and in the classroom is minimal at best.
"A good foundation in visual thinking during college can give the next generation of scientists and engineers the ability to
take advantage of the visualization tools of the early twenty-first century for scientific applications and research." p49
"Animation, just now available, is an exciting opportunity for educational visualization." p52
"Although most instructors do not understand visual learning, cognitive scientists are working on models of learning and
understanding that include visual techniques." p53
"Most engineering students learn CAD as part of their studies, but more advanced computing tools are now part of
every practicing engineer's needs." p54
Courses in art and color theory are also important for developing student's visual skills.
Numerous software packages for the PC and the Mac are outlined in the article.
8/91.
8. Ellis, S. R. (1989). Visions of visualization aids: design philosophy and observations. In SPIE (Eds.),
Three-dimensional visualization and display technologies (pp. 220-227). Visualization tools are useful for understanding
naturally 3D databases like those used by pilots and astronauts.
"Dynamic visual displays can help understand how these parameters change with time and conditions but purely visual
analysis is dependent upon a subjective perceptual assessment of the display."
The author states the importance of trying to match the dimensionality of the display the that of the phenomena. Usually
this involves trying to compress higher dimensional data down.
Pilots need to make immediate decisions based on quick, subjective spatial impressions. Often intentional distortions can
be used to improve the chances of a correct response to the image.
8/91.
9. Haber, R. N., & Wilkenson, L. (1982). Perceptual components of computer displays. IEEE CG&A, 2(3), 23-35.
"An understanding of the psychological principles of visual perception can help computer scientists improve
man-machine communication."
"[According to Miller] The number of items a person can compare, retain, or respond to at one time is limited to about
seven unrelated items of information [chunks]." This rule seems to be unaffected by the size of the chunk. Therefore the
larger the chunks can be built, the more info can be retained or compared. This rule seems to be unaffected by the time
delay between presentation and response, the difference is that long term memory is unlimited by the quantity of info that
is retained. Miller's work was based on verbal/textual materials and the authors contend that visual data is not nearly as
limiting. Visual information is processed in visual terms alone and is governed by different rules. Each picture acts as its
own coherent chunk so that no additional chunking procedure is needed to make it unitary. Pictures can have easily
recognizable features that allow them to be perceived and stored in familiar global schemas.
"The implication of these two principles is that visual information conveys structure directly....It is this direct tie of
perceived structure to visual stimulation that undoubtably accounts for the excellent comprehension, retention, and
retrieval of visual information as compared to nonvisual information."
They give a number of suggestions of how these principles of perceptual organization can be used to represent structure:
- Spatial organization. They specifically mention the use of this technique to represent time.
- Visual features (e.g. color, texture, shading, etc.).
- Movement. We are able to integrate patterns of motion from many different parts of the display into a coherent
organization (e.g. rotation on a 2D display to reveal a 3D object). Movement can of course represent time as a 4th
dimension. Apparent motion is effective for locating discrepancies between two data sets. They mention flicker boxes
used by astronomers to locate anomalous points on photos.
- The 3rd dimension. They mention the problems that arise when perspective illusions alter the size and shape of the
data.
- Meaning. Visual displays have different meanings in different contexts which may enhance or interfere with the
interpretation of the data. They mention that easily recognizable geometric forms might be used to help chunk the data.
8/91.
10. Kaiser, M. K., & Proffitt, D. R. (1989). Perceptual issues in scientific visualization. In SPIE (Eds.),
Three-dimensional visualization and display technologies (pp. 205-211). "In order to develop effective tools for scientific
visualization, consideration must be given to the perceptual competencies, limitations, and biases of the human operator."
The computer display screen can be thought of as a special impoverished visual environment. It is often hard with the
display of data to remove all of the ambiguity associated with the interpretation of the display.
The article outlines two views about perception:
- we perceive the world accurately (Gibsonian). This is the common sense approach put forward by Gibson. We see the
world as a dynamic stream of information, and that this dynamism removes ambiguity. most illusion arise from static 2D
images.
- perception is inherently unreliable (constructivist). This view arises from scientific experiments showing that perception
must be augmented by cognitive processes to correctly interpret the scene.
These two theories lead us to different strategies for designing Sci Vis tools. The authors contend there are ways of
bridging the gap.
Dynamic imaging techniques can help remove some of the ambiguities found in impoverished displays. Simulated
observer movement can create depth cues whereas simulated object movement can reveal change in structure over time.
Rarely ever do you want to change both object and observer at the same time, it can be be very disorienting.
The difficulties of using color to map interval scales is raised. Color in ordinal scales must also be applied carefully.
The Gibsonian view of motion perception is outlined. Depth order specification is unambiguously specified with motion
as is 3D form. Object displacement relative to the observer's point of view is almost impossible with static cues or
conventions. When a number of objects are moving in a visual field, we attend to the relative motions of the objects and
the common motion the objects share.
"Relative rotations are used to perceptually define 3-D form, whereas common motions are residual to form analysis and
define observer relative displacements."
They mention the power of dynamic imaging for understanding kinematic events.
"Further, if my understanding of a phenomenon is contingent upon my appreciation of that structure, my understanding
will be greatly aided if I can perceptually experience the structure."
8/91.
11. Kuhn, W., & Egenhofer, M. J. (1991). CHI '90 workshop on visual interfaces to geometry. SIGCHI Bulletin,
23(2), 46-55.
Workshop focused on:
"...how can geometric information be presented and manipulated at the user interface, optimizing the user's 'memory'
requirements and 'execution speed'."
This issue was addressed from a number of different perspectives, including:
- cognitive psychology
- computer science
- engineering
- industrial design
Interesting comments on the HCI requirements for engineers and designers.
Participant Leone Barnett pointed to the importance of "context" to provide clarification and meaning. "Context could be
used to help hide complexity, but not necessarily require the removal of complex features. This could shorten the learning
curve on a complex system."
David Maulsby: "Fundamental to a user's mental model of activity, such as editing a geometric design, is the association
of context with action....The ultimate context is the user's intention, a hierarchy of goals." An interactive system should
provide more than just a primitive set of tools, they should be organized with the way they are used.
8/91.
12. Lund, Y. I. (1990). Visualization of three-dimensional objects using computer animation. CoED, 10(4).
A somewhat thin article on the use of animation in teaching engineering graphics at Iowa State U. Animations have been
created on an SGI workstation to assist in the teaching of projection theory and the visualization of 3D form. Hidden-line
removed wireframe and shaded models were used. Currently the animations are only being used in the freshman course
but there are plans to have students be able to create their own animations in a sophomore course. The author claims
that animation helps the students visualization skills and to understand complex phenomena such as problems found in
scientific visualization.
7/91.
13. Majchrzak, A., Chang, T., Barfield, W., Eberts, R., & Salvendy, G. (1987). Human aspects of computer-aided
design. Philadelphia: Taylor & Francis.
Sections of this book copied:
- Section 1.4. Effective use of CAD.
- Chapter 2. CAD hardware and graphics technology. A solid introduction to I/O devices and display technology.
- Section 4.2. CAD software-to-I/O device interface.
- Chapter 5. Workstation and interface design for CAD. Full treatment of such issues as human factors and ergonomics
of the workstation proper and the interface. Some discussion of mental models but most psych material left for Chapter
6.
- Chapter 6. Cognitive and perceptual aspects of CAD. Reviewed elsewhere.
- Chapter 8. Social and organizational aspects of CAD. Covers issues such as workplace changes resulting from CAD
and individuals' reactions to job changes.
Extensive bibliography. 10/91.
14. McCuistion, P. J. (1991). Static vs. dynamic visuals in computer-assisted instruction. EDG Journal, 55(2).
An experiment was conducted to determine which of two methods of presenting graphic images (static or dynamic)
would enhance spatial abilities. This would be measured through scores on performance tests and/or mental rotation
tests. The author cites literature stating that spatial abilities rests on the ability to develop mental schemas and it is
important to present material in a way that is easily internalized. The computer using computer-assisted instruction may
be a good vehicle for that. The results showed that students viewing static presentations achieved slightly higher scores
on the performance tests whereas the students viewing the dynamic presentation made larger gains on the mental rotation
tests.
Presented at EDGD Mid-year 1990.
Also Presented at ASEE Annual Meeting (abstracted below)
The article starts with the contention that the ability to "think in 3 dimensions" is critical in design related courses. This
ability is a good predictor of successful engineering and science careers. The development of a mental schema of spatial
manipulation must coincide with the ability to work with traditional EDG techniques such as orthographic projection.
The study was set up to look at the differences in the use of static and dynamic visuals in a Computer Assisted
Instruction (CAI) environment. Students were given a pre and post mental rotation test and their achievement scores in
class were compared. In addition, they were given an exit written survey.
The written comments showed the students preferred CAI over reading the textbook. Those testing low initially on the
mental rotations test gained more from dynamic visuals but did not necessarily achieve better in class.
7/91.
15. McWhorter, S. W., & et al (1990). Evaluation of 3-D display techniques for engineering design visualization. In
Proceedings of the ASEE Engineering Design Graphics Division mid-year meeting, (pp. 121-129). Tempe, AZ: This
article evaluated a number of different display techniques as to their utility in helping students understand the geometry of
the object. Orthogonal display was the only non-pictorial type display and was rated the most inferior.
Three other types of display techniques were explored using both stereoscopic and non-stereoscopic viewing technology
in a SGI workstation.
- wireframe
- flat shaded
- hidden line removed (HLR)
The stereo displays performed better than non-stereo displays and the display technique preferences by the students
were rated in the order shown above. The authors error in stating that the Necker illusion results from a HLR cube
lacking geometric information. The authors also state that the wireframe is less ambiguous than HLR because HLR has
incomplete geometry whereas wireframe shows it all. I find this hard to buy this argument. They state that stereoscopic
viewing introduces the orientation cue that a normal wireframe view is lacking and that's why it rated the best. More
believably, they explained the lack of preference for orthogonal views to the fact that none of the students had any
training in engineering graphics and found 3rd angle projection techniques confusing.
8/91.
16. Meyer, G. W., & Greenberg, D. P. (1980). Perceptual color spaces for computer graphics. (ACM) Computer
Graphics, 14(3), 254-261.
"Perceptually uniform color spaces can be a useful tool for solving computer graphics color selection problems."
The idea is to define a color system where there is a perceptually uniform distance separating all of the colors. They state
that most such systems are based on the CIE color notation system.
A number of color systems are explored including Munsell's and the Optical Society of America (OSA) standard.
Techniques such as false coloring are outlined. Color maps can be built from perceptual color systems such as Munsell
or OSA though their application can violate many of the restrictions of size and spacing of individual color samples. The
accuracy with which the user interprets the data has been called into question, but the author's contend that the technique
is very useful for showing trends.
Frequently cited base article. 8/91.
17. Muir, D. W. (1985). Computer animation in engineering. In T. L. Kunii (Eds.), Computer graphics: technology and
art (pp. 299-307). NY: Springer-Verlag.
The article is a concise overview of the advantages, uses, and features available in computer animation for the engineer.
The author contends that static imagery is restrictive and that animation allows the engineer's ideas come to life and
communicate more effectively. One area where animation is particularly effective is in the illustration of complex motion
and complex assembly.
Figures in article somewhat dated. 8/91.
18. Nielson, G. M. (Ed.). (1990). Visualization in scientific computing. Washington: IEEE Computer Society Press.
The book is a compilation of current articles on the titled subject. This book also contains an extensive bibliography both
current and historical, including all of the articles in the book. At the end of the bibliography is a Dirichlet Tesselation of
all of the citations and a listing a important conferences and workshops.
6/91.
19. Nishimura, Y., & Sato, K. (1985). Dynamic information display. Visible Language, XIX(2), 251-271.
This article explores the use of new media which allows the use of dynamic layout techniques by visual designers. The
goal of this research is twofold: first to identify the characteristics of the new media and secondly, to develop a
framework within which the designer can be assisted in typographic layout generation.
Its crucial to understand the different modes of communication possible. They can be classified as:
• Static - Dynamic
• Sequential - Presentational
This leads to four possible combinations:
• Static Sequential (i.e. a book)
• Static Presentational (i.e. a poster or a photograph)
• Dynamic Sequential (i.e. a movie or a TV program)
• Dynamic Presentational
The last is the subject of this article. Whereas dynamic-sequential is essentially a predetermined linear process, in a
dynamic-presentational mode there is a non-linear distribution of information. This latter mode allows the user to
determine the pattern of information presentation through interactive devices. The real time interaction with the media is
classified by Owen as either transposition or transformation. In a transpositional operation allows for the viewer to view
the model from different points in time and space. In a transformational operation, the viewer alters the features of the
model to see "what-if".
The information structure is a hybrid of both a hierarchical and a spatial system. First, the different components of
graphic information is organized into a hierarchy. Then an individual component can be arranged on a grid or similar
spatial arrangement system. This layout is a form of structural representation of the underlying information and defines the
nature of the system.
The viewer observes a three dimensional layout of typographic information: two dimensions shown on an individual
viewing screen and the third represented by time. Though static and dynamic information display share many variables,
there are some inherent only in this dynamic mode of display (i.e. blinking, zoom rate, output rate, etc.).
When design situations are complex, the designer's manipulation needs the assistance by logical rules. Both these logical
rules and the designer's decisions need to be incorporated into the system.
The time variable can be represented with a linear pattern whereas the other variables can be represented by a network
pattern. These patterns then give rise to a higher order pattern that defines the process.
2/87.
20. Norman, D. (1982). Learning and memory. San Francisco: W. H. Freeman.
This text was designed as an introductory text into the cognitive aspects of learning and memory for researchers both in
and outside of the psychological profession. The author sees himself closer to the perceptual than to the psychophysical
camp of cognition. He also has a strong interest in HCI and AI. Chapters read include:
- Chapter 1 - intro
- Chapter 2 - visual short-term memory.
- Chapter 3 - primary memory
- Chapter 4 - secondary memory
- Chapter 8 - organization structure of memory
- Chapter 9 - semantic networks
- Chapter 10 - schemas
- Chapter 11 - ritualization of behavior (scripts)
- Chapter 12 - mental images
- Chapter 15 - intro to learning
- Chapter 19 - HCI
7/91.
21. Norman, D. A. (1987). Some observations on mental models. In R. M. Baecker & W. A. S. Buxton (Eds.),
Readings in human-computer interaction: a multidisciplinary approach (pp. 241-244). Los Altos, CA: Morgan
Kaufmann.
"In interacting with the environment, with others, and with artifacts of technology, people form internal, mental models of
themselves and of the things with which they are interacting. these models provide predictive and explanatory power for
understanding the interaction."
This article elaborates on a few general observations about mental models.
- Mental models are incomplete.
- People's abilities to 'run' their models are severely limited.
- Mental models are unstable. Details of the system are forgotten esp. if the model is not used for a while.
- Models do not have firm boundaries.
- Models are 'unscientific'. People maintain 'superstitious' behavior patterns.
- Models are parsimonious. People are willing to trade off extra physical action for reduced mental complexity.
He makes some interesting points about modeling mental models. The researcher needs to distinguish between his
conceptualization of the the model and the actual model. Conceptual models are excellent for teaching and understanding
physical systems. Often, though, the correspondence between the mental model and the conceptual model is less than
obvious.
In the ideal world, there will be a close correspondence between the conceptual and mental models and that the
conceptual model will govern all aspects of the design of the physical system. Rarely does this happen.
8/91.
22. Pratt, M. J. (1989). Kinematic analysis. In J. Rooney & P. Steadman (Eds.), Principles of computer-aided design
Englewood Cliffs, NJ: Prentice Hall.
Chapter 11
Good intro to subject. Outlines the various levels of modeling sophistication that kinematics systems can incorporate:
- user has to guarantee legality of incremental movement and visually checks for conflict.
- system allows user to define constraints.
- system performs geometric interogation such as Booleans for collision checking.
- system manages movement of mechanism, with the most sophisticated being in a time-continuous manner.
Links need to be represented geometrically with a solid model rep being the ideal. Geometry needed for collision
checking of which there are a number of techniques:
- Discrete time steps
- Modeling swept out volumes of the space the link passes through
- Modeling in a continuous four dimensional manner.
Each of these have their one disadvantages. Discrete steps may miss collisions whereas swept volumes may identify false
collisions because it doesn't synchronize with time dimension. 4D modeling is the best but visually difficult to understand
and very computationally intenstive. Another factor that is very important is "near misses" that may cause collisions
depending on part fabrication tolerances and distortion under load.
Good illus. 5/91.
23. Rooney, J. (1989). Geometry in motion. In Principles of computer-aided design Englewood Cliffs, NJ: Prentice Hall.
Chapter 15
Further expands on Chapter 11 and the concept of developing tools to visualizing "dynamic" geometry. Since all objects
are embedded in both space and time, a system may require both a spatial and temporal description as it moves,
changes, or evolves in form. Of particular importance is the spatio-temporal interrelationships of all of the components of
a system. Indicates that spatio-temporal modelling research is still in its infancy.
A number of examples are explored where 1 or 2-D objects have a time dimension added. These examples are both
single body representations and multiple body interrelationships.
Good illus. 5/91.
24. Rosenblum, L. J., & Nielson, G. M. (1991). Guest editors' introduction: visualization comes of age. IEEE
CG&A(May), 15-17.
Intro to an issue that has a number of good articles in it. The guest editors are also the editors of an important collection
of scientific visualization articles published in1990. Many of the articles in this issue are follow-ups to work reported in
the book.
The editors outline past history, current issues, and future goals in scientific visualization research. They note that solids
modeling techniques have been very valuable in displaying 3D data sets, but that the issue of what to do with data sets of
higher dimensionality is much less mature. They also point out that understanding how cognitive perception operates will
help define the displays of the future.
6/91.
25. Salomon, G. (1990). New uses for color. In B. Laurel (Eds.), The art of human-computer interface design (pp.
269-278). Reading, MA: Addison-Wesley Publishing.
This article offers a collection of new interface ideas using color to explore. The article starts off with a number of
examples of why color is so hard to to use. For instance, colors are heavily context dependent, not only in terms of the
other colors around it but also in terms of individual differences in physiological and psychological response to color.
From there the article breaks down into two parts: how color can be used to impart information to the user (despite the
above mentioned problems) and how interfaces can be designed to choose their own colors. (The first one of these two
areas is of most interest to me.)
The dimensionality of color is often broken down into hue, saturation, and value. She states that hue-based codes are
mainly associated with qualitative rather than quantitative change, therefore not as useful. Where hues can sometimes be
used is when they are used for real-world metaphors (this could be dangerous).
Color can be used as a redundant cue, reinforcing what is already coded with shape or alphanumerics. Color could also
be used as a mnemonic; aiding in recognition or recall of information previously seen or indicating a location where it
could be found.
A number of examples are given as how color can be used to help visualize data and processes. It pointed out that using
color for data visualization doesn't depend on the all of the users to recognize and/or respond similarly to a specific
color; rather, it depends on human pattern recognition. Color can be used to identify both static and dynamic patterns in
the data. Outside of patterns, dynamic color changes can be a powerful tools for depicting temporal transitions.
Desaturating colors ( in tandem with defocusing shapes) can be used to imply spatial depth.
6/91.
26. Sanford, J., Barfield, W., & Foley, J. (1987). Empirical studies of interactive computer graphics: perceptual and
cognitive issues. In Proceedings of the Human Factors Society - 31st Annual Meeting, (pp. 519-523). This paper, like
the Booth paper, looks at the effects of computer graphics realism cues on cognitive tasks such as mental rotation. The
main goal (again) is to find out if there is a point of diminishing returns in terms of speed and accuracy of this task. A
secondary goal is to test two competing models -- the analog and propositional -- of how visual information is
represented in long term memory. The authors are supportive of a hybrid model that states different models are
appropriate at different times.
The method is modeled directly after the Shepard and Metzler mental rotation test. The subjects are shown moderately
complex pairs of objects and asked whether they are the same or mirror images of each other. The pairs are also rotated
between 0 and 180 degrees to one another, the theory being that the subjects will "mentally rotate" one of the pair until
they line up. To this classic setup, another variable was added: varying levels of rendering. They were:
- wireframe w/ hidden lines removed
- flat shading and single light source
- smooth shading and single light source
- smooth shading and two light sources
There results showed support for the hybrid model of visualization. They seem to see a break where at pairs opposed
120 degrees or less the analog model is used but at 180 degrees, the propositional model is used.
Solid shaded images yielded significantly better results than the wireframe in terms of both error rates and response
times. Distinctions between flat and smooth shading and one or two lights yielded mixed results though it seems that
computational power is better spent on two light sources rather than smooth shading as far as a person's perceptions are
concerned.
6/91.
27. Truckenbrod, J. R. (1981). Effective use of color in computer graphics. (ACM) Computer Graphics, 15(3), 83-90.
The article outlines that basic principles of color theory to provide a fundamental understanding of the characteristics of
color and how they can be effectively applied in computer graphic displays.
In b+w, we can only distinguish between 13 levels of gray from black to white, whereas we can distinguish between
many more colors.
"Color expands the capability of the visual composition in communicating ideas by providing more detailed information."
The 3D color models of Ostwald and Munsell are outlined and how hue, value, and croma can be manipulated with
these models. Alber's color theory is also introduced by stating that color is dynamically perceived as they are
continuously interacting with one another. Itten is quoted in a discussion on color contrast. It is pointed out that warm
colors advance (and grow larger) in space while cool colors retreat.
The author states that for color harmony, small areas of intense color are balanced against larger areas of less intense
colors (a central tenet of Tufte's too).
Frequently cited base article on color usage. 8/91.
28. Tufte, E. R. (1990). Envisioning information. Cheshire, CT: Graphic Press.
In a chapter entitled "Escaping Flatland", Tufte talks about how one can increase the density and dimensionality of the
information on the printed page (or on the computer screen I suppose). One of the strategies involves Small Multiples:
small, multiple diagrams indexed by time. Another way is to enhance the micro/macro properties of the diagram; allow
the user to both look at the "big picture" and have enough detail that the focused, specific information can also be
gathered.
The small multiple concept is expanded upon in a number of chapters. Basic rules of successful small multiples includes
using the same structure in all of the diagrams so that an economy of perception occurs. Once the viewer decodes the
design for one slice of data, it can be applied to all of the others. The focus can then be on the change in information, not
the graphical design. An example of dimensional expansion of this technique: there is a 2-D matrix of small multiple
diagrams; each diagram shows the two-space location of a third variable. In this way a total of 5 dimensions is
displayed. When the visual task is contrast, comparison, and choice; high information displays like this are appropriate.
The key is keeping all of the information within eyespan. Another important aspect of high-density designs are that it
allows the user control over synthesizing the data for themselves. Thus control of the information is given over to the
user.
Tufte also tackles issues of color and contrast. With small multiples and other types of diagrams, it is important to let the
important (usually dependent) variable be the focus of attention. Strong borders and grids usually distract us from the
information of interest. Almost all color systems describe color in three-space. This natural dimensionality can be used to
your advantage to increase the dimensionality of your information. He warns that high saturation colors and high contrast
combinations should be used sparingly. High saturation colors are wonderful for focusing attention against muted
backgrounds or counterpointed against colors like black.
6/91.
29. Weghorst, S., & Stuetzle, W. (1991). Jeepers: an interface perception research tool. SIGCHI Bulletin, 23(2),
75-80.
30. Wickens, C. D., & Todd, S. (1990). Three dimensional display technology for aerospace and visualization. In
Proceedings of the Human Factors Society - 34th Annual Meeting, (pp. 1479-1483). The similarities and contrasts
between scientific visualization and tasks imposed on the pilot and air traffic controller are highlighted. The authors
explore various methods of displaying higher dimensional data sets. Issues such as rendering techniques and relative
motion cues are brought up as is Wicken's proximity compatibility principle. The principle asserts that tasks requiring a
more integrative, holistic approach will benefit form a more "object-like" display whereas tasks requiring more focused
attention on a single dimension will benefit from digital/bargraph/separated displays. He wants to show the relevance of
this principle to scientific visualization.
The experiment showed that "object-like" displays only showed usefulness to the extent they produced emergent
properties that supported the task. The authors consider this work very preliminary with future work using more
complex and possibly dynamic data sets.
6/91.
31. Zsombor-Murray, P. J. (1990). 2-D and 3-D CAD: complements to visualization. EDG Journal, 54(3), 17-29.
The author begins the article with the proposition that engineering problems are conceived in design spaces of a higher
dimension in a parallel processing mode but are solved in subspaces of fewer dimensions in a sequential processing
mode. The iterative process leading to a solution can be thought of as a series of excursions between higher design
spaces and lower solution spaces. The author outlines visualization tools that enhance this process.
The last example given is that of the deformation of a cylindrical tube to form a tee joint. He suggests a number of
visualization techniques that enhances the change over time. The techniques include the stroboscopic superimposition of
an image sequence, spatial layout of a temporal sequence, a straightforward animation.
7/91.