CS 524 Lecture 1

Back in 1987 the field of scientific computing was organized through the Workshop on Visualization in Scientific Computing, held February 9-10, 1987 in Washington DC, and then disseminated through the report from that workshop - 1987's Visualization in Scientific Computing.

available in pdf format here

Today this workshop report still provides a good overview of the field, and looking back at the predictions made over 20 years ago should give you a better idea about the predictions made a few years ago in the reports we will be looking at next.

But first, a quick primer on 1987:

A typical desktop computer had:

10 mhz CPU
512 Kilobytes RAM
640 x 480 displays
800 Kilobyte floppy drives for storage
first 'laptops' were coming out and hard drives started to become available

a Supercomputer like the Cray-2 from 1985 with 4-8 processors, 2 to 40 gigabytes of disk storage, and 512MB to 4 gigabytes of RAM could do 500 Megaflops per CPU (millions of floating point operations per second) and cost roughly $15,000,000. 27 were sold.

Today's CPUs can do 100 gigaflops and GPUs can go to 500 Gigaflops to 1000 Gigaflops and supercomputers are reaching 2 petaflops (2,000,000 Gigaflops) ... so our desktop machines are 25 to 250 times faster than the supercomputers of the time, and our supercomputers are 500,000 times faster than the supercomputers of the times.

There was no public internet, but we had 1200 bits/sec modems to call bulletin board systems, or CompouServe, or log into mainframe computers over the phone lines, but those 'on-line' services were not connected to each other.

Software was distributed on floppy discs or CD-ROMs (680 megabytes). Internal hard drives could hold 20-40 megabytes.

Most music was being bought on casette for people to play on their Sony Walkmans. LP records had almost disappeared but CDs were only a couple years old and starting to catch on.

NES was the main videogame console (2 Mhz CPU, 2 kilobytes of ram, but game 'lived' on the 16K cartridge)

Now back to the report.

Visualization is a method of computing. It transforms the symbolic into the geometric, enabling researchers to observe their simulations and computations. Visualization offers a method for seeing the unseen. It enriches the process of scientific discovery and fosters profound and unexpected insights. In many fields it is already revolutionizing the way scientists do science.

Visualization embraces both image understanding and image synthesis. It studies those mechanisms in humans and computers which allow them in concert to perceive, use and communicate visual information Visualization unifies the largely independent but convergent fields of:

Richard Hamming observed many years ago that "The purpose of computing is insight, not numbers." The goal of visualization is to leverage existing scientific methods by providing new scientific insight through visual methods.

An estimated 33-50% of the brain's neurons are associated with vision. Visualization in scientific computing aims to put that neurological machinery to work.

Today's data sources are such fire hoses of information that all we can do is gather and warehouse the numbers they generate.

High-volume data sources include:

There is every indication that the number of sources will multiply, as will the data density of these sources.

Scientists involved in the computational sciences require these data sources in order to conduct significant research. They are deluged by the flood of data generated. Using an exclusively numerical format, the human brain cannot interpret gigabytes of data each day, so much information now goes to waste.

Scientists need improved visual interaction

With their data
With each other

Scientists not only want to analyze data that results from super-computations; they also want to interpret what is happening to the data during super-computations. Researchers want to steer calculations in close-to-real-time; they want to be able to change parameters, resolution or representation, and see the effects. They want to drive the scientific discovery process; they want to interact with their data.

The most common mode of visualization today at national supercomputer centers is batch. Batch processing defines a sequential process:

compute
generate images and plots
record on paper, videotape, or film

On the other hand, immediate visual feedback can help researchers gain insight into scientific processes and anomalies, and can help them discover computational errors.

The application of visualization to scientific computing will undoubtedly face a type of cultural inertia well exhibited by the pre-computer history of visual technology. Over the past 100 years, each newly developed visual medium first mimicked the old.

Still cameras were first used to replace landscape and portrait artists.
Movies first captured theater from fifth row center; it took 20 years to discover the vocabulary to move a camera around.
Television neatly adopted the style and content of film; only now are its real-time and interactive capabilities being developed, as in the use of instant replays and graphic overlays.
Visualization, the new interactive visual medium, has great potential for new modes of use beyond its origins in rotating logos for television.

Most people see the end result of visualization — reproduced still color photographs or movies. With the exception of flight simulator trainees and video game players, all those not actually in the process of producing visualization see it as one-way and non-interactive. One cannot publish interactive systems in a journal.

The process of scientific discovery is essentially one of error recovery and consequent insight. The most exciting potential of wide-spread availability of visualization tools is not the entrancing movies produced, but the insight gained and the mistakes understood by spotting visual anomalies while computing. Visualization will put the scientist into the computing loop and change the way science is done.

Scientists need an alternative to numbers. A technical reality today and a cognitive imperative tomorrow is the use of images. The ability of scientists to visualize complex computations and simulations is absolutely essential to insure the integrity of analyses, to provoke insights and to communicate those insights with others.

So far, however, scientists and academics have been largely untouched by this revolution in computing. Secretaries who prepare manuscripts for scientists have better interactive control and visual feedback with their word processors than scientists have over large computing resources which cost several thousand times as much.

Traditionally, scientific problems that required large-scale computing resources needed all the available computational power to perform the analyses or simulations. The ability to visualize results or guide the calculations themselves requires substantially more computing power. Where will this power come from?

Workstations, minicomputers and image computers are significantly more powerful and effective visualization tools than supercomputers. It is a waste of supercomputer cycles to use them to convert model data into new pictures. Specialized graphics processors are more cost-effective than supercomputers for specialized picture processing and/or generation. Workstations should be placed on the desks of each and every researcher to give them immediate access to local graphics capabilities. Every scientist and engineer should have a personal workstation.

Every research center should provide on-site capabilities and facilities for high-end visualization. Visualization equipment and projects should be considered in competition with more memory, more disks, more networking, and soon, to provide a balanced response to user needs.

To encourage the development of visualization tools for scientific and engineering research, interactions must be fostered between scientists, engineers and visualization experts. These interdisciplinary groups should be expected to develop, document, use, and publish both (1) useful results in their discipline, and (2) effective visualization software and documentation. Scientists and engineers need to rely on the experience and intuition of visualization experts to anticipate which representations best convey the information distilled from a cascade of numbers from a supercomputer; the visualization experts need to rely on scientists and engineers to point out the crucial information which flows from the underlying science of a problem.

We encourage the support of interdisciplinary research teams, rather than just facilities, to ensure that long-term visualization developments be grounded in real problems. Close interaction between scientific investigators and visualization technologists will foster better communication between researcher and computer, and between researcher and researcher. The resulting effective and reusable tools can then be shared with scientists and engineers in other research areas, and within the research community at large.

It is expected that all teams engaged in visualization in scientific computing have a mix of skills, and that the development of the tools and techniques of visualization will be an iterative process with different skills contributing at different stages Here is a list of team members and their associated skills.

Visualization and Interdisciplinary Teams

Computational scientists and engineers conduct research in one of the discipline sciences (molecular modeling, medical imaging, brain structure and function, geosciences, space exploration, and so forth) or engineering (computational fluid dynamics, finite element analysis, and so forth). Given visualization tools, these researchers are responsible for making advances in their areas of expertise.
Visualization scientists and engineers are involved with the development of visualization software, hardware, networks, languages, operating systems and databases. These technologists are developing tools and techniques that have broad applicability to advanced scientific and engineering research.
Systems support personnel understand computer networks, operating systems, programming environments on supercomputers, workstation hardware and software, and device drivers. They integrate systems and new technologies, and configure and maintain visualization facilities for the computational scientist.
Artists have a formal education in visual communication, including an understanding of color theory, composition, lighting, view angles and related aesthetics. They propose effective visual representations for analysis and communication.
Cognitive scientists, particularly those with formal training in psychology, have experience in human / computer interfaces and visual perception. At the level of basic research, their theories point the field in new conceptual directions. At the task level, they might work with scientists and technologists to develop better user interfaces; for example, they might select input devices or design window management systems.

Visualization Benefits

Scientific breakthroughs depend on insight. In our collective experience, better visualization of a problem leads to a better understanding of the underlying science, and often to an appreciation of something profoundly new and unexpected.

Better visualization tools would enhance human productivity and improve hardware efficiency. We believe advanced capabilities for visualization may prove to be as critical as the existence of supercomputers themselves for scientists and engineers.

If properly designed and structured, tools and interfaces developed for one discipline science or engineering application would be portable to other projects in other areas.

Long Term Goals

In the 1980's, visualization communication in the United States is hobbled by lack of standards, mired in the intellectual frustration of making interconnections across incompatible media, and held up a the gateways by disparate transmission protocols never designed with visualization in mind. Visual communication cannot be shared among users across a distributed network of incompatible workstations with idiosyncratic user interfaces and no common layering of portable software or hardware.

Scientific communication is changing. Traditionally scientific publication has meant English language print, line drawings and a few static images. Increasingly, science cannot be done in print; in fact, essentially all of the visualization research opportunities described in this report require visualization networking and visualization - compatible electronic media for publication.

Needs for visualization-enhanced scientific communication in a number of areas:

Visual publication
Human / computer interfaces
Televisualization
Visualization software and hardware
Standards
Education

Visual Publication:

Contemporary scientific communications media are predominantly language-oriented. Printed media are coupled weakly, if at all, to the visual world of space-time. By contrast, half the human neocortex is devoted to visual information processing. In other words, current scientific communication leaves out half — the right half — of the brain. An integral part of our visualization task is to facilitate visual communication from scientist to scientist, from engineer to engineer, through the intermediary of visualization-compatible communications media.

While interaction today describes the scientist's potential to direct his computation and synthesize new insights dynamically, interaction has a social meaning as well. "Do you see what I see?" one researcher asks another. In this way, hypotheses can be tested and validated or falsified in minutes instead of years. Changing the scale and pace of visualization alone would affect research profoundly. But we can predict with certainty that such changes in modality will also lead to immense conceptual advances as well.

Scientific research requiring computationally intensive visualization is in danger of becoming Babelized and thereby incommunicable. Much of modern scientific research cannot be expressed in print — DNA sequences, molecular models, medical imaging scans, brain maps, simulated flights through a terrain, simulations of fluid flow, and so on. If poorly communicated, such research cannot stimulate new discoveries and new disciplines.

The end results of selected visualization — photographs, films and videotapes - are what most people see With the exception of flight simulator trainees and video game players, all visualization seen by those not involved in producing it is one-way it is non-interactive. A scientist cannot publish "interaction" in a journal.

Electronic media, such as videotapes, laser disks, optical disks and floppy disks, are now necessary for the publication and dissemination of mathematical models, processing algorithms, computer programs , experimental data and scientific simulations The reviewer and the reader will need to test models, evaluate algorithms and execute programs themselves, interactively, without an author's assistance. Scientific publication needs to be extended to make use of visualization-compatible media.

Reading and writing were only democratized in the past 100 years, and are the accepted communication tools for scientists and engineers today. A new communication tool, visualization, in time will also be democratized and embraced by the great researchers of the future.

The introduction of visualization technology will profoundly transform the way science is communicated and will facilitate the commission of large-scale engineering projects. Visualization and science go hand in hand as partners. No one ever expected Gutenberg to be Shakespeare as well. Perhaps we will not have to wait 150 years this time for the geniuses to catch up to the technology.

Scientists in computing environments need to adopt a new mode of human/computer interaction:

The field of visualization requires a new formulation of the problems of transforming symbolic and numerical descriptions into graphic renderings and simulated behaviors. What if the symbols and simulated behaviors are in the minds of scientists? How can the computer get high-level access to the mental imagery of what the researcher wants to express? New concepts are required that define human capabilities for analogical and metaphorical visualization, transformations of visually representable concepts, and generation of intuitive understandings that lead to conceptual scientific breakthroughs.

Concurrent processing of symbolic information and visual imagery

Many modes of thought require concurrent processing of symbolic information and visual imagery. For example, when we read a textual description, perhaps of a room in a mountain inn, we can often visualize the scene. Textbooks provide both text and supporting illustrations, enabling us to reference either or both for better comprehension.

Knowledge-based systems, however, do not currently have this capability. The concurrent processing of verbal and visual information will be of critical importance if computer-based intelligent systems are to capture any significant portion of the world's knowledge as archived in published literature.

Televisualization

The sheer scale of graphics and image data sets challenges the current bandwidth and interactivity of networks.

The application of networks to visualization is termed televisualization. Televisualization requires a major enhancement over existing network capabilities in the following areas:

Value-added processing at nodes - Transmitting 3D images or volumes over an wide-area network is not practical; transmission costs and times are too high. Instead, we need a scheme whereby we can balance transmission costs with loca l computing costs.

A major problem to be solved as we enter the 1990's is the need for portable, robust, useful visualization software. Currently there is no standard set of visualization tools and graphics interface conventions appropriate for complex tasks, the definition of which is an open research problem.

Over the next 2-5 years, highly parallel architectures will evolve to the point where image computers will be able to approach speeds offering real-time interaction.

We believe that it is too early in the life of the visualization discipline to impose standards, but we can recommend categories of visualization which should be standardized in order for tool makers and tool users to move forward on a national scale.

Most research centers have disregarded thes e standards as being too restrictive or old -fashioned to consider. Consequently, there currently exist an abundance of graphics packages which are only standard where they were written.

High-bandwidth networks will require new, visualization-capable, networking protocols. These still have not been proposed.

Education

Visual training, if any, is left to the commercial television and entertainment industries. Our culture is simply reluctant to accept, much less favor, any level of non-verbal communication in science. Our culture talks of the insight necessary for science, but only favor s that form of seeing which can be reduced to typed symbols an d statistics.

Today's academic and scientific reward systems overwhelmingly favor those who manipulate numbers and symbols. We have used computers to replace much phenomenological observation and rely mainly on symbol processing to gain insight. We trust numbers and symbols; they are easy to compare and publish. Academia has an enormous installed base of numerical data acquisition and analysis machinery, word processing and printing technology - all in support of getting science in print. Publishing and grants, and therefore tenure, have rarely come t o those whose productivity depends upon or produces visualization results.

CS 422	User Interface Design	developing effective user interfaces	Every spring
CS 424	Visualization & Visual Analytics	interactive 2D visualization of different types of data	Every fall
CS 426	Video Game Programming	creating complete audio visual interactive (and fun) experiences	Every spring
CS 488	Computer Graphics I	basics of how computers create images on screens, OpenGL	Every fall
CS 522	Human Computer Interaction	interaction and evaluation of interactive environments	Every 2 years
CS 523	Multi-Media Systems	creation of Educational Worlds	Every 2 years
CS 524	Visualization & Visual Analytics II	3D Visualization	Every 2 years
CS 525	GPU Programming	shaders and parallel processing	Every 2 years
CS 526	Computer Graphics II	current topics in computer graphics	Every 2 years
CS 527	Computer Animation	creating realistic motion	Every 2 years
CS 528	Virtual Reality	immersion	Every 2 years

Communications Media	Number of Years Old
Sight	500,000,000
Speech	500,000
Writing	5,000
Print Broadcasting	500
Visual Broadcasting	50
Visualization	5

Lecture 1

Intoduction to the Course

Information about the Course - Syllabus, presentations, projects etc.

How this class relates to to other similar / related CS courses

Coming Next Time

Visualization Basics