| Back | Map | Glossary | Information |
Take the following stream of numbers (34 KB).
Do they make any sense?
Now see what happens when the numbers are visualized in an image.
Thunder Storm Simulation
JPEG Image (7 KB); Credits and Copyrights
Scientific visualization is the process by which numerical data is represented visually to the researcher,. enabling him or her to make sense of it all.
Until recently, the end-products of scientific visualization have ranged from simple graphs to sophisticated animations. But all of these forms share a drawback: the researcher can't interact with the underlying process by which the data behind the visualization is produced, whether via computer simulations or instruments.
But advances in high-performance computers, high-speed networks, and data storage technologies are making it possible for scientists to use graphical representations as interfaces to examine, navigate and manipulate data dynamically. For example, a cosmologist might change the parameters controlling how galaxies form in simulation, as the calculations evolve in time. Or a radio astronomer could "see" an object while it was being observed, a feat that is still not possible.
CAVE T-Storm Simulation
JPEG Image (17.8 KB); Credits and Copyrights
At NCSA, the display and feedback devices that will make such interactivity possible are collectively termed "virtual environments."
These environments are characterized by any kind of representation that allows the researcher to interact visually, aurally, or even tactually with the process generating the underlying data, and alter that process in realtime through the representation itself. Because we live in a 3-D world, it helps to represent the data likewise in three dimensions plus time, thus allowing the researcher to become literally immersed in his or her work.
But "total immersion" in data poses a new set of computational challenges: transforming data into 3-D images that you can see, hear, even touch; fast connections with high-performance computers; and instantaneous interactions between researchers and their computers, instruments or both.
In short, the technologies driving "virtual environments" are fast becoming integral to effective metacomputing, and vice versa.
Return to the Parts of a Metacomputer