Extract's Table of Contents:
3-D Computing
Three-dimensional graphics and videos display information on a computer screen so that it looks as if it shows objects in the three-dimensional space
 that we take for granted in everyday life but lose when we look at a two-dimensional book, television screen, photograph, or computer monitor. Efforts to add back the missing third dimension without computers have been primitive and cumbersome, such as 3-D movies that required you to put on special cardboard glasses with one red and one green eyepiece.
*One puzzling or irritating feature of many terms in my Glossary may be the number of strange-looking names of computer products and information technology jargon that take the form of single words, with no spaces: CompuServe, IntelliWeb, MetaTools, MacroMedia, NetWare, and the like. There's a reason for this - sort of. Spaces matter to computers, in that software on the Internet, operating systems and telecommunications systems read a block of text looking for a space or special symbol, such as @ or /, that marks the end of a command or file name. They can't handle Macro Media and instead need something like Macro_Media, Macro@Media, Macro.Media, or Macro/Media. Techies are used to connecting words with dots, dashes, and slashes in this way and are cavalier about upper and lower case. You see the vestiges of this practice on the Internet, where addresses on the World Wide Web are full of dots and other connecting symbols that never, ever follow or are followed by a space for instance,
"http://www.choices.client/server.uiuc.edu/research/Vosaic/vosaic2.html" (a real address, for NCSA's Vosaic site. Vosaic actually stands for Visual Mosaic, but no spaces, please).
Applications of 3-D computing are growing rapidly, to the extent that Business Week made the topic its cover story in September 1995, with the somewhat breathless subheading "From medical imaging to virtual reality games, the race is on to capture computing's next dimension." The article is thin on examples, suggesting that 3-D is still in the take-off stage. It includes a military flight simulator, the "biggest ever" simulation, which left an Air Force officer saying, "You got sweat beading up on your forehead . . . you could hear anxiety in your wingman's voice when you got separated." It later states that "like no other field of computing, 3-D hits people in the gut." It cites engineering applications such as use of 3-D by the designers of New Zealand's winning entry for the America's Cup yachting competition, X-ray crystallography, architecture, radiology, and computer games, the single area in which it is expected to have most immediate impact.
It is as easy for a multimedia computer to process and display 3-D as any other type of graphic image, given enough processing power, speed of data transmission, and storage capacity. Those were not at all givens until the mid-1990s. Even today, although generating 3-D still graphics can easily be done on a PC, creating 3-D video graphics requires a powerful and expensive computer workstation, typically a Silicon Graphics machine. Silicon Graphics has become Hollywood's favorite enabler of ultraspecial effects; its sales tripled between 1991and 1994. To some extent, it's a prestige item, equivalent to a Porsche. Half a million dollars can be the entry fee for this top end of special effects. You know you've arrived when you drive a Silicon Graphics.
You can make do with less (and Silicon Graphics sells machines for a few thousand as well as a million dollars), with many competitors such as Hewlett-Packard driving prices down rapidly, but the cost is still high in comparison with that for basic personal computers. The first commercially available plug-in card for PCs that offers full 3-D video graphics capabilities, introduced in late 1995, cost $24,000. (The feature "set" includes "24-bit texture acceleration, scaleable textual memory, bilinear and trilinear MIP-mapping, specular highlights, full-score antialiasing, transparency and depth cue.")
"The world is complex, dynamic, multidimensional; the paper is static, flat. How are we to represent the rich visual world of experience and measurement on mere flatland?" (Edward Tufte, Envisioning Information)
The top end of the top end of the PC hardware range rarely costs more than $12,000, fully loaded with monitor, high-quality printer, and multimedia goodies. Spending $24,000 for a card to plug into a PC may not look like much of a deal. It may well be an excellent deal for such applications as engineering design; simulation of physical operations, including store design, manufacturing facilities, and office layout; and any planning or decision process that
fundamentally depends on effective visualization of movement. Training pilots is an obvious example. Wherever visualization of movement in physical space is natural in decision making, 3-D computing is a productive direction for companies. As one expert comments, "The brain absorbs three-dimensional information like a sponge. We've been doing it since birth."
The real world is three-dimensional. Depth perception is a critical element of how our brains work. It's natural to us. We have been denied it by available media and perhaps as a result we don't expect to get out of Flatland. Given that the $24,000 will cost at most $5,000 five years from now (that figure assumes just a 25 percent-per-year drop in price, routine in the PC hardware field), this may be a technology worth exploring today to locate opportunities that can be realized when the costs make them affordable or to locate ones for which $24,000 is well worth paying for now.
One very successful example of how 3-D can augment existing business processes comes from physical plant operations. Safety and environmental legislation increasingly requires full and frequent surveys of manufacturing plants, hazardous materials storage centers, oil refineries, utilities, and many others. In addition, how a plant changes as it is being built, modified, and upgraded has to be carefully documented and its environmental impacts interpreted. All this documentation fills entire filing cabinets. Much of it will be in computer form, developed and updated by computer-assisted design (CAD) systems. Constructing a physical model of the plant, an essential requirement because two-dimensional diagrams do not provide enough data for effective analysis and decision making, is highly time consuming and labor intensive. A multimedia extension of both CAD and physical modeling is the system marketed by a UK firm, CADCenter, which takes video input from a series of on site cameras and produces a complete 3-D model, wing "photogrammetry. techniques, which calculate the exact location of each image and derive from them the precise position and size of each machine, piece of pipe, opening, and the like that is in the field of vision.
Toy Story has attracted a lot of attention as the first computer-generated, full-length, 3-D cartoon. It marks a significant milestone in the evolution of multimedia. To put it in context, though, it was by no means simple multimedia. Its 114 ,000 frames occupied 600 billion bytes-1,000 hard disk equivalents-and required 800, 000 hours of machine processing time to produce. The producer states that, in assembling the largest number of Ph.D.s to work on any film in history. Disney created history. And the firm saved a lot of money as well as producing a film even the snobbiest critics praise.
Toy Story cost just $25 million to make, compared with the$40 million for
The Lion King. The team that made it numbered 110, versus The
Lion King's 800.
The system was developed by University College, London, and illustrates two of the major emerging, if easily overlooked, trends in multimedia: (1) the use of scientific software development theory and tools to create easy so-use applications that are of value to business, and (2) the increasing readiness, even need, for university researchers to turn those tools into marketed products. The CADCenter product is the result of successful academic research combined with an already-successful commercial company's own software and marketing capabilities to create something many businesses need but do not know about. They do know about or can easily find out about multimedia games, office technology, "edutainment," and CD-ROMs aimed at consumers.
A major opportunity for both research organizations and multimedia services providers and businesses is thus to find out about each other. Scientists look for every opportunity to turn developments in computer technology into useful tools; World Wide Web and Internet browsers were the creation of physicists looking to make the Net more of a community of researchers, for example. They have taken the lead in many areas of 3-D because of the importance of visualization in their work. Well over half the articles and ads in any 1995 issue of Scientific Computing World were about visualization tools that can be used with the same standard graphical user interfaces employed by businesses and individuals. Three years ago, comparable publications developed at most 5 percent of their pages to these topics.
See also: Virtual Reality
Access Time
Access time is the waiting time between a computer instruction's requesting data from a CD-ROM hard disk, or other device and the data's being located. To that must be added the time to transfer the data. Think of yourself looking up the number of a business contact in your Rolodex. The time to locate the name is the access time, and the time to write down the phone number on a piece of paper is the transfer time. Your Rolodex access time is generally far slower than the transfer rate. That is typically the case with computers. For example, the first CD-ROM drives took almost a second to locate data and then transferred it at 6 printed page equivalents per second (150,000 bits).The newer 4x drives (four times the speed of the original) that are found in most standard multimedia PCs have an access time of a quarter of a second and a transfer rate of 24 page equivalents per second. Go on to the next generation of drives, 6x, and the access time is just 0.15 seconds and the transfer rate, 36 printed page equivalents per second. If you take these figures from an ad, it seems reasonable to conclude that, because the 6x drive's transfer rate is six times that of the basic drive, it will transfer, say, 2 printed book equivalents (15 million bits) of data six times faster. The 6x drive should take a total of 16 seconds; the 4x, 25seconds; and the base drive, 100 seconds.
No way. Consider two very different combinations of data access amounting to 15 million bits: (1) 10,000 accesses of short items, each of which is 1,500 bits (a paragraph) in size and (2) 5 accesses of items, each of which is 3 million bits (half a book) in size Here are the time factors in seconds:
The same total volume of data takes between 105 and 10,100 seconds to access and transfer on the base drive 2 minutes versus almost 3 hours, a hundred fold difference. The 4-times-faster 4x drive is only 2.5 times faster when handling frequent accesses to small data items. The 6x drive takes 40 minutes to process all the short items and just 18 seconds to handle the longer ones. So, the overall difference in performance is 18 seconds to 3 hours for exactly the same number of bits.
There are instructive lessons for managers n this oversimplified example (it doesn't include the effects of software processing time and the exact type of operating system, which can further change the relative performance ratios between the 4x and 6X drives). One lesson may simply be not to believe ads. Just as horsepower figures don't tell you much about a car, neither do figures on the raw speed of hardware. More general lessons are that access time can dramatically affect overall performance and that it is the mix of applications, number of accesses involved, and size of the data items they transfer that determine the overall performance of the computer. This poses many challenges to designers of multimedia systems. How to cut access time for its CD-ROM encyclopedia consumed much of the Microsoft development team's personal lives for a year, for instance, as they probed for ways to save fractions of a second.
You can easily see, from the numbers used in the above example, how a system that performs very quickly in handling some requests can be abominably slow in handling others. CD-ROM is poorly suited to short, bursty data traffic, such as 10,000 accesses of 1,500 bits. Hard disks have a shorter access time and outperform CD-ROM in this area By contrast, CD-ROM performs relatively well in accessing the lengthy files characteristic of multimedia That said, it still takes half a minute to access just 5 photographs using a 4x drive. The typical hard disk access time is now around 20 milliseconds, with newer devices offering under 10 milliseconds. Transfer rates are 2 to 3 million bytes per second, half a printed book, around four times faster than even a 6X CD-ROM drive. The advantage of CD-ROM is simply low-cost storage capacity.
The heads that read the data on a disk float above the disk as it rotates. The gap is so tiny that a wisp of smoke would hit it like an avalanche. It's the equivalent of a Boeing 747 flying along at full cruising speed four inches above the ground. That's why disks are so slow. They have to move so far to get the
data.
Access time is mostly idle time to a computer, although there are ways to overlap operations so that the machine does not sit totally spinning its electrons. Access time is generated by the simple reality of physical movement. Hard disks, floppy disks, and CD-ROM disks all rotate, just like a compact disc or its predecessor, the rapidly disappearing long-playing record. There is thus a small delay as either the reading part of the drive has to move to the track where the data are located, or as it waits for the section of that track to spin around to it. On a hard disk, the heads have to move, albeit a tiny distance, to read the data.
In many instances, computer access time does not affect your own waiting time, called "response time," the gap between your inputting a request and getting a response. Acceptable response time is 3 to 5 seconds, depending on the application. When it takes you a second or so to absorb information, you won't even notice 100 milliseconds (thousandths of a second) for a single transaction. The problem is that your request may generate many computer disk reads-dozens, hundreds, or thousands-to locate and move needed data. That's why searching for information on a CD-ROM can be so slow. Each individual access and transfer is faster than your awareness of the elapsed time, but they can add up.
Senator Everett Dirksen, a power player for decades in Congress, once commented about the federal budget,
"A billion dollars here, a billion there and pretty soon you're talking about real money.. The equivalent for access time is "A millisecond here, a millisecond there, and pretty soon you wish you'd never bought this damned computer."
|
