Some readers of our publication must have admired the extreme thoroughness with which we have thus far ignored the Virtual Reality Modeling Language (VRML, pronounced vermel - suggesting a small rodent of some sort, possibly midway between vole and gerbil). As our readers know, RTG covers only the most serious, heavy duty, industrial strength graphics and hence the fact that we have ignored VRML is quite possibly a sign of its being fundamentally too light weight to pique our interest. Another possibility, of course, is that we are behind the curve in understanding VRML and have thus far been in fear that if we said anything that that situation would be self evident. Whatever the case, it is time to acknowledge the subject.

If the Internet is this year's hot stuff and virtual reality was last year's hot stuff, then virtual reality over the Internet is potentially the hot stuff of the decade. In the present context, virtual reality is taken mainly in the small sense of being an interactive simulation viewed on a monitor, not in the large sense of necessarily being immersed in real time with a head-mounted display.

To do a simulation over a network, each participant must have at least three things: a database that describes the shapes and characteristics of objects in the simulated world, a near-real-time datastream from the network that describes where moving objects (including those objects representing other participants) are currently located, and viewing apparatus by which the virtual world is observed. Often, the participant also would need to place current position data onto the network so that other participants could include everyone in the world view. Our prejudice towards graphics is evident, but keep in mind that sounds, touch, and odors might be included. The paradigm of database, datastream, and "viewing" apparatus holds up reasonably well for other sensory stimuli.

The database is a large amount of data that is transmitted only infrequently, maybe just once at the start of the simulation. The near-real-time data-stream is continuous, but relatively low bandwidth, because only the positions and orientations of objects need be sent, and extrapolation is helpful in minimizing the bandwidth required. The viewing apparatus goes a long way toward determining the quality of the experience. A PC with slow graphics might operate with the same database and datastream as a RealityEngine™ driving a head-mounted display. The point is that the generated pixels do not travel over the network, they are produced and consumed locally so there is no challenge to network bandwidth.

VRML is a standard format for the data that must be transferred over the network, both for the database and the datastream. Recently, a standards committee adopted VRML 2.0. The first version of VRML was rather limited; one could basically only view static objects. There was some controversy in the development of the second version of the standard, with Microsoft promoting one approach and a group of companies led by Silicon Graphics promoting another. The differences between the approaches centered on the implementation of "behaviors," the descriptions of how objects behave when interacting with other objects. The Microsoft approach was viewed by many as technologically more sophisticated, but also more difficult to use. The SGI-favored approach was adopted by the standards committee. Microsoft, however, has kept its version alive, saying that it will be appropriate for use in certain types of applications and that it intends to support both versions of the standard.

What will VRML be used for? Fun-and-games comes to mind. The latency (i.e., transmission delays) in the Internet will limit the types of fun-and-games to those in which there is no fast back-and-forth interaction. Chess is fine; ping-pong is out of the question, at least for now. Developing complicated VRML databases will be expensive, and so far Internet providers have not developed good mechanisms for extracting money from players in ways comparable to, say, coin-operated arcade games. Highly motivated great minds are no doubt working on these problems, but little has yet emerged.

Another use for VRML is to show products. This has much potential. High speed interactivity is generally not required, and the pay-off to the producer is in sales of the product. One might ponder automobiles or electronic gizmos via the Internet, opening doors or pushing buttons to ones heart's content. Some industry observers predict that so-called "product-visualization" will soon be the number one application for VRML.

A third potential use for VRML is for interactive education. The potential seems to be there, but the present capabilities of the Internet for multi-media style educational material have been little exploited. Perhaps VRML will prove so exciting that it will spark a flood of educational applications, perhaps not.

VRML is still very much in the realm of enthusiasts. The VRML 2.0 standard has only been out for a few months. There are a dozen or so viewing programs for PC's, and a few authoring packages. The number of VRML sites on the net is closer to a hundred than a thousand. Fast rendering will make the technology more appealing to users, but it will be a year before there is a large installed base of PC's having 3-D graphics accelerators. Until there are well-equipped users, there will be limited interest in developing the sites. In sum, you haven't missed much yet, but the pieces are being put into place.

Interactive simulation for military training has been under development for more than a decade. The Distributed Interactive Simulation (DIS) standards predated VRML, and it would not be much of an exaggeration to say that VRML was developed without reference to DIS. A few people have worked in both worlds, but the mindset seems to be to keep them separate. VRML is further along than DIS in ways related to object behaviors, although there is work being done to advance the DIS capabilities in this area. DIS, unlike VRML, does not include a database standard or mechanisms for on-line database transmission. DIS is edging towards standards for moving models, real time changes in databases, and real time environment databases (for weather and the like). DIS deals mainly with the near-real-time datastream. Separately, the aerospace community uses a few de facto database standards, and in most ways simulation database standards are more sophisticated than VRML. Most interactive simulation demands lower latency than the Internet presently provides.

While VRML and DIS have obvious similarities, the differences now predominate. Expect to see VRML pressing ahead in issues related to object behaviors and real-time changes in databases. Expect DIS to press ahead in issues related to complex environments, and, now especially, ties to artificial intelligence-based simulation and engineering-model simulations. Expect that over time the common ground between VRML and DIS will increase. Expect both to prosper. However, don't expect the two development communities to talk much to each other. They have different agendas.

A Festival of Undefined Abbreviations (FUA)

Is the military fond of acronyms? Are Alaskan bears fond of salmon? Well of course they are; they are masters of the art.

From time to time I am plunged into some project or other that makes me feel like I'm drowning in alphabet soup. I call such a project a festival of undefined abbreviations - let's see, that would be a FUA. I recently saw a chart that had an alphabet soup abbreviation at the top of the chart, and under it two bulletized columns of ponderous FUA, save one bullet that said "fax communications." How does this happen?

FUA is one response to the need for ever-more-specialized terminology in a world with increasing numbers of specialties. The military is certainly not the only profession with such specialties. Computer people, business people, and many branches of government have there own FUA. Military FUA is especially interesting, because sometimes they spell out to things making obvious sense, which you can rarely say about the computer field. Also, notions to the contrary notwithstanding, the military often makes a valiant effort to speak with candor and precision.

Let's warm up with a few simple ones. How about CBS and NBC? Americans know these as television networks, but CBS is also the U.S. Army's Corps Battle Simulation and NBC is nuclear, biological, and chemical - as in NBC threats. A CINC, pronounced sink, is a Commander in Chief. A JEDI is a Joint Exercise Driver for Intelligence.

Moving along, there is OOTW, operations other than war, or, now sometimes, MOOTW, military operations other than war. Disaster relief or drug interdiction or hostage rescue efforts carried out by the military are examples of an OOTW. Years ago there was not much OOTW, but now it is on the rise and it needed a name. By the way, OOTW is pronounced oot-wah, though pronunciations are almost never written down. It is a challenge to discover how many seemingly unpronounceable combinations of letters have actually been assigned pronunciations for use by the cognoscenti.

For alphabet soup abbreviations to serve well as words, they should have a pronunciation. A clever abbreviator will make the abbreviation obviously pronounceable. How about JECEWSI? It seems designed to be pronounced like Jacuzzi, the maker of pumps for spas. Trouble is, JECEWSI is supposed to be an abbreviation for Joint Electronic Warfare Exercise, but the letters in the abbreviation are not in the order of phrase from which it was derived. It is a higher achievement if the letters preserve the order of the original, and a higher achievement still if they are initial letters. Thus radar, radio detection and ranging is a high achievement. Once it is so obviously pronounceable, it becomes a word and is dubbed an acronym.

The pinnacle of achievement is to get the pronounced word to have a meaning which at least hints at the concept it represents. Fans of radar point out that it is spelled the same backwards and forwards, which hints that reflection is involved in the operating principle - a bit of a stretch. In private life, the pronounceable and suggestive abbreviation is achieved with some frequency - Mothers Against Drunk Driving are MADD, and so forth. Military people rarely spend the time needed to achieve this peak. Indeed the achievement might draw undo attention and suspicions of gimmickry. Nonetheless, some might suppose that whoever coined CFDB/MSDS for Conventional Forces Database/Master Simulation Data System might have tried just a little harder.

Sometime abbreviations get perversely out of hand and become just as long or longer than a phrase for which they substitute. The Continental United States is the CONUS, pronounced cone-us. Outside the Continental United States is OCONUS. But some operations encompass both CONUS and OCONUS. This is certainly no more or no less than worldwide, but the phrase CONUS and OCONUS is widely preferred in the military context. Extra precision was paid for but not achieved.

Politicians are fond of decorously elliptical language. Someone in the Pentagon said that when generals get more than two stars they are, of necessity, politicians. However, most of the military is admirably frank, and you will not find them calling a war a peacekeeping action. Most abbreviations are made below the political levels. The combination of frank expression, which is revealing, together with abbreviation, which is concealing, can have a certain charm. Thus there is a JULLS, a Joint Universal Lessons Learned System, and a JMETL, a Joint Mission Essential Task List. In these abbreviations, by the way, "joint" means shared by the Army, Navy, and Air Force. The JDWP is the Joint Doctrine Working Party. A NIEX is a No-notice Interoperability Exercise.

An alternative to using abbreviations is to give things names, like Operation Desert Storm. Naming provides an immediate challenge to the poetic sense that abbreviating evades. I happen to have a list of training exercises that include Tempest Express, Keen Edge, Agile Lion, and Atlantic Resolve. Not bad. But one wonders what Ponderous Ballerina could be about.

copyright CGSD Corp., updated Jan 28, 1997, www.cgsd.com/july_96.html