Building HMDs is not rocket science. Back in the early 90s (when I did this kind of research) we used LCD displays purloined from the new portable TVs that were coming on to the market. Sega used 2 320x240 displays (left and right). The focusing system for the eyes (so you don't need to wear glasses, which you can't with most HMDs) is very similar to what you might find in a pair of binoculars. Add an orientation sensor (yaw pitch roll) and that's really just about it. Oh, and you may want a pair of headphones.
Seriously, this isn't rocket science. We manufactured test HMDs in my garage for a year and a half, using off-the-shelf components. That said, my focal plane can now do things that are downright unnatural - because we used some very odd lensing stages which, well, didn't always work perfectly.;-)
But again, these things really aren't safe for any sort of extended use. 20 minutes, a few times a week, is all that anyone could really hope to tolerate without producing unpleasant side-effects.
All things considered, I think that heads-up displays are far better, because they only add to the real world, rather than substituting something for it. That's where I'd like to see this sort of development go.
Surprisingly, this is not the case. People learn how to adapt themselves quite well to stereo-blindness. There are other cues for depth-perception than stereo (five others, to be precise) and the brain just fills in the gaps.
About 10% of the population is stereo-blind, and most stereo-blind folks don't even know it. It's not really a serious disability - except in cases like this.
Back in the early 1990s I started a company designing consumer VR ware for video games. We did lots of design work, got a few patents (hardware) on our work, then got a big contract from Sega to design key portions of their Virtua VR system - which was an HMD (head-mounted display, the real name for these things) that would plug into the Sega Genesis system to give you full immersion into the game playing world. We worked out all of the technical details, got the prices on the parts down to where it could sell for about $199 retail, and sat back and waited, and waited, and waited. Then Sega killed the project. We never heard from Sega why they killed the project. But, a year later, I heard a lecture from some researchers at SRI, who had done the testing for this helmet on Sega's behalf. Sega wouldn't allow them to release the results, so they did the tests again - on their own dime - and released them. And here's why you won't be wearing an HMD anytime soon:
1) Binocular dysphoria: when you wear a stereo HMD, your eyes/brain are getting one clue for depth perception (parallax) whereas in reality, there are six different cues for depth perception (focus plane, shadowing, etc) which your eye/brain uses to sort out what's going on in the z-plane of reality. While you're in the HMD, the brain adapts to this. Trouble is, when you take the HMD off, your brain is _still_ adapted to this. Now your brain eventually goes back to normal, but this takes some time, and it varies from individual to individual along a bell-curve distribution. Some people come back almost immediately. Others come back very slowly. That is an enormous problem if you go out and get into a car right away, or - as would be the case with the kids using the Virtua VR - getting onto a bike, walking down stairs, etc.
2) Torque: The Virtua VR was, like most early VR HMDs, closer to Darth Vader's helmet than to a pair of eyeglasses. That puts pressure onto the neck, and the neck can't really handle more weight than the head's already putting on it. Adding weight adds a lot of torque to the neck's movements, resulting in much more frequent neck strain.
3) Barfogenics: Although computers are more than fast enough to update images at 30 fps (even the Genesis could do this), the sensors which are used to calculate the yaw/pitch/roll of the head - in order to keep the image aligned with your proprioceptive sense of where your head is - generally don't work nearly as quickly. Most cheaper tracking systems, the kinds you'd find in consumer electronics, have some hysteresis associated with them. And that's bad, because if the image lags the movement by more than 50 msec, almost everyone will end up getting motion sickness. (Technically, this is known as "simulator sickness".) But the sensitivity of people to simulator sickness is also distributed on a bell-curve. Some folks get it very quickly, others don't get it at all.
So there you have it: Sega was told that they'd be selling a device that would cause kids to ruin their depth perception, would give them neck sprains, and would make them puke. Sega didn't even want to think about those kinds of lawsuits...
Today HMDs are lighter, but these fundamental issues remain, and remain unresolved. Yes, you can use optical tracking these days, because comupters are much faster with optical processing, but it's difficult to set up. HMDs are lighter, but they're still bad for your eye/brain. And until those problems get resolved, don't expect to be putting your head into an HMD.
One of the persistent worries about GM croops that have been introduced to the wild is whether or not they can "leak" their GM-ness. There wasn't much worry on this point when GM crops were introduced in North America but there is growing evidence that pollens produced by GM plants can fertilize non-GM crops and thus transfer their GM genes. Thus, whether we like it or not, releasing a plant with GM genes means that we are _always_ taking _some_ risk that the entire species will become "contaminated" by these GM genes. If all mustards were suddenly selenium sinks, we might find that table mustard had become too selenium rich for safe human consumption.
The biggest problem we're likely to have over the next thousand years or so is getting rid of our waste heat, not conserving it in the oceans. If, say, China comes to btu/person equivalence with the United States, the planet will cook like an egg. Then we'll be _praying_ for the oceans to cool down. While I agree that this will convert some kinetic energy in the oceans into radiated IR, it pales in comparison to the overal translation of all forms of energy usage into waste heat.
The energy is still there, in the form of waste heat, after the electricity is used to _do_ something. It may be taken out of the ocean, but it ain't taken out of the Earth.
As somone who was intimately involved in the Sega "Virtua VR" HMD, I know a few things about why there is no VR HMD in wide-scale use.
Reason 1: It's very bad for your brain.
Turns out that using the "false" stereo generated by perspective displays makes your brain interpret the signals coming through your eyes in the real-world incorrectly. This effect occurs in nearly everyone who dons an HMD, but goes back to normal after a period of time. Trouble is, that period of time varies from person to person on a bell curve distribution. Some folks come back to normal nearly immediately, while others take hours or days. This means you *can not* trust your depth perception after using a fully immersive VR system. (Don't believe me? Check out the SRI report on this, published in 1996.)
Reason 2. It's bad for your neck.
Adding even just a little weight to your head increases torque forces on your neck, and can result in neck strains. The old generation of "Darth Vader" HMDs were particularly bad in this respect. Now that they're little more than heavy eyeglasses, this isn't so much a problem, but it does need to be considered - particularly in children, whose necks can support a lot less weight.
Reason 3: Bad VR will make you sick.
As had been noted by another poster, any desynchronization between your body and your brain of greater than 50 msec *will* make you motion sick. And some people are more sensitive to motion sickness than others - this too varies along a bell-curve distribution. Now while the computer can nearly always respond in less than 50 msec., it is difficult and expensive to develop sensors which can give you accurate yaw/pitch/roll readings that quickly.
Given these three conditions, it isn't surprising at all that immersive VR hasn't taken off - it's bad for you!
One of the things that both programmers (TV programmers, that is) and consumers usually fail to "get" about IPTV is that it takes us completely away from the channel model of programming. A channel is a set of programs - just like a DJ's set is a selection of tracks. There's nothing intrinsic about the programming - it exists because TV spectrum is limited, so programmers pick the programs that they feel will get them the highest ratings in the market.
But when you move to IPTV, where you can send a highly individualized, per-program stream to each user's STB, why do you need a channel? Can't the customer just directly select the programs they're interested in - from a very, very, very long list of available programs - and watch those? Why do you need a TV programmer at that point?
Of course, there are all sorts of licensing and copyright issues which need to be observed in that situation (so that everyone involved gets to make some money) but that's just a legal nicety.
The idea of having "channels" as anything besides a convienence when we move to IPTV is ludicrous. I mean, we don't have channels on the web, do we?
The basic idea of the Cell is that the programs are unaware of which Cell they're working upon. The Cell can be closely coupled or over a slow network link; the dispatching software (running on the PMU, I believe) decides where to send each Cell process. This certainly makes tasks such as render farming much easier.
The devil's in the details, but the broad strokes illustrated by what we already know seem very impressive. The Cell revolution won't happen overnight - it might not even happen at all. But if it does, we're going to think about computers a lot differently...
The Cell uses a portable software architecture so that code written for any Cell can execute on any other Cell. No re-coding required. As for the kinds of tasks which can be safely offloaded from the PS3 to your HDTV, iCell, etc., I suppose that's limited by the developer's imagination. Some tasks need to be kept close (rendering), others (AI, etc.) can be moved more freely.
As for theoretical increases, the kinds of things that people treasure on their Macs - AV processing - are the kinds of tasks that are most highly parallelizable. Getting near-to-theoretical performance improvements on those sorts of tasks isn't really difficult if the code is written well. It won't be a continuous 10x improvement, but you'd get it where it matters most...
That said, Apple did change their ISA once already - it was said that it couldn't be done, and they did it fairly effortlessly. Since very little code is written in assembler these days, adapting to a new ISA shouldn't be very difficult from "Carbonizing" an old OS 9 app... In theory. Debugging a single-chip multiprocessor, on the other hand, is going to be hairy...
You can't really consider a 4Ghz PPC a low performance computer. It's over 3 times faster than the one I'm using, and even if its instruction execution is less efficient, I doubt we'd lose 3x.
So, let's consider: what is the CPU on my Macintosh actually doing? How much of that can't be handled by an under-powered processor? Certainly the 1.25 Ghz G4 in my iBook is just about sufficient for any normal computing tasks. (Firefox, Entourage X, Word, iChat AV) The only time I really wish for a faster processor is when I do encoding or those sorts of things, i.e. highly parallelizable tasks. I do think that it would present a less-than-obvious transition from G5 architecture to Cell architecture, but, by the same token, the speedup in graphics performance, encoding - the media tasks that make the Mac such an alluring machine for consumers - would benefit greatly...
Although the article (which is quite clear) indicates that the AltiVec architecture is closer to G4 than G5, won't the speed increase of having 8 fully-parallel processors (9 if you count the main CPU) more than make up for the issues associated with the loss of the G5's advanced features? It seems to me that this is a natural for Apple - it will give them a 5x - 10x performance boost over anything that's on the drawing boards over at Intel.
Even so, I doubt we'd see Cell-based Macs until at least 2007 - but wouldn't it be great to run PS3 games on your Mac? (As if that'll ever happen.) But then again, given the Cell architecture, your PS3 could use your Mac to make its games run faster! A whole new reason to have an XServe-based supercomputer...
Jupiter emits many times more infrared radiation than it absorbs. So, by your definition, Jupiter would be a star - and it clearly isn't, 2010 aside...
The sad part is that most of the folks using Kazaa simply didn't know any better. Yes, there are lots of better alternatives to bloated sacks of spyware - but many of them are still to hard for most people to use. Or people simply haven't heard about them. (Some of the world still does not read/. -- I know this is obscene, but we've got to play the hand we're given.)
There ought to be a campaign "Geeks don't let friends use spyware" or something. Heh. Make a vow today to give your friends safe P2P software - it's the gift that keeps on giving!
Ten years on, VRML is still being used as a common data format for 3D graphics. Try to move data between Vectorworks (a common design program) and Maya, for example - you'll need VRML in the pathway. And beyond that, VRML handles net-based resources like no other 3D tool. Renderware, which is used in games like GTA:Vice City will push bazillions of polygons around for you, but has no capabilities for sound or motion graphics. And Renderware's nearly good as it gets for commercial real-time 3D libraries.
VRML's been an international standard since 1997, it's part of the MPEG-4 spec (which would have gone somewhere, but for those...ahem...minor licensing issues) and it's been out there being *improved* for the last ten years while every other file format has come and gone, gone, gone.
Could you imagine putting models of the earth online with anything other than VRML? At least with VRML people have some chance of viewing it.
So let them try to float another "universal" 3D format. BRING IT ON! Because, in another year's time, when someone is talking about another universal 3D format, VRML will still be chugging along.
Disclaimer: I'm biased, because, working with Tony Parisi, I invented VRML.
A friend who grew up in Orange County went to see the film on opening night, just for giggles. He reported that rather than being filmed in Orange County (which he expected) it was quite obviously filmed in the San Fernando valley, about 20 miles to the north (and where the film studios are located). So they couldn't even get the location right.
True, only Orange Countians (and Valley girls and boys) will notice this, but would it have cost so much more to shoot it on location?
According to a lot of folks in the early days of Virtual Reality, the Vernor Vinge novella "True Names" had a more profound effect on their thinking than the later (and more renowned) "Neuromancer". "True Names and the Cyberspace Frontier" is being released in December by TOR Books, with a set of essays from scientists & thinkers about how the novella influenced their own thinking about virtual reality, cyberspace, and whatnot...
Although it may seem that VRML-97 has faded away, it is still very widely used in architecture, in science, and education. Because it is a general-purpose format, and not optimized for the very simple worlds of most games - such as Quake - programmers like John Carmack dismissed it immediately upon its initial release, a body slam from which VRML has never recovered, in many eyes. But that was 5 years ago. These days, VRML displays rich environments at 30+ fps with appropriate hardware.
Although pundits were quick to announce the death of VRML, it will have a longer lifespan than any of its competitors - including Java3D. VRML-97, with a few extensions, is the 3D compositing layer of MPEG-4, which is slowly making its way into the consumer marketplace. You can expect that, by the end of this decade, VRML will be the most widespread and widely used 3D technology in consumer electronics - a market of more importance and economic power than the Web.
That's not a bad thing. Standards should not be arbitrarily pulled apart - particularly by competing commercial organizations (reference my XML article on FEED from a few years ago for points on this matter). The VRML97 ISO spec is "owned" by the Web3D consortium, in fact to make spec changes basically "illegal". Whatever that means.
That's a good question. I don't know why Nature adopted GEML - it may be a case of a "tipping point" where enough geneticists and genomics firms finally realized that there was a need for a standard and some "cheerleader" got out there and started waving the GEML flag. If anyone knows the whys of this, I'd be interested to hear...
Slashdot Mirror
Building HMDs is not rocket science. Back in the early 90s (when I did this kind of research) we used LCD displays purloined from the new portable TVs that were coming on to the market. Sega used 2 320x240 displays (left and right). The focusing system for the eyes (so you don't need to wear glasses, which you can't with most HMDs) is very similar to what you might find in a pair of binoculars. Add an orientation sensor (yaw pitch roll) and that's really just about it. Oh, and you may want a pair of headphones.
;-)
Seriously, this isn't rocket science. We manufactured test HMDs in my garage for a year and a half, using off-the-shelf components. That said, my focal plane can now do things that are downright unnatural - because we used some very odd lensing stages which, well, didn't always work perfectly.
But again, these things really aren't safe for any sort of extended use. 20 minutes, a few times a week, is all that anyone could really hope to tolerate without producing unpleasant side-effects.
All things considered, I think that heads-up displays are far better, because they only add to the real world, rather than substituting something for it. That's where I'd like to see this sort of development go.
Surprisingly, this is not the case. People learn how to adapt themselves quite well to stereo-blindness. There are other cues for depth-perception than stereo (five others, to be precise) and the brain just fills in the gaps.
About 10% of the population is stereo-blind, and most stereo-blind folks don't even know it. It's not really a serious disability - except in cases like this.
Back in the early 1990s I started a company designing consumer VR ware for video games. We did lots of design work, got a few patents (hardware) on our work, then got a big contract from Sega to design key portions of their Virtua VR system - which was an HMD (head-mounted display, the real name for these things) that would plug into the Sega Genesis system to give you full immersion into the game playing world. We worked out all of the technical details, got the prices on the parts down to where it could sell for about $199 retail, and sat back and waited, and waited, and waited. Then Sega killed the project. We never heard from Sega why they killed the project. But, a year later, I heard a lecture from some researchers at SRI, who had done the testing for this helmet on Sega's behalf. Sega wouldn't allow them to release the results, so they did the tests again - on their own dime - and released them. And here's why you won't be wearing an HMD anytime soon:
1) Binocular dysphoria: when you wear a stereo HMD, your eyes/brain are getting one clue for depth perception (parallax) whereas in reality, there are six different cues for depth perception (focus plane, shadowing, etc) which your eye/brain uses to sort out what's going on in the z-plane of reality. While you're in the HMD, the brain adapts to this. Trouble is, when you take the HMD off, your brain is _still_ adapted to this. Now your brain eventually goes back to normal, but this takes some time, and it varies from individual to individual along a bell-curve distribution. Some people come back almost immediately. Others come back very slowly. That is an enormous problem if you go out and get into a car right away, or - as would be the case with the kids using the Virtua VR - getting onto a bike, walking down stairs, etc.
2) Torque: The Virtua VR was, like most early VR HMDs, closer to Darth Vader's helmet than to a pair of eyeglasses. That puts pressure onto the neck, and the neck can't really handle more weight than the head's already putting on it. Adding weight adds a lot of torque to the neck's movements, resulting in much more frequent neck strain.
3) Barfogenics: Although computers are more than fast enough to update images at 30 fps (even the Genesis could do this), the sensors which are used to calculate the yaw/pitch/roll of the head - in order to keep the image aligned with your proprioceptive sense of where your head is - generally don't work nearly as quickly. Most cheaper tracking systems, the kinds you'd find in consumer electronics, have some hysteresis associated with them. And that's bad, because if the image lags the movement by more than 50 msec, almost everyone will end up getting motion sickness. (Technically, this is known as "simulator sickness".) But the sensitivity of people to simulator sickness is also distributed on a bell-curve. Some folks get it very quickly, others don't get it at all.
So there you have it: Sega was told that they'd be selling a device that would cause kids to ruin their depth perception, would give them neck sprains, and would make them puke. Sega didn't even want to think about those kinds of lawsuits...
Today HMDs are lighter, but these fundamental issues remain, and remain unresolved. Yes, you can use optical tracking these days, because comupters are much faster with optical processing, but it's difficult to set up. HMDs are lighter, but they're still bad for your eye/brain. And until those problems get resolved, don't expect to be putting your head into an HMD.
One of the persistent worries about GM croops that have been introduced to the wild is whether or not they can "leak" their GM-ness. There wasn't much worry on this point when GM crops were introduced in North America but there is growing evidence that pollens produced by GM plants can fertilize non-GM crops and thus transfer their GM genes. Thus, whether we like it or not, releasing a plant with GM genes means that we are _always_ taking _some_ risk that the entire species will become "contaminated" by these GM genes. If all mustards were suddenly selenium sinks, we might find that table mustard had become too selenium rich for safe human consumption.
Then what would we put on our pretzels?
The biggest problem we're likely to have over the next thousand years or so is getting rid of our waste heat, not conserving it in the oceans. If, say, China comes to btu/person equivalence with the United States, the planet will cook like an egg. Then we'll be _praying_ for the oceans to cool down. While I agree that this will convert some kinetic energy in the oceans into radiated IR, it pales in comparison to the overal translation of all forms of energy usage into waste heat.
The energy is still there, in the form of waste heat, after the electricity is used to _do_ something. It may be taken out of the ocean, but it ain't taken out of the Earth.
As somone who was intimately involved in the Sega "Virtua VR" HMD, I know a few things about why there is no VR HMD in wide-scale use.
Reason 1: It's very bad for your brain.
Turns out that using the "false" stereo generated by perspective displays makes your brain interpret the signals coming through your eyes in the real-world incorrectly. This effect occurs in nearly everyone who dons an HMD, but goes back to normal after a period of time. Trouble is, that period of time varies from person to person on a bell curve distribution. Some folks come back to normal nearly immediately, while others take hours or days. This means you *can not* trust your depth perception after using a fully immersive VR system. (Don't believe me? Check out the SRI report on this, published in 1996.)
Reason 2. It's bad for your neck.
Adding even just a little weight to your head increases torque forces on your neck, and can result in neck strains. The old generation of "Darth Vader" HMDs were particularly bad in this respect. Now that they're little more than heavy eyeglasses, this isn't so much a problem, but it does need to be considered - particularly in children, whose necks can support a lot less weight.
Reason 3: Bad VR will make you sick.
As had been noted by another poster, any desynchronization between your body and your brain of greater than 50 msec *will* make you motion sick. And some people are more sensitive to motion sickness than others - this too varies along a bell-curve distribution. Now while the computer can nearly always respond in less than 50 msec., it is difficult and expensive to develop sensors which can give you accurate yaw/pitch/roll readings that quickly.
Given these three conditions, it isn't surprising at all that immersive VR hasn't taken off - it's bad for you!
One of the things that both programmers (TV programmers, that is) and consumers usually fail to "get" about IPTV is that it takes us completely away from the channel model of programming. A channel is a set of programs - just like a DJ's set is a selection of tracks. There's nothing intrinsic about the programming - it exists because TV spectrum is limited, so programmers pick the programs that they feel will get them the highest ratings in the market.
But when you move to IPTV, where you can send a highly individualized, per-program stream to each user's STB, why do you need a channel? Can't the customer just directly select the programs they're interested in - from a very, very, very long list of available programs - and watch those? Why do you need a TV programmer at that point?
Of course, there are all sorts of licensing and copyright issues which need to be observed in that situation (so that everyone involved gets to make some money) but that's just a legal nicety.
The idea of having "channels" as anything besides a convienence when we move to IPTV is ludicrous. I mean, we don't have channels on the web, do we?
Reality is that which will kill you if ignored long enough.
That which is not manditory is forbidden.
That which is not forbidden is manditory.
At that point.
The basic idea of the Cell is that the programs are unaware of which Cell they're working upon. The Cell can be closely coupled or over a slow network link; the dispatching software (running on the PMU, I believe) decides where to send each Cell process. This certainly makes tasks such as render farming much easier.
The devil's in the details, but the broad strokes illustrated by what we already know seem very impressive. The Cell revolution won't happen overnight - it might not even happen at all. But if it does, we're going to think about computers a lot differently...
The Cell uses a portable software architecture so that code written for any Cell can execute on any other Cell. No re-coding required. As for the kinds of tasks which can be safely offloaded from the PS3 to your HDTV, iCell, etc., I suppose that's limited by the developer's imagination. Some tasks need to be kept close (rendering), others (AI, etc.) can be moved more freely.
As for theoretical increases, the kinds of things that people treasure on their Macs - AV processing - are the kinds of tasks that are most highly parallelizable. Getting near-to-theoretical performance improvements on those sorts of tasks isn't really difficult if the code is written well. It won't be a continuous 10x improvement, but you'd get it where it matters most...
That said, Apple did change their ISA once already - it was said that it couldn't be done, and they did it fairly effortlessly. Since very little code is written in assembler these days, adapting to a new ISA shouldn't be very difficult from "Carbonizing" an old OS 9 app... In theory. Debugging a single-chip multiprocessor, on the other hand, is going to be hairy...
You can't really consider a 4Ghz PPC a low performance computer. It's over 3 times faster than the one I'm using, and even if its instruction execution is less efficient, I doubt we'd lose 3x.
So, let's consider: what is the CPU on my Macintosh actually doing? How much of that can't be handled by an under-powered processor? Certainly the 1.25 Ghz G4 in my iBook is just about sufficient for any normal computing tasks. (Firefox, Entourage X, Word, iChat AV) The only time I really wish for a faster processor is when I do encoding or those sorts of things, i.e. highly parallelizable tasks. I do think that it would present a less-than-obvious transition from G5 architecture to Cell architecture, but, by the same token, the speedup in graphics performance, encoding - the media tasks that make the Mac such an alluring machine for consumers - would benefit greatly...
Although the article (which is quite clear) indicates that the AltiVec architecture is closer to G4 than G5, won't the speed increase of having 8 fully-parallel processors (9 if you count the main CPU) more than make up for the issues associated with the loss of the G5's advanced features? It seems to me that this is a natural for Apple - it will give them a 5x - 10x performance boost over anything that's on the drawing boards over at Intel.
Even so, I doubt we'd see Cell-based Macs until at least 2007 - but wouldn't it be great to run PS3 games on your Mac? (As if that'll ever happen.) But then again, given the Cell architecture, your PS3 could use your Mac to make its games run faster! A whole new reason to have an XServe-based supercomputer...
Jupiter emits many times more infrared radiation than it absorbs. So, by your definition, Jupiter would be a star - and it clearly isn't, 2010 aside...
There ought to be a campaign "Geeks don't let friends use spyware" or something. Heh. Make a vow today to give your friends safe P2P software - it's the gift that keeps on giving!
VRML's been an international standard since 1997, it's part of the MPEG-4 spec (which would have gone somewhere, but for those...ahem...minor licensing issues) and it's been out there being *improved* for the last ten years while every other file format has come and gone, gone, gone.
Could you imagine putting models of the earth online with anything other than VRML? At least with VRML people have some chance of viewing it.
So let them try to float another "universal" 3D format. BRING IT ON! Because, in another year's time, when someone is talking about another universal 3D format, VRML will still be chugging along.
Disclaimer: I'm biased, because, working with Tony Parisi, I invented VRML.
True, only Orange Countians (and Valley girls and boys) will notice this, but would it have cost so much more to shoot it on location?
According to a lot of folks in the early days of Virtual Reality, the Vernor Vinge novella "True Names" had a more profound effect on their thinking than the later (and more renowned) "Neuromancer". "True Names and the Cyberspace Frontier" is being released in December by TOR Books, with a set of essays from scientists & thinkers about how the novella influenced their own thinking about virtual reality, cyberspace, and whatnot...
Although pundits were quick to announce the death of VRML, it will have a longer lifespan than any of its competitors - including Java3D. VRML-97, with a few extensions, is the 3D compositing layer of MPEG-4, which is slowly making its way into the consumer marketplace. You can expect that, by the end of this decade, VRML will be the most widespread and widely used 3D technology in consumer electronics - a market of more importance and economic power than the Web.
Mea culpa.
That's not a bad thing. Standards should not be arbitrarily pulled apart - particularly by competing commercial organizations (reference my XML article on FEED from a few years ago for points on this matter). The VRML97 ISO spec is "owned" by the Web3D consortium, in fact to make spec changes basically "illegal". Whatever that means.
That's a good question. I don't know why Nature adopted GEML - it may be a case of a "tipping point" where enough geneticists and genomics firms finally realized that there was a need for a standard and some "cheerleader" got out there and started waving the GEML flag. If anyone knows the whys of this, I'd be interested to hear...