Since the "payload" of an X-Prize vehicle is three x 200 lb people, needing 400 pounds of landing propellant turns our 850 gallon tank vehicle from a three person vehicle into a one person vehicle.
In the most negative light, you could say that powered landing (with a low performance propellant like we use) takes away two thirds of our payload capacity, but that is a poor metric to base decisions off of, because operational issues have historically been orders of magnitude more important to cost effectiveness than propellant consumption.
We can get the 400 pounds back by either going to a carbon fiber tank instead of a fiberglass tank (cost: $40k up front design fee, then $25k per tank, compared to $9k for the fiberglass tank), or by upsizing everything to a 1600 gallon fiberglass tank (cost: $17k for the tank plus more for bigger engine plumbing, catalysts, and nozzles).
Upsizing the tank is lower risk, because it only uses suppliers that we already buy from, while the carbon fiber tank job would be custom from ATK, and I have already had two other vendors back out on me for big tank work. We already have a 1600 gallon fiberglass tank on hand.
Our mixed monoprop has a measured sea level Isp of 145, with normal increases with altitude. Our big vehicles have a mass ratio of about five, takes off with somewhat under one positive G of acceleration, and has a somewhat regressive thrust profile from partial blowdown pressurization. That combination is sufficient for suborbital flights. A 200+ Isp biprop can do it with a mass ratio of three, but the vehicle gets a lot more complicated to build and operate.
> I figure you need about four times as much fuel at liftoff for a vertical rocket-borne landing as you would for a parachute- or wing-borne landing.
No, not even remotely close.
You only need enough propellant to kill the terminal velocity of the vehicle to land it safely. A vehicle that is stable reentering base first has a Cd right around 1.0, and any high performance rocket vehicle is going to be coming in pretty light after it has burned most of the propellant. The V2 impacted the ground still supersonic because it was aerodynamically stable nose first, so it maintained its 0.15 or so Cd on descent. A reasonably stubby base first reentry will have a terminal velocity of only 200 mph or so.
Killing that speed with a comfortable safety margin takes about 400 pounds of propellant in our vehicle, compared to 8000 pounds of propellant burned on ascent. A higher performance rocket engine could do it with propotionately smaller amounts. A parachute / drogue / ejection system for this weight vehicle is indeed lighter, coming in at about 100 pounds, but that brings a number of disadvantages with it, like coming down tens of miles away and still needing final impact attenuation.
We wanted to use parachtues as a quick hack for the X-Prize, but the test range where we were planning to fly was going to require a half million dollars of "engineering support" and wanted us to carry a thrust termination system (bomb) on the vehicle to satisfy themselves that it won't drift out of the range.
Long term, there is no question in our minds that powered landing is the way to go. We just were given a pretty strong incentive to go there earlier than we were planning.
We don't expect to win the X-Prize, both because Burt probably has it in the bag, and we are behind schedule. We still plan on continuing our development, because our designs are nearly an order of magnitude cheaper to fabricate and operate than Space Ship One, and orders of magnitude matter. If SS1 crashes on Monday, we will throw more time and resources at an attempt, because there really is no other contender, but it will be a long shot.
We could have flown an unguided rocket to very high altitudes a long time ago, but we have instead concentrated on control systems, which is where the important work needs to be done. A team that was busy flying rockets to hundreds of thousands of feet altitude, then decided to add a guidance and control system to their rockets would be in for many rude surprises at high energy levels.
This isn't immediately obvious, but an X-Prize class vehicle pretty much requires an active control system (a trained pilot with appropriate controls is also an active control system). A short burn time rocket, like the recent CSXT 100 km shot, can live with just aerodynamic stabilization (note that it also landed 20 miles away), but the G forces are far too high for people. As the burn time lengthens with lower acceleration forces, the vehicle will gravity turn away from vertical, making it almost impossible to keep a 60 second burn time even accelerating upwards.
People that harp on about propellant specific impulse in the context of suborbital rockets are like programmers that obsessively optimize a function that isn't a hot spot. The goal of a rocket ship is not to deliver specific impulse, it is to move a payload reliably and cost effectively. Isp can always be traded away for mass fraction, and quite often you can improve operability or reliability by doing so. With our new vehicle designs using a single engine and jet vanes instead of four differentially throttled engines we are more likely to consider trading some engine and system complexity for performance, but issues like the requirement for deep throttling still make it a complex decision.
I do Armadillo work on Tuesdays, weekends, and late at night. At Id lately I have been working on next-generation rendering technology while the rest of the company manages the Doom beta process.
I don't issue press releases. I just publicly write about what I am working on, and other people find it noteworthy enough to talk about. All of our development work, including the dead ends and mistakes, is fairly well documented on the Armadillo Aerospace website.
For those of you that are underwhelmed by the 310 pound vehicle, do note that the big vehicle (1500 lbs) that can actually carry people is also flying. Look back in the Armadillo updates around April 19 for testing video. We have since reworked the propulsion system to follow what has worked so well on the subscale vehicle, and should be testing it this weekend. If it works well, we will be repeating the boosted hop with the big vehicle next week.
The flight time is currently limited by federal law to 15 seconds of rocket burn time. We have a waiver coming to extend that to 120 seconds, but beyond that we will need a full launch license.
The significance of all this is that the vehicles are intended to fly up, come back down and land right where they took off from, all without ablating, expending, or seperating anything. It should be possible to have turn around times under one hour even for quite large vehicles.
Someone that has done "some good things (NeXT, the first iMac, OS X)" in their career gets my respect.
Most of the negative tales about Jobs probably have some grounding in truth -- it was almost amusing watching him berate the stage people before a show for glitches in the prop moving systems: "What the hell is this??? Did you guys pick up these parts at Home Depot???". However, he did always listen when I was talking about a technical issue, even when I was saying something that didn't sit with his current understanding of graphics cards / APIs / gaming.
When I was considering setting up to demo Doom 3 at macworld, all of the Apple people were going on about how we needed to sanitize it because "Steve won't let there be any blood or killing". I finally went to him directly, and he replied "If you think you can make it great, then let's do it. I trust you, so you'll have to decide." Not quite the overbearing micromanager he is sometimes portrayed as.
I'm not a regular mac user, but I'm glad Steve Jobs is still around.
John Carmack
We aren't being held up by regulatory issues.
on
X-Prize Progress Update
·
· Score: 5, Informative
We have a good working relationship with AST, the division of the FAA that handles launch license, and we are one of only three companies (along with Scaled and XCOR) currently in the RLV launch license process. We have found all the people there helpful and eager to work with us. There is a lot of paperwork to be done, but we are working through it, and do not see a problem satisfying them. Things like calculating and minimizing expected third party casualty rates are obviously necessary and sensible.
The environmental aspects are less rational, with no analytical sense of scale.
Still, I'm only mildly concerned about the regulatory side of things. I think it will work out. None of our work is held up by any of this, so the worst case is that we have a vehicle built and tested repeatedly at the 200,000 lb-sec waivered impulse limit, with no launch license to allow us to fill the tank the rest of the way up. If that happens, THEN we get peeved about the situation, but continue flight testing with what we can.
Let me repeat: In no way have we been hampered by regulatory burden. Yet. We have been VERY hampered by commercial companies being too worried about liability exposure to work with us - peroxide companies, filament winders, and parachute companies have all caused us significant problems.
The supply issue with 90% peroxide basically cost us almost the entire year of flight testing. We spent the last six months developing a propellant combination that could conveniently replace the 90% peroxide based on widely available chemicals instead of the ultra-specialized propulsion grade. We are in the final optimizing and scale up phase of that. Instead of being irate about it, I try to look on the bright side - it is lots cheaper, easer to handle, and even a bit higher performance.
There are lots of problems still to be worked, but everything is coming along fine. We are behind schedule and somewhat over budget, but no worse off than any other project I have ever worked on...
High concentration hydrogen peroxide all by itself makes a low performance, but very convenient, rocket propellant. All hydrogen peroxide is in solution with some amount of water, because even if you had 100% peroxide, some of it would start decomposing to water (and oxygen) as you stored it.
Drugstore peroxide is 3% concentration. If you pour it on a catalyst, like silver or platinum, you will see bubbles forming in the solution (released oxygen), and the liquid will get somewhat warmer due to the released energy. Above roughly 70% concentration, the heat released is enough to vaporize all the water content, so if you pass it through a good catalyst, you will get all gas coming out the other side, and gas can be accelerated through a rocket nozzle to produce thrust. At 70%, the gas is only just above the boiling point of water, but as the concentration goes up, the temperature goes up fast. 90% peroxide, the most common grade used for propulsion, produces gas at about 1400 F temperature. Going all the way to 98% peroxide, the highest concentration produced, gives a few hundred degrees more temperature, but at a significant price increase. Higher temperature lets you use less propellant for a given amount of thrust-time, because it maintains a given chamber pressure with a less dense, but hotter, mixture (a simplification).
"Real" rocket propellants have temperature several thousand degrees higher, which does indeed increase performance, but the engines have to be cooled, and you need to manage both a fuel and an oxidizer in some form. One of our fundamental system trades is that it is better for an X-Prize class vehicle to use a propellant that simplifies vehicle engineering, even if you have to use more of it.
We use 90% peroxide from a small specialty supplier for all of our flight vehicles, but they closed shop a while ago, and we haven't been able to come to terms with the only domestic supplier of 90% peroxide, FMC chemical corp. Because of this, we have been working on alternate propellant schemes for a good part of this year, in parallel with building the full size X-Prize vehicle. If we had been able to just buy 90% peroxide like we buy all of our other industrial chemicals, we never would have bothered with the research.
Just about every week, someone asks why we don't concentrate it ourselves. True, dozens of people have made a few gallons of high concentration peroxide at various times, but there have only been two large scale concentrators operated in the US outside of the official manufacturers - Rotary Rocket had a concentrator, but it only went to 85% concentration, and it didn't do purification, and Beal Aerospace had a large scale concentrator operational after the blew up their first one. Sure, we could figure out how to do it, but then we would be in the chemical plant business instead of the rocket business, and that's not what we want to do. I am funding an operator in Houston to produce a few thousand pounds of 90% for us, but he is six months behind schedule on delivery, which proves my point about it not being as simple as people think.
The direction we have been pursuing is using a combination of 50% peroxide, which is readily available through distributors from multiple manufacturers, and a small amount of miscible fuel (methanol in our current work). 50% peroxide by itself doesn't work as a rocket propellant, because you can't boil all the water, which makes even decomposing most of the peroxide difficult. Adding a fuel and (the tricky part!) getting it to burn with the released oxygen gives you the energy necessary to vaporize the water and get everything up to a high temperature. Mixing fuels with high concentration oxidizers usually makes a touchy and deadly explosive (we have intentionally detonated a mix of 90% peroxide and alcohol - Very Scary), but buffered with 50% water, and running off of stoichemetric mixture ratio, the risk is not very high. We have a study report from the Department of Mines in the late 50's investigating th
Just building the vehicle costs less than $100k, most of the money is in building multiple iterations of everything as you figure out exactly how you actually need to spend the money:
$ 6k 850 gallon fiberglass tank $ 2k High pressure carbon fiber pressurant tank and regulator $ 1k Honeycomb composite panels $ 5k Aluminum fabrication for cabin $15k Redundant parachutes, drogues, drogue cannons, releases $13k Fiber optic gyro based IMU $ 8k Unrestricted (supersonic / high altitude) GPS $ 2k PC104 systems $ 5k video, audio, and data communications $20k Engine machining, catalysts, laser cut plates $ 5k Plumbing, valves, etc $ 5k Fastblock external insulation
For powered landings instead of parachute landings, delete the parachutes and add:
You could trivially spend an order of magnitude more by just using "space certified" versions of everything, but the important point is that standard industrial versions of many things are perfectly adequate. In many cases, todays standard industrial practice is far ahead of the best that could be done at any price in the early sixties.
This is all with free labor for assembly and testing, but that is still only a couple hundred man hours for a full vehicle. We are expecting to destroy the first vehicle in some (unmanned) testing mishap along the way, and build another one mostly from scratch. That will take less than two months, depending on lead times for some items.
You probably mean "Burt Rutan", the aircraft designer at Scaled. Dick Rutan is his brother, who piloted the voyager, and was the test pilot for XCOR's EZ-Rocket, but doesn't have anything to do with Space Ship One, the X-Prize vehicle.
I have always maintained that Burt is the odds-on favorite to win the X-Prize, but it isn't over yet. His design requires a pilot on board for all tests, so there is a non-negligable chance that there could be a fatality, which would almost certainly end the effort in the X-Prize timeframe.
>Inconel, the best commonly used alloy has a single use temperature limit of about 1030K, or 757C. It melts at ~1400C
The refractory metals are better, but less commonly used. Columbium/niobium is reasonable to form. Molybdenum and alloys like TZM take a bit more heat, but have a potentially annoing ductile to brittle transition point for systems that will cold soak. The state of the art is irridium coated rhenium, which doesn't melt until 2466 C / 4471 F.
We fabricated a TZM chamber a while ago at fairly high expense, but still burned through it after an extended length run:
This experience has convinced me that active cooling methods, like transpiration cooling, are probably a good idea for high reusability reentry vehicles.
>Based on the first two responses to this post, you'd think people had never heard of inertial >navigation. With MEMS accelerometers it ought to be pretty light, too.
Pure 3 axis inertial navigation with a strapdown inertial measuring requires extreme precision. MEMS inertial units aren't even in the right ballpark. Mechanical stable platform inertial systems that actually rotated inside the vehicles didn't require awesomely accurate sensors, but they are big, heavy, and not as reliable.
It is a useful programming exercise to write a simulation of a strapdown inertial system and play with bias, noise, and nonlinearity errors (add cross axis coupling and acceleration effects for micromachined gyros for bonus points). Pick reasonable ranges and quantize to 12 bits, then integrate at 100 hz or so. You can start the simulation motionless, but in a minute it will be cruising along at 60 mph in some random direction, hundreds of feet from the start position. An hour later, it will be heading for Mars.
The low end inertial systems that have been moderately soccessful are done by removing gravity from the equation and just doing 2D navigation, and often using other sensors, like magnetometers instead of rate gyros for heading, or odometer readings instead of double integrating accelerometers. Double integration of interrelated noisy sensors with an implicit 1G acceleration is really more demanding than it would initially seem.
The only reason you wouldn't want to use GPS in an ocean crossing is if you are afraid a Bad Guy might be jamming the signals.
Arguably the most professional and widely viewed machinima so far is the music video for Zero 7's "In the Waiting Line", produced by my wife's company, Fountainhead Entertainment. This was a real, commercial production using machinima tools.
It was neat to see the Q3 engine playing on MTV, but it made me greatly regret the quantized normals in Q3 models, which resulted in a noticeable popping on the environment maps. This was largely my motivation for adding per-pixel environment map calculation to the new Doom engine (under the ARB2 path, at least).
>IAARS. (I Am A Rocket Scientist.) >... >The Shuttle uses LOX and LH2, both of which are f'nasty to deal with and are economical only to >generate the immense thrust necessary to achieve orbit. While in orbit, the Orbiter maneuvers >using (relatively) small hydrazine thrusters. N2H4 is also f'nasty, but somewhat less so than >either LOX or LH2.
???
The OMS uses hydrazine / nitrogen tetroxide, which is way, WAY more nasty than LOX / LH2.
LOX / LH2 are cryogens, and contact with them will give you frostbite. Hydrazine is carcinogenic and toxic, but nitrogen tetroxide is roughly as poisonous as the best war gasses from WWI. Plus, it has very low surface tension, so when it spills, it spreads extremely rapidly, which causes it to vaporize even faster than the already high vapor pressure would indicate. The various oxides of nitrogen are famous for the "BFRC" ( big red cloud ) that results from spills, which you should run away from very fast.
Rewriting shaders behind an application's back in a way that changes the output under non-controlled circumstances is absolutely, positively wrong and indefensible.
Rewriting a shader so that it does exactly the same thing, but in a more efficient way, is generally acceptable compiler optimization, but there is a range of defensibility from completely generic instruction scheduling that helps almost everyone, to exact shader comparisons that only help one specific application. Full shader comparisons are morally grungy, but not deeply evil.
The significant issue that clouds current ATI / Nvidia comparisons is fragment shader precision. Nvidia can work at 12 bit integer, 16 bit float, and 32 bit float. ATI works only at 24 bit float. There isn't actually a mode where they can be exactly compared. DX9 and ARB_fragment_program assume 32 bit float operation, and ATI just converts everything to 24 bit. For just about any given set of operations, the Nvidia card operating at 16 bit float will be faster than the ATI, while the Nvidia operating at 32 bit float will be slower. When DOOM runs the NV30 specific fragment shader, it is faster than the ATI, while if they both run the ARB2 shader, the ATI is faster.
When the output goes to a normal 32 bit framebuffer, as all current tests do, it is possible for Nvidia to analyze data flow from textures, constants, and attributes, and change many 32 bit operations to 16 or even 12 bit operations with absolutely no loss of quality or functionality. This is completely acceptable, and will benefit all applications, but will almost certainly induce hard to find bugs in the shader compiler. You can really go overboard with this -- if you wanted every last possible precision savings, you would need to examine texture dimensions and track vertex buffer data ranges for each shader binding. That would be a really poor architectural decision, but benchmark pressure pushes vendors to such lengths if they avoid outright cheating. If really aggressive compiler optimizations are implemented, I hope they include a hint or pragma for "debug mode" that skips all the optimizations.
To leap 50' in the air, you must be going 56.6 ft/s when leaving the ground, disregarding air resistance. Apogee will be in 1.77 seconds.
Assuming a linear acceleration, and a four foot period of acceleration from crouching to leaving the ground with legs extended, the average speed must be 28.3 ft/s over the four feet, for 0.14 seconds of acceleration, or 404 ft/s^2. 12.6 G's of acceleration isn't at all unreasonable for arm / leg contraction at light loads. You can make a >50G acceleration with a pitching motion of your arm.
12.6 G's of acceleration for an 800 pound hulk is only 10080 pounds, divided by two 24" long by 8" wide feet give a mere 26.25 psi force on the pavement.
If I botched these calculations, everyone is surely going to take the opportunity to say how the Armadillo vehicles will crash and burn...:-)
We have obviously been eagerly waiting for this unveiling. Nobody has denied that Rutan is the odds-on favorite for the X-Prize, but I take a positive thing away from this unveiling -- I have always contended that being an "airplane guy" is going to hurt Rutan in the X-Prize, and this is definitely a "winged thing". I would have been more concerned if it was just a purely ballistic capsule being air launched. I have little doubt that they will fairly rapidly have successful zoom climbs to somewhat above 100,000', but it is far from the simplest design to go to 350,000'. It is certainly true that complex designs can be made to work with enough talent, experience, testing, and money, which Rutan has all of, but there is plenty of room for things to screw up.
I don't expect that they will make any flights to 100km this year, but I can certainly be proven wrong...
I am quite happy with our current design, and we are committed to following through irrespective of what Rutan does. Even if he makes it, we have a different ecological niche in terms of vehicle capabilities -- our entire launch infrastructure can be towed by a light truck, and launched from anywhere. If he does win the X-Prize before us, we will ditch the monopropellant propulsion system and move to something more cost effective (at the expense of more development time) for the long term. We may be forced to do that anyway, if our peroxide situation doesn't resolve itself.
There is an interesting annecdote related to this.
At the world space congress last year, I was talking to Buzz Aldrin's son, who is head of acquisitions at Boeing. He really didn't believe that cheap, reusable launchers were possible (he thinks "billions of dollars in development"), but he said that if we win the X-Prize, demonstrating cheaper launch for even suborbital lobs, Boeing would "just buy us".
From our short discussion, it was clear that we have quite different world views, so I hesitate to read much into his statements one way or the other, but it was a bit curious.
I agree with some comments that this isn't exactly general interest news.
I am not interested in hearing from every chem major that is interested in starting a business (already heard from a couple, that's how I found out about the slashdot story). However, if anyone here does happen to have a brother-in-law that is a VP at FMC or some such, a little nudge wouldn't hurt.
The full story:
Rocket grade peroxide is stabilizer free, and 85% - 100% concentrated, as opposed to drug store peroxide at about 3% concentration. You can get up to 70% peroxide reasonably easily, but the high concentration stuff is a specialty item.
When we started our development work a bit over two years ago, we were doing some concentration of the peroxide ourselves, which is fine for making small test batches, but you really don't want to be making drums of the stuff, or you wind up spending as much time messing with that as you do building rockets.
We had some initial discussions with FMC about that time, but they weren't terribly encouraging. Shortly thereafter, we made contact with X-L Space Systems, a small company that was producing 98% concentration peroxide and selling it reasonably to several small outfits, as well as NASA and the USAF. I wound up buying a dozen or so drums from X-L, and everything was going well.
The owner of X-L was having such a hard time getting the government to pay their bills on time (he never had complaints about his small commercial customers) that he finally decided it just wasn't worth the headache, and he closed the company down. I was in discussion with him to make a large enough order to justify keeping production open, but we wouldn't need all that much peroxide for nearly eight months, so the storage logistics were looking troublesome. In hindsight, I should have worked something out, even if it was expensive or difficult.
About six months ago, we started contacting FMC again. The details haven't been very pleasant, largely because we keep thinking we are almost there, and it keeps not being the case. If they would just tell me exactly what I have to buy to make them happy, I would gladly do it, but they keep finding new things. That is the "stringing us along" part. They are mumbling again about lawyers and liability at the moment, which we thought had been worked through previously.
We have also spoken to Degussa about production, but they won't sell in drums, only large storage tanks (they supposedly have some drums in the US, but they are "promised to" NASA, and they won't sell them to us). We could live with that, but we broke off contact with them a while ago because FMC was sounding reasonable, but insisting that they be our sole supplier.
This is one of the unfortunate tradeoffs in modern society -- in the 70's, FMC would just ship drums of peroxide to the guys doing rocket powered dragsters without any hassles (one of them sent me a scan of some of his old shipping invoices). Today, fears of liability are larger than basic business drives like making money with your product. I'm not a "back in the good old days" sort, I fully recognize that the other advantages of modern society outweigh the nanny-state disadvantages, but one can always hope for across-the-board improvements.
Other than being almost out of peroxide, things are going very well for Armadillo. We rescheduled a lot of our development now that the X-Prize is fully funded, so we are parallel tracking full scale vehicle development with subscale flight testing.
>But he mentioned something about next gen cards having less bandwidth. Does that make sense to anyone?
The RATIO of bandwidth to calculation speed is going to decrease. It is nothing short of miraculous that ram bandwidth has made the progress is has, but adding gates is cheaper than adding more pins or increasing the clock on external lines.
Bandwidth will continue to increase, but calculation will likely get faster at an even better pace. If all calculations were still done in 8 bit, we would clearly be there with this generation, but bumping to 24/32 bit calculations while keeping the textures and framebuffer at 8 bit put the pressure on the calculations.
Yes, we are upset about it, and it will have some impact on how we deal with some companies in the future, but nothing drastic is going to change in terms of what support is going to be available.
Making any judgements from a snapshot intended for a non-interactive demo is ill advised.
> NASA has launched more missions than anybody else
NASA has launched more manned missions than anybody else, but the Russians have launched nearly TEN TIMES as many space mission.
This is when someone adds "Yeah they had to, because their electronics suck, so they need to replace their sats more often", but that doesn't change the point about launches.
My comment specifically regards the "shelf life" of a rendering engine. I think that an upcoming game engine, either the next one or the one after that, will have a notably longer usable life for content creation than we have seen so far. Instead of having to learn new paradigms for content creation every couple years, designers will be able to continue working with common tools that evolve in a compatible way. Renderman is the obvious example -- lots of things have improved and evolved, but its fundamental definition is clearly the same that it was over a decade ago.
This is only loosly related to the realism of the graphics. I don't think a detailed world simulation that is indistinquishable from reality will be here in the next decade, except for tightly controlled environments. You will be able to have real-time flythroughs that can qualify as indistinguishable, but given the ability to "test reality" interactively, we have a lot farther to go with simulation than with rendering.
I was fairly pleased with how that article turned out - when I first heard about it, I dreaded seeing a trivialized simplification of the issues, but it turned out as representative as you can be in that space.
However, I really dislike discussions of the attribution of techniques to a particular programmer. Everything is derived from things before it, and I make no claims of originality. I would say that one of my talents is the ability to be aware of what sources are feeding into my work, and be able to backtrack to them. Also, there are always lots of other possible answers for any given problem that can be made to work. BSP vs sector list, Portals vs PVS vs scan line occlusion, tilted constant Z rasterization vs block subdivision vs background divides, etc. Looked at in the proper perspective, individual techniques just aren't all that important. Sometimes it sounds like "Dude, he INVENTED needle nose pliers!!!"
Heck, I somewhat deride the very concept of originality. Creativity is just synthesis without the introspection. Lots of people will catch on that and start a rant about how Id games aren't original, but they are missing the point - it is possible to set out and develop something that will be received as "original" without ever having an "original" idea spring into your mind.
The best way to get answers is to just keep working the problem, recognizing when you are stalled, and directing the search pattern. Many of the popular notions of innovation and creativity are in some ways cop-outs that keep people from being as effective as they could be. The little document I wrote about developing a part of the shadow algorithm for Doom that Nvidia has on their website was a pretty good example of my process. Don't just wait for The Right Thing to strike you - try everything you think might even be in the right direction, so you can collect clues about the nature of the problem.
Slashdot Mirror
Actually, you underestimated this one.
Since the "payload" of an X-Prize vehicle is three x 200 lb people, needing 400 pounds of landing propellant turns our 850 gallon tank vehicle from a three person vehicle into a one person vehicle.
In the most negative light, you could say that powered landing (with a low performance propellant like we use) takes away two thirds of our payload capacity, but that is a poor metric to base decisions off of, because operational issues have historically been orders of magnitude more important to cost effectiveness than propellant consumption.
We can get the 400 pounds back by either going to a carbon fiber tank instead of a fiberglass tank (cost: $40k up front design fee, then $25k per tank, compared to $9k for the fiberglass tank), or by upsizing everything to a 1600 gallon fiberglass tank (cost: $17k for the tank plus more for bigger engine plumbing, catalysts, and nozzles).
Upsizing the tank is lower risk, because it only uses suppliers that we already buy from, while the carbon fiber tank job would be custom from ATK, and I have already had two other vendors back out on me for big tank work. We already have a 1600 gallon fiberglass tank on hand.
Our mixed monoprop has a measured sea level Isp of 145, with normal increases with altitude. Our big vehicles have a mass ratio of about five, takes off with somewhat under one positive G of acceleration, and has a somewhat regressive thrust profile from partial blowdown pressurization. That combination is sufficient for suborbital flights. A 200+ Isp biprop can do it with a mass ratio of three, but the vehicle gets a lot more complicated to build and operate.
John Carmack
> I figure you need about four times as much fuel at liftoff for a vertical rocket-borne landing as you would for a parachute- or wing-borne landing.
No, not even remotely close.
You only need enough propellant to kill the terminal velocity of the vehicle to land it safely. A vehicle that is stable reentering base first has a Cd right around 1.0, and any high performance rocket vehicle is going to be coming in pretty light after it has burned most of the propellant. The V2 impacted the ground still supersonic because it was aerodynamically stable nose first, so it maintained its 0.15 or so Cd on descent. A reasonably stubby base first reentry will have a terminal velocity of only 200 mph or so.
Killing that speed with a comfortable safety margin takes about 400 pounds of propellant in our vehicle, compared to 8000 pounds of propellant burned on ascent. A higher performance rocket engine could do it with propotionately smaller amounts. A parachute / drogue / ejection system for this weight vehicle is indeed lighter, coming in at about 100 pounds, but that brings a number of disadvantages with it, like coming down tens of miles away and still needing final impact attenuation.
We wanted to use parachtues as a quick hack for the X-Prize, but the test range where we were planning to fly was going to require a half million dollars of "engineering support" and wanted us to carry a thrust termination system (bomb) on the vehicle to satisfy themselves that it won't drift out of the range.
Long term, there is no question in our minds that powered landing is the way to go. We just were given a pretty strong incentive to go there earlier than we were planning.
John Carmack
A variety of responses:
We don't expect to win the X-Prize, both because Burt probably has it in the bag, and we are behind schedule. We still plan on continuing our development, because our designs are nearly an order of magnitude cheaper to fabricate and operate than Space Ship One, and orders of magnitude matter. If SS1 crashes on Monday, we will throw more time and resources at an attempt, because there really is no other contender, but it will be a long shot.
We could have flown an unguided rocket to very high altitudes a long time ago, but we have instead concentrated on control systems, which is where the important work needs to be done. A team that was busy flying rockets to hundreds of thousands of feet altitude, then decided to add a guidance and control system to their rockets would be in for many rude surprises at high energy levels.
This isn't immediately obvious, but an X-Prize class vehicle pretty much requires an active control system (a trained pilot with appropriate controls is also an active control system). A short burn time rocket, like the recent CSXT 100 km shot, can live with just aerodynamic stabilization (note that it also landed 20 miles away), but the G forces are far too high for people. As the burn time lengthens with lower acceleration forces, the vehicle will gravity turn away from vertical, making it almost impossible to keep a 60 second burn time even accelerating upwards.
People that harp on about propellant specific impulse in the context of suborbital rockets are like programmers that obsessively optimize a function that isn't a hot spot. The goal of a rocket ship is not to deliver specific impulse, it is to move a payload reliably and cost effectively. Isp can always be traded away for mass fraction, and quite often you can improve operability or reliability by doing so. With our new vehicle designs using a single engine and jet vanes instead of four differentially throttled engines we are more likely to consider trading some engine and system complexity for performance, but issues like the requirement for deep throttling still make it a complex decision.
I do Armadillo work on Tuesdays, weekends, and late at night. At Id lately I have been working on next-generation rendering technology while the rest of the company manages the Doom beta process.
I don't issue press releases. I just publicly write about what I am working on, and other people find it noteworthy enough to talk about. All of our development work, including the dead ends and mistakes, is fairly well documented on the Armadillo Aerospace website.
John Carmack
For those of you that are underwhelmed by the 310 pound vehicle, do note that the big vehicle (1500 lbs) that can actually carry people is also flying. Look back in the Armadillo updates around April 19 for testing video. We have since reworked the propulsion system to follow what has worked so well on the subscale vehicle, and should be testing it this weekend. If it works well, we will be repeating the boosted hop with the big vehicle next week.
The flight time is currently limited by federal law to 15 seconds of rocket burn time. We have a waiver coming to extend that to 120 seconds, but beyond that we will need a full launch license.
The significance of all this is that the vehicles are intended to fly up, come back down and land right where they took off from, all without ablating, expending, or seperating anything. It should be possible to have turn around times under one hour even for quite large vehicles.
BTW, Doom beta testing is going very well.
John Carmack
Someone that has done "some good things (NeXT, the first iMac, OS X)" in their career gets my respect.
Most of the negative tales about Jobs probably have some grounding in truth -- it was almost amusing watching him berate the stage people before a show for glitches in the prop moving systems: "What the hell is this??? Did you guys pick up these parts at Home Depot???". However, he did always listen when I was talking about a technical issue, even when I was saying something that didn't sit with his current understanding of graphics cards / APIs / gaming.
When I was considering setting up to demo Doom 3 at macworld, all of the Apple people were going on about how we needed to sanitize it because "Steve won't let there be any blood or killing". I finally went to him directly, and he replied "If you think you can make it great, then let's do it. I trust you, so you'll have to decide." Not quite the overbearing micromanager he is sometimes portrayed as.
I'm not a regular mac user, but I'm glad Steve Jobs is still around.
John Carmack
We have a good working relationship with AST, the division of the FAA that handles launch license, and we are one of only three companies (along with Scaled and XCOR) currently in the RLV launch license process. We have found all the people there helpful and eager to work with us. There is a lot of paperwork to be done, but we are working through it, and do not see a problem satisfying them. Things like calculating and minimizing expected third party casualty rates are obviously necessary and sensible.
The environmental aspects are less rational, with no analytical sense of scale.
Still, I'm only mildly concerned about the regulatory side of things. I think it will work out. None of our work is held up by any of this, so the worst case is that we have a vehicle built and tested repeatedly at the 200,000 lb-sec waivered impulse limit, with no launch license to allow us to fill the tank the rest of the way up. If that happens, THEN we get peeved about the situation, but continue flight testing with what we can.
Let me repeat: In no way have we been hampered by regulatory burden. Yet. We have been VERY hampered by commercial companies being too worried about liability exposure to work with us - peroxide companies, filament winders, and parachute companies have all caused us significant problems.
The supply issue with 90% peroxide basically cost us almost the entire year of flight testing. We spent the last six months developing a propellant combination that could conveniently replace the 90% peroxide based on widely available chemicals instead of the ultra-specialized propulsion grade. We are in the final optimizing and scale up phase of that. Instead of being irate about it, I try to look on the bright side - it is lots cheaper, easer to handle, and even a bit higher performance.
There are lots of problems still to be worked, but everything is coming along fine. We are behind schedule and somewhat over budget, but no worse off than any other project I have ever worked on...
John Carmack
High concentration hydrogen peroxide all by itself makes a low performance, but very convenient, rocket propellant. All hydrogen peroxide is in solution with some amount of water, because even if you had 100% peroxide, some of it would start decomposing to water (and oxygen) as you stored it.
Drugstore peroxide is 3% concentration. If you pour it on a catalyst, like silver or platinum, you will see bubbles forming in the solution (released oxygen), and the liquid will get somewhat warmer due to the released energy. Above roughly 70% concentration, the heat released is enough to vaporize all the water content, so if you pass it through a good catalyst, you will get all gas coming out the other side, and gas can be accelerated through a rocket nozzle to produce thrust. At 70%, the gas is only just above the boiling point of water, but as the concentration goes up, the temperature goes up fast. 90% peroxide, the most common grade used for propulsion, produces gas at about 1400 F temperature. Going all the way to 98% peroxide, the highest concentration produced, gives a few hundred degrees more temperature, but at a significant price increase. Higher temperature lets you use less propellant for a given amount of thrust-time, because it maintains a given chamber pressure with a less dense, but hotter, mixture (a simplification).
"Real" rocket propellants have temperature several thousand degrees higher, which does indeed increase performance, but the engines have to be cooled, and you need to manage both a fuel and an oxidizer in some form. One of our fundamental system trades is that it is better for an X-Prize class vehicle to use a propellant that simplifies vehicle engineering, even if you have to use more of it.
We use 90% peroxide from a small specialty supplier for all of our flight vehicles, but they closed shop a while ago, and we haven't been able to come to terms with the only domestic supplier of 90% peroxide, FMC chemical corp. Because of this, we have been working on alternate propellant schemes for a good part of this year, in parallel with building the full size X-Prize vehicle. If we had been able to just buy 90% peroxide like we buy all of our other industrial chemicals, we never would have bothered with the research.
Just about every week, someone asks why we don't concentrate it ourselves. True, dozens of people have made a few gallons of high concentration peroxide at various times, but there have only been two large scale concentrators operated in the US outside of the official manufacturers - Rotary Rocket had a concentrator, but it only went to 85% concentration, and it didn't do purification, and Beal Aerospace had a large scale concentrator operational after the blew up their first one. Sure, we could figure out how to do it, but then we would be in the chemical plant business instead of the rocket business, and that's not what we want to do. I am funding an operator in Houston to produce a few thousand pounds of 90% for us, but he is six months behind schedule on delivery, which proves my point about it not being as simple as people think.
The direction we have been pursuing is using a combination of 50% peroxide, which is readily available through distributors from multiple manufacturers, and a small amount of miscible fuel (methanol in our current work). 50% peroxide by itself doesn't work as a rocket propellant, because you can't boil all the water, which makes even decomposing most of the peroxide difficult. Adding a fuel and (the tricky part!) getting it to burn with the released oxygen gives you the energy necessary to vaporize the water and get everything up to a high temperature. Mixing fuels with high concentration oxidizers usually makes a touchy and deadly explosive (we have intentionally detonated a mix of 90% peroxide and alcohol - Very Scary), but buffered with 50% water, and running off of stoichemetric mixture ratio, the risk is not very high. We have a study report from the Department of Mines in the late 50's investigating th
Just building the vehicle costs less than $100k, most of the money is in building multiple iterations of everything as you figure out exactly how you actually need to spend the money:
$ 6k 850 gallon fiberglass tank
$ 2k High pressure carbon fiber pressurant tank and regulator
$ 1k Honeycomb composite panels
$ 5k Aluminum fabrication for cabin
$15k Redundant parachutes, drogues, drogue cannons, releases
$13k Fiber optic gyro based IMU
$ 8k Unrestricted (supersonic / high altitude) GPS
$ 2k PC104 systems
$ 5k video, audio, and data communications
$20k Engine machining, catalysts, laser cut plates
$ 5k Plumbing, valves, etc
$ 5k Fastblock external insulation
For powered landings instead of parachute landings, delete the parachutes and add:
$ 4k Laser altimeter
$ 4k Wire rope isolator landing gear
You could trivially spend an order of magnitude more by just using "space certified" versions of everything, but the important point is that standard industrial versions of many things are perfectly adequate. In many cases, todays standard industrial practice is far ahead of the best that could be done at any price in the early sixties.
This is all with free labor for assembly and testing, but that is still only a couple hundred man hours for a full vehicle. We are expecting to destroy the first vehicle in some (unmanned) testing mishap along the way, and build another one mostly from scratch. That will take less than two months, depending on lead times for some items.
John Carmack
You probably mean "Burt Rutan", the aircraft designer at Scaled. Dick Rutan is his brother, who piloted the voyager, and was the test pilot for XCOR's EZ-Rocket, but doesn't have anything to do with Space Ship One, the X-Prize vehicle.
I have always maintained that Burt is the odds-on favorite to win the X-Prize, but it isn't over yet. His design requires a pilot on board for all tests, so there is a non-negligable chance that there could be a fatality, which would almost certainly end the effort in the X-Prize timeframe.
John Carmack
The refractory metals are better, but less commonly used. Columbium/niobium is reasonable to form. Molybdenum and alloys like TZM take a bit more heat, but have a potentially annoing ductile to brittle transition point for systems that will cold soak. The state of the art is irridium coated rhenium, which doesn't melt until 2466 C / 4471 F.
We fabricated a TZM chamber a while ago at fairly high expense, but still burned through it after an extended length run:
burned TZM
This experience has convinced me that active cooling methods, like transpiration cooling, are probably a good idea for high reusability reentry vehicles.
John Carmack
>Based on the first two responses to this post, you'd think people had never heard of inertial
>navigation. With MEMS accelerometers it ought to be pretty light, too.
Pure 3 axis inertial navigation with a strapdown inertial measuring requires extreme precision. MEMS inertial units aren't even in the right ballpark. Mechanical stable platform inertial systems that actually rotated inside the vehicles didn't require awesomely accurate sensors, but they are big, heavy, and not as reliable.
It is a useful programming exercise to write a simulation of a strapdown inertial system and play with bias, noise, and nonlinearity errors (add cross axis coupling and acceleration effects for micromachined gyros for bonus points). Pick reasonable ranges and quantize to 12 bits, then integrate at 100 hz or so. You can start the simulation motionless, but in a minute it will be cruising along at 60 mph in some random direction, hundreds of feet from the start position. An hour later, it will be heading for Mars.
The low end inertial systems that have been moderately soccessful are done by removing gravity from the equation and just doing 2D navigation, and often using other sensors, like magnetometers instead of rate gyros for heading, or odometer readings instead of double integrating accelerometers. Double integration of interrelated noisy sensors with an implicit 1G acceleration is really more demanding than it would initially seem.
The only reason you wouldn't want to use GPS in an ocean crossing is if you are afraid a Bad Guy might be jamming the signals.
John Carmack
Arguably the most professional and widely viewed machinima so far is the music video for Zero 7's "In the Waiting Line", produced by my wife's company,
Fountainhead Entertainment. This was a real, commercial production using machinima tools.
It was neat to see the Q3 engine playing on MTV, but it made me greatly regret the quantized normals in Q3 models, which resulted in a noticeable popping on the environment maps. This was largely my motivation for adding per-pixel environment map calculation to the new Doom engine (under the ARB2 path, at least).
John Carmack
>IAARS. (I Am A Rocket Scientist.) ...
>
>The Shuttle uses LOX and LH2, both of which are f'nasty to deal with and are economical only to
>generate the immense thrust necessary to achieve orbit. While in orbit, the Orbiter maneuvers
>using (relatively) small hydrazine thrusters. N2H4 is also f'nasty, but somewhat less so than
>either LOX or LH2.
???
The OMS uses hydrazine / nitrogen tetroxide, which is way, WAY more nasty than LOX / LH2.
LOX / LH2 are cryogens, and contact with them will give you frostbite. Hydrazine is carcinogenic and toxic, but nitrogen tetroxide is roughly as poisonous as the best war gasses from WWI. Plus, it has very low surface tension, so when it spills, it spreads extremely rapidly, which causes it to vaporize even faster than the already high vapor pressure would indicate. The various oxides of nitrogen are famous for the "BFRC" ( big red cloud ) that results from spills, which you should run away from very fast.
John Carmack
Rewriting shaders behind an application's back in a way that changes the output under non-controlled circumstances is absolutely, positively wrong and indefensible.
Rewriting a shader so that it does exactly the same thing, but in a more efficient way, is generally acceptable compiler optimization, but there is a range of defensibility from completely generic instruction scheduling that helps almost everyone, to exact shader comparisons that only help one specific application. Full shader comparisons are morally grungy, but not deeply evil.
The significant issue that clouds current ATI / Nvidia comparisons is fragment shader precision. Nvidia can work at 12 bit integer, 16 bit float, and 32 bit float. ATI works only at 24 bit float. There isn't actually a mode where they can be exactly compared. DX9 and ARB_fragment_program assume 32 bit float operation, and ATI just converts everything to 24 bit. For just about any given set of operations, the Nvidia card operating at 16 bit float will be faster than the ATI, while the Nvidia operating at 32 bit float will be slower. When DOOM runs the NV30 specific fragment shader, it is faster than the ATI, while if they both run the ARB2 shader, the ATI is faster.
When the output goes to a normal 32 bit framebuffer, as all current tests do, it is possible for Nvidia to analyze data flow from textures, constants, and attributes, and change many 32 bit operations to 16 or even 12 bit operations with absolutely no loss of quality or functionality. This is completely acceptable, and will benefit all applications, but will almost certainly induce hard to find bugs in the shader compiler. You can really go overboard with this -- if you wanted every last possible precision savings, you would need to examine texture dimensions and track vertex buffer data ranges for each shader binding. That would be a really poor architectural decision, but benchmark pressure pushes vendors to such lengths if they avoid outright cheating. If really aggressive compiler optimizations are implemented, I hope they include a hint or pragma for "debug mode" that skips all the optimizations.
John Carmack
To leap 50' in the air, you must be going 56.6 ft/s when leaving the ground, disregarding air resistance. Apogee will be in 1.77 seconds.
:-)
Assuming a linear acceleration, and a four foot period of acceleration from crouching to leaving the ground with legs extended, the average speed must be 28.3 ft/s over the four feet, for 0.14 seconds of acceleration, or 404 ft/s^2. 12.6 G's of acceleration isn't at all unreasonable for arm / leg contraction at light loads. You can make a >50G acceleration with a pitching motion of your arm.
12.6 G's of acceleration for an 800 pound hulk is only 10080 pounds, divided by two 24" long by 8" wide feet give a mere 26.25 psi force on the pavement.
If I botched these calculations, everyone is surely going to take the opportunity to say how the Armadillo vehicles will crash and burn...
John Carmack
We have obviously been eagerly waiting for this unveiling. Nobody has denied that Rutan is the odds-on favorite for the X-Prize, but I take a positive thing away from this unveiling -- I have always contended that being an "airplane guy" is going to hurt Rutan in the X-Prize, and this is definitely a "winged thing". I would have been more concerned if it was just a purely ballistic capsule being air launched. I have little doubt that they will fairly rapidly have successful zoom climbs to somewhat above 100,000', but it is far from the simplest design to go to 350,000'. It is certainly true that complex designs can be made to work with enough talent, experience, testing, and money, which Rutan has all of, but there is plenty of room for things to screw up.
I don't expect that they will make any flights to 100km this year, but I can certainly be proven wrong...
I am quite happy with our current design, and we are committed to following through irrespective of what Rutan does. Even if he makes it, we have a different ecological niche in terms of vehicle capabilities -- our entire launch infrastructure can be towed by a light truck, and launched from anywhere. If he does win the X-Prize before us, we will ditch the monopropellant propulsion system and move to something more cost effective (at the expense of more development time) for the long term. We may be forced to do that anyway, if our peroxide situation doesn't resolve itself.
John Carmack
There is an interesting annecdote related to this.
At the world space congress last year, I was talking to Buzz Aldrin's son, who is head of acquisitions at Boeing. He really didn't believe that cheap, reusable launchers were possible (he thinks "billions of dollars in development"), but he said that if we win the X-Prize, demonstrating cheaper launch for even suborbital lobs, Boeing would "just buy us".
From our short discussion, it was clear that we have quite different world views, so I hesitate to read much into his statements one way or the other, but it was a bit curious.
John Carmack
I agree with some comments that this isn't exactly general interest news.
I am not interested in hearing from every chem major that is interested in starting a business (already heard from a couple, that's how I found out about the slashdot story). However, if anyone here does happen to have a brother-in-law that is a VP at FMC or some such, a little nudge wouldn't hurt.
The full story:
Rocket grade peroxide is stabilizer free, and 85% - 100% concentrated, as opposed to drug store peroxide at about 3% concentration. You can get up to 70% peroxide reasonably easily, but the high concentration stuff is a specialty item.
When we started our development work a bit over two years ago, we were doing some concentration of the peroxide ourselves, which is fine for making small test batches, but you really don't want to be making drums of the stuff, or you wind up spending as much time messing with that as you do building rockets.
We had some initial discussions with FMC about that time, but they weren't terribly encouraging. Shortly thereafter, we made contact with X-L Space Systems, a small company that was producing 98% concentration peroxide and selling it reasonably to several small outfits, as well as NASA and the USAF. I wound up buying a dozen or so drums from X-L, and everything was going well.
The owner of X-L was having such a hard time getting the government to pay their bills on time (he never had complaints about his small commercial customers) that he finally decided it just wasn't worth the headache, and he closed the company down. I was in discussion with him to make a large enough order to justify keeping production open, but we wouldn't need all that much peroxide for nearly eight months, so the storage logistics were looking troublesome. In hindsight, I should have worked something out, even if it was expensive or difficult.
About six months ago, we started contacting FMC again. The details haven't been very pleasant, largely because we keep thinking we are almost there, and it keeps not being the case. If they would just tell me exactly what I have to buy to make them happy, I would gladly do it, but they keep finding new things. That is the "stringing us along" part. They are mumbling again about lawyers and liability at the moment, which we thought had been worked through previously.
We have also spoken to Degussa about production, but they won't sell in drums, only large storage tanks (they supposedly have some drums in the US, but they are "promised to" NASA, and they won't sell them to us). We could live with that, but we broke off contact with them a while ago because FMC was sounding reasonable, but insisting that they be our sole supplier.
This is one of the unfortunate tradeoffs in modern society -- in the 70's, FMC would just ship drums of peroxide to the guys doing rocket powered dragsters without any hassles (one of them sent me a scan of some of his old shipping invoices). Today, fears of liability are larger than basic business drives like making money with your product. I'm not a "back in the good old days" sort, I fully recognize that the other advantages of modern society outweigh the nanny-state disadvantages, but one can always hope for across-the-board improvements.
Other than being almost out of peroxide, things are going very well for Armadillo. We rescheduled a lot of our development now that the X-Prize is fully funded, so we are parallel tracking full scale vehicle development with subscale flight testing.
John Carmack
>But he mentioned something about next gen cards having less bandwidth. Does that make sense to anyone?
The RATIO of bandwidth to calculation speed is going to decrease. It is nothing short of miraculous that ram bandwidth has made the progress is has, but adding gates is cheaper than adding more pins or increasing the clock on external lines.
Bandwidth will continue to increase, but calculation will likely get faster at an even better pace. If all calculations were still done in 8 bit, we would clearly be there with this generation, but bumping to 24/32 bit calculations while keeping the textures and framebuffer at 8 bit put the pressure on the calculations.
John Carmack
No, this was not leaked on purpose.
Yes, we are upset about it, and it will have some impact on how we deal with some companies in the future, but nothing drastic is going to change in terms of what support is going to be available.
Making any judgements from a snapshot intended for a non-interactive demo is ill advised.
John Carmack
The number I saw quoted was $150k, not $250k.
John Carmack
> but similar attempts in the past have cost about $20,000
Ky has said it cost over $250,000. Just building the rocket and motor probably cost $20k, but everything else adds up.
John Carmack
> NASA has launched more missions than anybody else
NASA has launched more manned missions than anybody else, but the Russians have launched nearly TEN TIMES as many space mission.
This is when someone adds "Yeah they had to, because their electronics suck, so they need to replace their sats more often", but that doesn't change the point about launches.
John Carmack
My comment specifically regards the "shelf life" of a rendering engine. I think that an upcoming game engine, either the next one or the one after that, will have a notably longer usable life for content creation than we have seen so far. Instead of having to learn new paradigms for content creation every couple years, designers will be able to continue working with common tools that evolve in a compatible way. Renderman is the obvious example -- lots of things have improved and evolved, but its fundamental definition is clearly the same that it was over a decade ago.
This is only loosly related to the realism of the graphics. I don't think a detailed world simulation that is indistinquishable from reality will be here in the next decade, except for tightly controlled environments. You will be able to have real-time flythroughs that can qualify as indistinguishable, but given the ability to "test reality" interactively, we have a lot farther to go with simulation than with rendering.
John Carmack
I was fairly pleased with how that article turned out - when I first heard about it, I dreaded seeing a trivialized simplification of the issues, but it turned out as representative as you can be in that space.
However, I really dislike discussions of the attribution of techniques to a particular programmer. Everything is derived from things before it, and I make no claims of originality. I would say that one of my talents is the ability to be aware of what sources are feeding into my work, and be able to backtrack to them. Also, there are always lots of other possible answers for any given problem that can be made to work. BSP vs sector list, Portals vs PVS vs scan line occlusion, tilted constant Z rasterization vs block subdivision vs background divides, etc. Looked at in the proper perspective, individual techniques just aren't all that important. Sometimes it sounds like "Dude, he INVENTED needle nose pliers!!!"
Heck, I somewhat deride the very concept of originality. Creativity is just synthesis without the introspection. Lots of people will catch on that and start a rant about how Id games aren't original, but they are missing the point - it is possible to set out and develop something that will be received as "original" without ever having an "original" idea spring into your mind.
The best way to get answers is to just keep working the problem, recognizing when you are stalled, and directing the search pattern. Many of the popular notions of innovation and creativity are in some ways cop-outs that keep people from being as effective as they could be. The little document I wrote about developing a part of the shadow algorithm for Doom that Nvidia has on their website was a pretty good example of my process. Don't just wait for The Right Thing to strike you - try everything you think might even be in the right direction, so you can collect clues about the nature of the problem.
John Carmack