Do you really blackhole IPs, though? While it might be computationally more expensive to run baysian filters on each message, surely this would reduce the number of legitimate messages rejected. Also, why not use a whitelist system...if the email fails a filter test, and that distinct message is only being sent to ONE aol email, then the message is put in a holding queue and the originator is sent a message saying his email failed the spam filter, if he wants it to reach its destination please enter the numbers found in this box in a reply.
No, but they have killed 10s of millions total. Certainly thousands were killed total in Tienamen square, though of course most of the deaths happened off camera after the protesters went home, and were arrested later. Even if you assume the subject executed in every American state government execution is innocent, it comes nowhere close to the number killed by the Chinese.
You're missing the point. For all the people in China to experience the advantages of a first world society, they would need the equivalent of much more than a car for each person in resources. Even if those resourfes went to something more practical, such as automated electric cars or vacuum tunnels (vac tunnels would require to construct MANY times the resources of cars, but the absense of friction would reduce the amount of energy to run it to near zero. If there's no friction, its possible to reclaim all the energy used to accelerate each vehicle...no matter how high a speed the vehicles are accelerated to. However, superconducting magnets are needed to stop the friction the electrons experience, and so there would still be cooling costs. Also switching costs. But the energy each car is given can be reclaimed when it is slowed down, even if said car reaches a speed of several miles/second)
Anyway, even if they invested those resources in more practical items the resources have to come from somewhere. Actually, getting them from outer space is far more difficult than expanding current technology, and using nuclear energy since it might damage the biosphere to burn the hydrocarbons required. (whether fusion or reprocessing fission)
I'm pretty certain that's Tux. I've got a stuffed penguin like that somewhere around the house, and the color and size is exactly the same. While the face microsoft shows the world is their marketing and management side, believe it or not to make (mostly) working software they have to have many thousands of geeks. While their corporate masters may encourage them to do the bad stuff they are famous for, that doesn't mean they can't joke around.
Umm, no... A $100 card in this case would be a Radeon 8500 which was the fastest card available a year ago... And while the $100 version is just the card, it has a long list of features that do loads more than just play games. It is a very high quality card.
Yep, its about "penis envy". A $100 or less card will run all the games you really want to play right now with essentially full quality (ok, so you might not be able to turn on every performance killing feature from the driver control panel but the increase in visual detail is negligible). Also, even when games do come out that finally use the power of a card more powerful than you have, they may not be games you really want to play. Take Unreal 2, for instance. While not a "bad" game per say...it was good enough for me to finish it...it certainly wasn't worth spending money on a video card to play it, if I didn't already have one. It offerred nothing new, basically the sterile non-interactive environments of Quake 2 (all you can do is shoot monsters and flip levers, and every lever must be flipped. Dialogue makes no difference, characters die in the plot but you cannot save any of them) dressed up with better graphics. I want more to DO in the game, like Deus Ex had. For instance, in the game there's an entire level of the inside of your spaceship that gets revisited periodically. You can wander around, even going into maintenance sections. But you cannot touch or interact with the environment in any way! Nothing you say to characters makes even the slightest bit of difference, and you cannot do anything but open doors.
Its rare that a game comes out that both has system destroying graphics prompting an upgrade AND is actually a game you want to play. While Doom 3 may look good, it is yet to be seen whether its even as good as half life single player. Also, its unlikely it comes even close to a classic like Battlefield 1942 or Half-Life mods for multiplayer.
You are aware, I hope, that you probably gave as good as you got. If I were one of those shy nerds, discovering you were actually pretty smart and trying to work up the courage to ask you out, I'd feel pretty shitty seeing you making out in the hall with some bonehead wearing torn jeans, a black teeshirt, and carrying his test paper with a C on it. I'd feel absolutely lousy in discovering not even the girls with similar interest were attracted to me. Hypothetically speaking, of course.
And what sounds like HUGE overhead costs. The program probably eats memory, bandwidth, and cpu like a hog. Look at how they designed it : probably by coding up the whizbang features in the highest level abstractions then getting their developer to somehow make it work.
So? AOL is one of the crappiest pieces of bloatware I've used. It has a HUGE number of features, animated widgets, ect. I was so happy to get it off my system. IRC may be clean, tiny, and powerful but guess which will get more use? Besides, as Bill has pointed out it people needing huge amounts of computing power and bandwidth so they can waste nearly all of it running bloatware is what drives the tech sector.
Well, this is a common argument and there's a simple escape : yes, the computer isn't any better than those who programmed it...if the programmers had the ability to think for the thousands of years of subjective time required to play chess like Deep Blue. (That is, it would take thousands of years for a human to work out the problem like it does)
Presumably a self improving AI would be subject to the same argument. Sure, it wouldn't be any smarter than the people who programmed it (and taught it)...if the people had a few spare million years and the ability to change the workings of their own minds. You can tell this argument to the kill-bot when it comes for you, I'm sure it'll be amused for the microsecond it takes to process it.
Unfortunatly I have to say that this argument is very tired, old, and verging on troll. People have been saying "the hardware is ahead of the software. I won't need to upgrade for sooo long". Its true, and its tiring. Games are sold to the mass market, to run on the machines the majority of players have. This means they tend to run fine on machines a year or two older than the best currently available. Right now there are games that will strain a Geforce3, which was released about 15 months ago.
Well, this isn't a new idea. But there's problems with it. The biggest is that the existing living computing cells we have to work with are very, very slow. Yes, they work in parallel and may actually do calculations using electrons at a quantum level. We should be able to duplicate this at some point with more conventional techniques, though. Also, while there would seem like living cells would have a cost and efficiency advantage, our current techniques for building microchips are themselves rather efficient. Lithography is somewhat similar to xeroxing endless copies based on a template. Most of the cost and complexity is involved in quality control, design, and contaiminant management....which wouldn't go away if we used neurons.
That actually sounds interesting. So its an MMORPG set in the real world, where you roadtrip around the country and duel other players? Course, while you're at it would be even more fun to add some GTA3 style elements...the ability to go on a cross country crime spree could be interesting, if well simulated. The only problem is part of the fun of GTA3 is being a thug to anyone without consequences. If EVERYONE in the game (who's a human player) is a thug, it might not be very fun...noone to beat up.
As for better lift tech : hey, I'm all for it. The 20 or 30 pathfinder example was what I think could be done 2 years from now (that is, using 3 billion dollars to produce in quantity existing mars probe designs), with the space shuttle museumed and no more resources spent on it except to finish finger pointing after the disaster.
Hmm. Thing is, no matter how much you improve lifting technology it will STILL BE CHEAPER to send unmanned vehicles (however they get up there). After all, if its unmanned a failure rate of 5% or less is fine. Heck, you could get away with higher than that (whatever the optimal ratio between quality control costs and losses due to failures). But if there are men aboard people won't be happy unless that rate is made as near to zero as humanly possible Thus every part has to be checked ten times, with more paperwork overhead generated than the space shuttle weighs. After every mission, every piece has to be exhaustively tested. Every imaginable contingency has to be thought of and planned against. Even the most unlikely disaster might need a safety system custom designed to prevent said disaster.
Also, another observation : if an unmanned mission fails : say my proposal to put factories up there completely flops. Its not a total disaster. The technology developed to build the factories now exists, and can be used elsewhere. There's not nearly the stigma of getting people killed, so perhaps next year the funders might try again. A far more productive outcome than a disaster with a manned mission, where all activity might be ceased for years while the fingers are pointed. Resources get dumped into even tighter quality control than ever, rather than new ideas and technology. A total mess, and a waste of millions of man hours of human lifespan. Imagine how much more the workers at NASA could contribute to society if they did something more productive than triple checking every part for the space shuttle. What if instead they are trying out radical new robots, changing the mission plan totally on the fly to use the space probe in a different way (all of which has happened with our probe programs) Mot to mention, if something goes wrong they merely feel guilty about losing a few big bucks of taxpayer money than a human being.
Do you understand the concept of "if it takes more resources to gather a resource than said resource provides, you lose resources". For instance, if it takes more energy to dig the oil out of shale than you get from burning this oil, you are getting nowhere. Current technology would require expending far more energy to construct orbital power stations than the station could ever gather.
Right on. You see the truth here. As for "funny stuff a human might notice" : well, in the short term before anything resembling real AI is built, human operators would analyze the output from the robots. The probes would drive around exploring on their own perhaps, and be ordered to investigate interesting phenomena by their human masters when they see it. And if you are just exploring the moon, the speed of light lag is low enough that direct remote control operation is practical. You'd have about good a senses through the probe's high res cameras and senors as you'd get inside a bulky, armored spacesuit and have more time to explore to boot.
Also, without all the "overhead" supporting humans most of our space program's resources would go to its stated purpose....developing and trying new techology and learning about our universe, not blasting test pilots into space.
Actually, sir, you are wrong here. Nuclear propulsion is inherently VERY, VERY, dangerous if its used in the boost phase. You have a hot, running nuclear pile. It has to have LIGHTWEIGHT SHIELDING. It has to produce an enormous amount of energy for the first few seconds during liftoff, to minimize the propellant used. If it melts down, you have hot radioactive debris everywhere. A fusion plant, even if possible, will be many decades, maybe centuries away before one with the power/weight ratio exists, if ever (think of all the lasers or magnets needed...much weight). There's an enormous difference between using a hot nuke plant to reach orbit and using a regular rocket carrying cool, freshly made fuel rods.
The costs to minimize this danger, and the liability if it fails, would make the space shuttle seem cheap.
However, there is in fact a third option you have not mentioned. A laser beamed from the ground would superheat an inert propellant block on the spaceship. Pulsed in the right timing, and it would generated a planar shockwave. No thrust nozzles or anything needed. Merely a heavy cube of propellant and the spacecraft bolted on top, as well as some sort of stabilization system. Much safer, nothing to explode, astronaut escape vehicles possible, and a far far better propellant/payload ratio (the laser would heat the propellant up at least 10 times hotter than a conventional rocket can reach). For the initial liftoff phase a short linear accelerator might be used to give the spacecraft its initial kick (and providing a safety delta V if the laser on the ground fails)
This is something that would make space travel feasible in mass. Since the main power system, and most of the complexity , is on the ground (plenty of room to have backups then) the maintainence and operation costs would be far lower. Its estimated that unmanned payloads might reach orbit for perhaps $60 a pound (just the electricity to run the laser).
Mod me down for saying this, but the honest truth may be the best next step in space exploration is to drop the manned program entirely, and spent the money on better remote probes and satellites. Three billion a year would buy at least 10, probably 20 or 30 pathfinder probes (or an improved model) per year. That's a lot of mars exploration. This isn't a popular view, but there are some convincing arguments.
First, one of the stated goals for the space program is to develop new technology. But when are you more likely to use the latest and greatest bleeding edge experimental engine? On a manned spacecraft where loss is catastophic to the whole program, or a relatively cheap robot? Fact is, the pathfinder mission used some of the fastest processors and lots of new off the shelf technology. They had some bugs with it, which is why it can't be used with a manned mission. Sometimes this approach (known to the press as "better faster cheaper") fails, but the point is its SO much cheaper than a single manned mission a failure is not really that big an issue. For the price of one year of shuttle launches we could send dozens of probes to mars (as said before).
Be honest here. While its said that manned exploration is a precursor to manned colonization, the hard fact is that it takes too much energy to put people in orbit. For a very long, long time it will be easier to use advancing technology to support more people on this earth than move them to space. Besides that, humans aren't adapted to live in space. The basic plan has always been to go to the final frontier...then build a huge enclosed, sheltered colony that the human colonists huddle in 99% of the time. Its like going to the Grand Canyon or Yellowstone then huddling in your Winnebago all week.
A far more realistic plan is to create a life that can live there. I imagine "big clanking replicators" : a huge factory with fairly familar machinery, all of it automated and only requiring human supervision to perform repairs. Mining machines, robotic rock haulers, nuclear power plants, smelters, presses, lathes, ect...most of the robotic tech similar to what you would find in a general motors plant. This facility would be built on the moon, remotely operated by people on earth. It would be capable of constructing the parts to build another facility (and so on). While expensive, it would be a fraction of the cost of human missions, and after enough replications be able to produce useful products.
Unmanned boosters blow up 4% of the time, and its nothing but a finanical nuisance. I've just described a plan that would develop far more advanced, bleeding edge tech than anything that could be used in a manned mission. The technology developed (better industrial automation, better artificial intelligence, better remote telepresense) would be immediatly useful on earth. A manned trip to mars would involve mostly old, proven technology, with a few exotic exceptions necessary for the mission. (such as a nuclear propulsion system, something NOT usable on earth)
I understand why noone will listen to me : there's an incredible glamour about blasting off our heroes into orbit, sending a man out in space to get the job done. Hell, I want to go too. But the truth is, without all the overhead associated with minimizing the risks to said heroes a lot more could be accomplished with the same money. In addition, the new tech and perhaps even real products from space would eventually provide a real return on investment, enriching us on the ground.
"the 180 gig HD's will do you just fine". Well, go join an anime divx sharing channel. Or fire up a p2p app for a few months. I know people who have over a terrabyte of divx anime movies, and have to burn new stuff onto dvds to keep up. Sure, right now it looks like 10 terrabytes would be enough to hold all the video you could watch (just like a few gigs can hold more text than you'll ever have time to read in your lifetime), but by that time uses for that much space should become apparent.
Something bugged me about your statement : "the bubble around the carrier is a 1.000 nautical miles radius (http://www.chinfo.navy.mil/navpalib/factfile/airc raft/air-fa18.html), so what's the point of making 35 more miles in an hour?".
The point is not making the circle LARGER, the point is how fast you can MOVE the whole circle. DOUBLING the carrier's speed would enable you to put that "threat circle" over a hostile country twice as fast. For patrol purposes, this might let you say, threaten Iraq one day then zip over to North Korea the next.
"that I don't see why one would immediately make the other totally irrelevant, they could very well complement each other". Conservatively, said sentient computer would have several advantages that would in fact make human brains irrelevant.
1. Said computer would have access to its own schematics, and the tools used by its designers to build it. It would be capable of self improvement to a far greater degree than humans can ever do. There's no reason this couldn't be a recursive process.
2. Said computer would be able to copy its underlying hardware and personality in a mass production facility, churning out fully knowledgable and skilled copies by the millions. (instead of 9 months plus 20 years training for humans and the process often fails at some point)
3. Said computer might think the same way as humans....1000 or one million times faster. That alone would make human thought irrelevant. I would wager you could accomplish more than anyone on earth if you had 1000 times the thinking speed and first accumulated all human knowledge.
4. The only bad sci fi here is the motives of the machine. Maybe it wouldn't try to do something destructive...but that doesn't mean human thinkers would have much meaning at this point.
However, to get to mars you don't need a "strong radioactive source". Freshly enriched plutonium is NOT...it is an alpha emitter, so a sheet of thin foil is plenty protection. It is chemically not much more toxic than any other heavy metal, and anyways as you'll see in the coming days big pieces of the shuttle wreckage will be nearly completly intact. Armored capsules were made for the nuclear generators brought to the moon in the Apollo program, a similar method could be used for a nuclear rocket program.
You are aware that there's a big difference between transporting enriched nuclear fuel rods to be used in a reactor (the pellets the rods are made of when new are only slightly radioactive, and safe to handle with gloves) inside sealed metal and ceramic capsules tested to take the shock of an explosion and crashing to the ground and live nuclear bombs able to be armed and detonated remotely.
Even if not every carrier brings along a few FA-18s, they generally also have maybe 1000 SAMS or more stored in the escort ships. I assume the best of them can probably reliably destroy any non-stealth aircraft made, F-18s included. If not, my tax dollars are going to waste, because there's absolutely no reason why a rocket that can take 30g turns would not be able to find and hit its target.
Slashdot Mirror
Do you really blackhole IPs, though? While it might be computationally more expensive to run baysian filters on each message, surely this would reduce the number of legitimate messages rejected. Also, why not use a whitelist system...if the email fails a filter test, and that distinct message is only being sent to ONE aol email, then the message is put in a holding queue and the originator is sent a message saying his email failed the spam filter, if he wants it to reach its destination please enter the numbers found in this box in a reply.
No, but they have killed 10s of millions total. Certainly thousands were killed total in Tienamen square, though of course most of the deaths happened off camera after the protesters went home, and were arrested later. Even if you assume the subject executed in every American state government execution is innocent, it comes nowhere close to the number killed by the Chinese.
You're missing the point. For all the people in China to experience the advantages of a first world society, they would need the equivalent of much more than a car for each person in resources. Even if those resourfes went to something more practical, such as automated electric cars or vacuum tunnels (vac tunnels would require to construct MANY times the resources of cars, but the absense of friction would reduce the amount of energy to run it to near zero. If there's no friction, its possible to reclaim all the energy used to accelerate each vehicle...no matter how high a speed the vehicles are accelerated to. However, superconducting magnets are needed to stop the friction the electrons experience, and so there would still be cooling costs. Also switching costs. But the energy each car is given can be reclaimed when it is slowed down, even if said car reaches a speed of several miles/second)
Anyway, even if they invested those resources in more practical items the resources have to come from somewhere. Actually, getting them from outer space is far more difficult than expanding current technology, and using nuclear energy since it might damage the biosphere to burn the hydrocarbons required. (whether fusion or reprocessing fission)
I'm pretty certain that's Tux. I've got a stuffed penguin like that somewhere around the house, and the color and size is exactly the same. While the face microsoft shows the world is their marketing and management side, believe it or not to make (mostly) working software they have to have many thousands of geeks. While their corporate masters may encourage them to do the bad stuff they are famous for, that doesn't mean they can't joke around.
Umm, no... A $100 card in this case would be a Radeon 8500 which was the fastest card available a year ago... And while the $100 version is just the card, it has a long list of features that do loads more than just play games. It is a very high quality card.
Yep, its about "penis envy". A $100 or less card will run all the games you really want to play right now with essentially full quality (ok, so you might not be able to turn on every performance killing feature from the driver control panel but the increase in visual detail is negligible). Also, even when games do come out that finally use the power of a card more powerful than you have, they may not be games you really want to play. Take Unreal 2, for instance. While not a "bad" game per say...it was good enough for me to finish it...it certainly wasn't worth spending money on a video card to play it, if I didn't already have one. It offerred nothing new, basically the sterile non-interactive environments of Quake 2 (all you can do is shoot monsters and flip levers, and every lever must be flipped. Dialogue makes no difference, characters die in the plot but you cannot save any of them) dressed up with better graphics. I want more to DO in the game, like Deus Ex had. For instance, in the game there's an entire level of the inside of your spaceship that gets revisited periodically. You can wander around, even going into maintenance sections. But you cannot touch or interact with the environment in any way! Nothing you say to characters makes even the slightest bit of difference, and you cannot do anything but open doors.
Its rare that a game comes out that both has system destroying graphics prompting an upgrade AND is actually a game you want to play. While Doom 3 may look good, it is yet to be seen whether its even as good as half life single player. Also, its unlikely it comes even close to a classic like Battlefield 1942 or Half-Life mods for multiplayer.
You are aware, I hope, that you probably gave as good as you got. If I were one of those shy nerds, discovering you were actually pretty smart and trying to work up the courage to ask you out, I'd feel pretty shitty seeing you making out in the hall with some bonehead wearing torn jeans, a black teeshirt, and carrying his test paper with a C on it. I'd feel absolutely lousy in discovering not even the girls with similar interest were attracted to me. Hypothetically speaking, of course.
And what sounds like HUGE overhead costs. The program probably eats memory, bandwidth, and cpu like a hog. Look at how they designed it : probably by coding up the whizbang features in the highest level abstractions then getting their developer to somehow make it work.
So? AOL is one of the crappiest pieces of bloatware I've used. It has a HUGE number of features, animated widgets, ect. I was so happy to get it off my system. IRC may be clean, tiny, and powerful but guess which will get more use? Besides, as Bill has pointed out it people needing huge amounts of computing power and bandwidth so they can waste nearly all of it running bloatware is what drives the tech sector.
Or found one cc number (his mom's) and spread it via trojan or worm to 2.2 million comps...
Well, this is a common argument and there's a simple escape : yes, the computer isn't any better than those who programmed it...if the programmers had the ability to think for the thousands of years of subjective time required to play chess like Deep Blue. (That is, it would take thousands of years for a human to work out the problem like it does)
Presumably a self improving AI would be subject to the same argument. Sure, it wouldn't be any smarter than the people who programmed it (and taught it)...if the people had a few spare million years and the ability to change the workings of their own minds. You can tell this argument to the kill-bot when it comes for you, I'm sure it'll be amused for the microsecond it takes to process it.
Unfortunatly I have to say that this argument is very tired, old, and verging on troll. People have been saying "the hardware is ahead of the software. I won't need to upgrade for sooo long". Its true, and its tiring. Games are sold to the mass market, to run on the machines the majority of players have. This means they tend to run fine on machines a year or two older than the best currently available. Right now there are games that will strain a Geforce3, which was released about 15 months ago.
Well, this isn't a new idea. But there's problems with it. The biggest is that the existing living computing cells we have to work with are very, very slow. Yes, they work in parallel and may actually do calculations using electrons at a quantum level. We should be able to duplicate this at some point with more conventional techniques, though. Also, while there would seem like living cells would have a cost and efficiency advantage, our current techniques for building microchips are themselves rather efficient. Lithography is somewhat similar to xeroxing endless copies based on a template. Most of the cost and complexity is involved in quality control, design, and contaiminant management....which wouldn't go away if we used neurons.
That actually sounds interesting. So its an MMORPG set in the real world, where you roadtrip around the country and duel other players? Course, while you're at it would be even more fun to add some GTA3 style elements...the ability to go on a cross country crime spree could be interesting, if well simulated. The only problem is part of the fun of GTA3 is being a thug to anyone without consequences. If EVERYONE in the game (who's a human player) is a thug, it might not be very fun...noone to beat up.
As for better lift tech : hey, I'm all for it. The 20 or 30 pathfinder example was what I think could be done 2 years from now (that is, using 3 billion dollars to produce in quantity existing mars probe designs), with the space shuttle museumed and no more resources spent on it except to finish finger pointing after the disaster.
Hmm. Thing is, no matter how much you improve lifting technology it will STILL BE CHEAPER to send unmanned vehicles (however they get up there). After all, if its unmanned a failure rate of 5% or less is fine. Heck, you could get away with higher than that (whatever the optimal ratio between quality control costs and losses due to failures). But if there are men aboard people won't be happy unless that rate is made as near to zero as humanly possible Thus every part has to be checked ten times, with more paperwork overhead generated than the space shuttle weighs. After every mission, every piece has to be exhaustively tested. Every imaginable contingency has to be thought of and planned against. Even the most unlikely disaster might need a safety system custom designed to prevent said disaster.
Also, another observation : if an unmanned mission fails : say my proposal to put factories up there completely flops. Its not a total disaster. The technology developed to build the factories now exists, and can be used elsewhere. There's not nearly the stigma of getting people killed, so perhaps next year the funders might try again. A far more productive outcome than a disaster with a manned mission, where all activity might be ceased for years while the fingers are pointed. Resources get dumped into even tighter quality control than ever, rather than new ideas and technology. A total mess, and a waste of millions of man hours of human lifespan. Imagine how much more the workers at NASA could contribute to society if they did something more productive than triple checking every part for the space shuttle. What if instead they are trying out radical new robots, changing the mission plan totally on the fly to use the space probe in a different way (all of which has happened with our probe programs) Mot to mention, if something goes wrong they merely feel guilty about losing a few big bucks of taxpayer money than a human being.
Do you understand the concept of "if it takes more resources to gather a resource than said resource provides, you lose resources". For instance, if it takes more energy to dig the oil out of shale than you get from burning this oil, you are getting nowhere. Current technology would require expending far more energy to construct orbital power stations than the station could ever gather.
Right on. You see the truth here. As for "funny stuff a human might notice" : well, in the short term before anything resembling real AI is built, human operators would analyze the output from the robots. The probes would drive around exploring on their own perhaps, and be ordered to investigate interesting phenomena by their human masters when they see it. And if you are just exploring the moon, the speed of light lag is low enough that direct remote control operation is practical. You'd have about good a senses through the probe's high res cameras and senors as you'd get inside a bulky, armored spacesuit and have more time to explore to boot. Also, without all the "overhead" supporting humans most of our space program's resources would go to its stated purpose....developing and trying new techology and learning about our universe, not blasting test pilots into space.
Actually, sir, you are wrong here. Nuclear propulsion is inherently VERY, VERY, dangerous if its used in the boost phase. You have a hot, running nuclear pile. It has to have LIGHTWEIGHT SHIELDING. It has to produce an enormous amount of energy for the first few seconds during liftoff, to minimize the propellant used. If it melts down, you have hot radioactive debris everywhere. A fusion plant, even if possible, will be many decades, maybe centuries away before one with the power/weight ratio exists, if ever (think of all the lasers or magnets needed...much weight). There's an enormous difference between using a hot nuke plant to reach orbit and using a regular rocket carrying cool, freshly made fuel rods.
The costs to minimize this danger, and the liability if it fails, would make the space shuttle seem cheap.
However, there is in fact a third option you have not mentioned. A laser beamed from the ground would superheat an inert propellant block on the spaceship. Pulsed in the right timing, and it would generated a planar shockwave. No thrust nozzles or anything needed. Merely a heavy cube of propellant and the spacecraft bolted on top, as well as some sort of stabilization system. Much safer, nothing to explode, astronaut escape vehicles possible, and a far far better propellant/payload ratio (the laser would heat the propellant up at least 10 times hotter than a conventional rocket can reach). For the initial liftoff phase a short linear accelerator might be used to give the spacecraft its initial kick (and providing a safety delta V if the laser on the ground fails)
This is something that would make space travel feasible in mass. Since the main power system, and most of the complexity , is on the ground (plenty of room to have backups then) the maintainence and operation costs would be far lower. Its estimated that unmanned payloads might reach orbit for perhaps $60 a pound (just the electricity to run the laser).
Mod me down for saying this, but the honest truth may be the best next step in space exploration is to drop the manned program entirely, and spent the money on better remote probes and satellites. Three billion a year would buy at least 10, probably 20 or 30 pathfinder probes (or an improved model) per year. That's a lot of mars exploration. This isn't a popular view, but there are some convincing arguments.
First, one of the stated goals for the space program is to develop new technology. But when are you more likely to use the latest and greatest bleeding edge experimental engine? On a manned spacecraft where loss is catastophic to the whole program, or a relatively cheap robot? Fact is, the pathfinder mission used some of the fastest processors and lots of new off the shelf technology. They had some bugs with it, which is why it can't be used with a manned mission. Sometimes this approach (known to the press as "better faster cheaper") fails, but the point is its SO much cheaper than a single manned mission a failure is not really that big an issue. For the price of one year of shuttle launches we could send dozens of probes to mars (as said before).
Be honest here. While its said that manned exploration is a precursor to manned colonization, the hard fact is that it takes too much energy to put people in orbit. For a very long, long time it will be easier to use advancing technology to support more people on this earth than move them to space. Besides that, humans aren't adapted to live in space. The basic plan has always been to go to the final frontier...then build a huge enclosed, sheltered colony that the human colonists huddle in 99% of the time. Its like going to the Grand Canyon or Yellowstone then huddling in your Winnebago all week.
A far more realistic plan is to create a life that can live there. I imagine "big clanking replicators" : a huge factory with fairly familar machinery, all of it automated and only requiring human supervision to perform repairs. Mining machines, robotic rock haulers, nuclear power plants, smelters, presses, lathes, ect...most of the robotic tech similar to what you would find in a general motors plant. This facility would be built on the moon, remotely operated by people on earth. It would be capable of constructing the parts to build another facility (and so on). While expensive, it would be a fraction of the cost of human missions, and after enough replications be able to produce useful products.
Unmanned boosters blow up 4% of the time, and its nothing but a finanical nuisance. I've just described a plan that would develop far more advanced, bleeding edge tech than anything that could be used in a manned mission. The technology developed (better industrial automation, better artificial intelligence, better remote telepresense) would be immediatly useful on earth. A manned trip to mars would involve mostly old, proven technology, with a few exotic exceptions necessary for the mission. (such as a nuclear propulsion system, something NOT usable on earth)
I understand why noone will listen to me : there's an incredible glamour about blasting off our heroes into orbit, sending a man out in space to get the job done. Hell, I want to go too. But the truth is, without all the overhead associated with minimizing the risks to said heroes a lot more could be accomplished with the same money. In addition, the new tech and perhaps even real products from space would eventually provide a real return on investment, enriching us on the ground.
"the 180 gig HD's will do you just fine". Well, go join an anime divx sharing channel. Or fire up a p2p app for a few months. I know people who have over a terrabyte of divx anime movies, and have to burn new stuff onto dvds to keep up. Sure, right now it looks like 10 terrabytes would be enough to hold all the video you could watch (just like a few gigs can hold more text than you'll ever have time to read in your lifetime), but by that time uses for that much space should become apparent.
Something bugged me about your statement : "the bubble around the carrier is a 1.000 nautical miles radius (http://www.chinfo.navy.mil/navpalib/factfile/airc raft/air-fa18.html), so what's the point of making 35 more miles in an hour?".
The point is not making the circle LARGER, the point is how fast you can MOVE the whole circle. DOUBLING the carrier's speed would enable you to put that "threat circle" over a hostile country twice as fast. For patrol purposes, this might let you say, threaten Iraq one day then zip over to North Korea the next.
"that I don't see why one would immediately make the other totally irrelevant, they could very well complement each other". Conservatively, said sentient computer would have several advantages that would in fact make human brains irrelevant.
1. Said computer would have access to its own schematics, and the tools used by its designers to build it. It would be capable of self improvement to a far greater degree than humans can ever do. There's no reason this couldn't be a recursive process.
2. Said computer would be able to copy its underlying hardware and personality in a mass production facility, churning out fully knowledgable and skilled copies by the millions. (instead of 9 months plus 20 years training for humans and the process often fails at some point)
3. Said computer might think the same way as humans....1000 or one million times faster. That alone would make human thought irrelevant. I would wager you could accomplish more than anyone on earth if you had 1000 times the thinking speed and first accumulated all human knowledge.
4. The only bad sci fi here is the motives of the machine. Maybe it wouldn't try to do something destructive...but that doesn't mean human thinkers would have much meaning at this point.
However, to get to mars you don't need a "strong radioactive source". Freshly enriched plutonium is NOT...it is an alpha emitter, so a sheet of thin foil is plenty protection. It is chemically not much more toxic than any other heavy metal, and anyways as you'll see in the coming days big pieces of the shuttle wreckage will be nearly completly intact. Armored capsules were made for the nuclear generators brought to the moon in the Apollo program, a similar method could be used for a nuclear rocket program.
You are aware that there's a big difference between transporting enriched nuclear fuel rods to be used in a reactor (the pellets the rods are made of when new are only slightly radioactive, and safe to handle with gloves) inside sealed metal and ceramic capsules tested to take the shock of an explosion and crashing to the ground and live nuclear bombs able to be armed and detonated remotely.
Even if not every carrier brings along a few FA-18s, they generally also have maybe 1000 SAMS or more stored in the escort ships. I assume the best of them can probably reliably destroy any non-stealth aircraft made, F-18s included. If not, my tax dollars are going to waste, because there's absolutely no reason why a rocket that can take 30g turns would not be able to find and hit its target.