How funny! I guess since I was learning about it at the time, and the Cray wasn't RISC, Cray got no credit for it. All the credit went to the MIPS guys, and the 68000 and PowerPC guys.
And here Cray invented it!
Can't say I'm totally surprised though. I still remember being sad when I heard he had passed... he was a great engineer.
Cray's innovative design was... Reduced Instruction Set
No - Cray didn't do RISC, he did Vector math. I don't have a cite - I did look - but it came up in my CPU design class way back when. (When I took the class, the Cray-1 was still pretty cool!)
If you honestly believe that "value brought to society" is the metric
I think that is the proper metric for basically any profession... and I do agree that software engineering as practiced today has nothing to do with computer science as taught today - and that this is a problem. I think where we disagree is that you seem to think that industry is wrong, and I think that academia is wrong.
To a certain extent, both are useful - I see the value in computer science in academia, with the trickle down effects. I just think that if the goal is to give students (that are going to earn a living creating code) a valuable education in exchange for their money, that goal is not achieved by moving education away from the practical and towards the theoretical. In other words, academia is being payed by students to prepare them for industry - the answer (in most cases) is not to say that industry is wrong.
(And just to clarify - I have written a LOT of code, used in a lot of different systems. I maintain it, and do virtually all the debugging. My code almost always works the first time, and I don't use formal proof methods. But I also take longer to examine procedures where my companies could lose money than to examine procedures that are less important - that sort of thing is the reason my salary is 3 standard deviations from the mean. I don't say this to brag, I say this to show the value in thinking about "maximizing value to society" rather than correctness.)
The tidal locking that makes one side of the moon face us is fairly well understood, I think. Essentially: you start with two large masses in orbit around each other. These masses are not perfectly uniformly distributed, and so "tides" appear - the Earth (and more visibly, the oceans) move "around" the equator following the moon. This slows the planet a tiny amount - and eventually the two objects always show the same side to each other.
Obviously, the moon completed this a long time ago - but the Earth will presumably tidally lock to the moon eventually.
Hmmm... thinking about this, what if there were 2 explosions - the first one separated the bodies, and the second one happened later, adding the orbital momentum to the moon?
Still, I think the two-body solution is the more likely one. But I recommend that we use this theory as an excuse to return to the moon and excavate the core! (Excavation will resolve all issues - proof left to the student as an exercise).
Well, when judging between two possible causes of an outcome, choosing the one that happens 99% of the time is smarter than choosing the one that is 1 in a million.
It is far more likely that an external object imparted angular momentum to the mass that became the moon than that the moon was ejected from the Earth and happened to hit a one in a million chance and get the angular momentum some other way.
Capture happens all the time. Explosions happen too - but as far as I know never have been seen to result in a stable, two body system.
Therefore, it seem much more likely that the moon was created from a two-body interaction (not capture - the moon was fried too thoroughly for that) than from a single body event.
The sun's gravity accelerates things near the Earth by 0.006 m/sec^2. Two objects close enough to be meaningfully gravitationally linked while orbiting the sun in the Earth's orbit will have a maximum differential acceleration of maybe a thousandth of that. So to get to 1000m/s takes 5 years.
So this essentially posits that an explosion had enough force to blown the planet apart, and send the pieces into space, but not to escape velocity (11.2 km/s) but instead to a velocity just short of that (11.19 km/s or so), so that the moon goes flying away for 2.5 years but 2.5 years later comes back and settles into a nice, circular orbit.
That would be hard to accomplish on purpose - saying an accident did it is beyond belief.
That would seem to require a fairly rapidly spinning ancient Earth, though. At least a possible explanation, but for momentum to be conserved the Earth would have to have been originally spinning at 2-3 RPM. That seems fast to me for a planet, but I guess larger things have been seen to spin faster (pulsars, etc).
True. I guess my real issue is that what students are taught in Computer Science degrees (mostly about the thing you mention - correctness proofs, computation theory, etc.) does not adequately prepare them for their real job (deciding when and what algorithms to use).
For example, saving a few bits of storage in the guidance system of a rocket is a bad optimization - as you say. But there are many times where incorrect optimizations are the correct way to solve a problem. For example, did you know that one of the early Cray computer hardware was designed (on purpose) using an algorithm (I think for division) that was very fast, but did not always produce the correct result. This was a good design - the incorrect result was a very rare case that could be checked for in software.
Often, the most important thing in software is not correctness - and the fact that you poo-poo Microsoft (and, yes, admittedly I poo-poo them too) really just seals my point. Microsoft was the best software developer on the planet - and it was not because of their ability to write good code. The metric of a software engineer is not "lines of bug free code", or something like that - it is "value brought to society".
So you have to balance many things - time, money, effort, and yes, correctness. But in our world 100% correctness is only marginally more valuable than 99% correctness - and the difference in resources required to achieve those standards is enormous.
I'm not sure I agree - The moon has an ungodly amount of angular momentum. I'm having trouble coming up with a method whereby a section of object a leaves object a, and then has enough thrust perpendicular to the direction of object a to get up to it's 1km/s orbital velocity.
Interestingly enough, you can just use the rocket exhaust to keep the chamber at vacuum using a properly shaped duct. Or just use really big pumps... but this kind of thing is done quite a bit.
But that is exactly the wrong way to think about things! If you only think about why it won't work, you'll never make anything new!
For example, you could say here: in order to get high efficiencies you would have to sit in a waveguide with nearly perfect reflectors on both ends. That could lead to buildings designed as such waveguides, etc. With new materials being developed with negative indexes of refraction at useful wavelengths, impossible waveguides are just more expensive.
not like to have my reproductive organs anywhere near such a device
Similarly, the obvious solution is to remove your reproductive organs at the door! Schnick!
Very hard to estimate, honestly. Most of the cost is not in the materials - it is in the manufacturing process. First, the Ariane 5 has two solid rocket motors, each 105 tons. You have to make a 105 ton solid object with no bubbles (the fuel is a lot like rubber - you pour it into a mold). You then have to inspect in, xray it, etc.
So what is "propellant" cost, and what is "structure" cost?
My rule of thumb is that solid propellant costs about $10/pound for the mix - but that is a lot more "estimaty" than the other rule of thumb. So that is about $300/pound delivered to orbit (propellant cost) if someone built as all-solid rocket.
I think the shuttle SRBs cost a few tens of millions. Ariane's SRBs are probably a few million dollars - but I don't know the Ariane's true pricing, of course.
If the elevator costs a lot and ends up only lasting 10 years, it'll be a lot more expensive than conventional rockets.
Just two things:
1) Comparing space elevators to "conventional rockets" is silly - carbon nanotubes used in space elevators will do wonders on rockets as well.
2) In another post, I share that a space elevator, due to it's length, can expect to be cut through by microscopic space junk about every ten days or so. So upkeep is a huge issue - probably swamps rocket costs over the long term.
I was assuming that the Ariane propellant costs would be similar.
Yes - very silly to launch a space station every time you want a satellite in orbit. For comparison, the Shuttle uses 630 tons of LOX and 100 tons of liquid hydrogen - about 4 times as much! Also, a lot of the propellant cost of the space shuttle is in the solid rockets - not only are solids more expensive, but they also require closer tolerances and inspections of the propellant. Solids are cheap to design and build, but relatively expensive to refuel - but as I say, people should ignore that, the cost of space access has nothing to do with fuel costs.
Maybe some day they'll figure out how to do a reusable launch vehicle that doesn't actually cost more than using an expendable launch vehicle.
I think that can be shortened to "maybe someday they'll make a reusable launch vehicle." The shuttle is basically rebuilt every launch... as you say.
(Of course, I must add the disclaimer that I am currently working on such a vehicle - which is, of course, way better that all the other vehicles with billion dollar budgets! Cuz weesums smart!)
But it can take 8 tonnes to geostationary, and the Shuttle can only take about 4 tonnes.
Well, the shuttle can't take anything to GEO - you would need a secondary booster of some kind (which they won't allow on the shuttle anymore, due to safety issues). I also don't know the numbers for GEO by heart - so here is the LEO analysis.
The shuttle can take 22,000 kg to LEO, the Ariane 5 can lift about the same, 16,000 kg. The Ariane 5 uses 130 tons of LOX, and 25 tons of liquid hydrogen. LOX is about $0.01 per pound in large quantities (remember, it is air!) - so that costs about $10,000! Liquid hydrogen costs about $10/lb - so that is $500K. Any solids used cost more, but this shows how ridiculously cheap propellant is for rockets!
Heh - I once proposed a rocket with mass approx 6*10^24 kg. Very efficient, amazing Isp to get to Leo, and very simple to build: you just take the parts of the Earth that remain after your rocket (only a few kg), and launch those using a rocket... (I don't think anyone got the joke, though)
Instead of taking a rocket to leave Earth, you send Earth away in a rocket... much more efficient!
1) Using a laser to provide power to an object 15,000 km away is tricky... possible, but very tricky. 2) Something like this would provide something (the wire would be turning with the Earth and thus the field, but the field strength would still be modulating). It would not provide much energy compared to a GW power plant, however. 3) Rocket fuel costs essentially nothing. The Space Shuttle costs half a billion per flight or so, the propellant costs are a few tens of millions. A good rule of thumb is that LEO is $20/kg in propellant. 4) Rockets and elevators would have very different advantages - they could easily work side by side. One caution - do not compare today's rockets to tomorrow's elevators. The technology that will enable space elevators will work wonders on rockets as well.
if people would just get up off their lazy butts and get working.;-}
The real problems with space elevator-like designs are maintenance. You have a 35,000km long ribbon, in an environment where a 1 cm diameter cable will be cut on average once every 1000 km/years - so you would expect the ribbon to be cut every ten days or so. Even if you come up with a design that you can repair fast enough to keep up, most of the up-mass is repair stuff.
I agree, but I like to compare the Isp of elevators to rockets - I use milli-earths-seconds!
More seriously, another way to think about it (that looks at mass efficiency) is to look at the expected lifetime of the elevator (in kg delivered) divided by the mass of the elevator, and compare that to the expected lifetime of a similarly non-existing RLV made with nanotubes (in kg delivered) divided by the mass of the RLV.
That's what kills the space elevator, really. A nanotube RLV will be pretty darn impressive!
Look closely at what I wrote - I do not disagree with you. I was saying that the same object in the same orbit has the same energy no matter how it got there.
The percentage of energy expended that goes into the final object varies wildly - but the final object energy is the same.
So as to your statement:
reduce it to gravitational potential energy
note that no matter how you get to GEO, you have to expend the same mgh (gravitational potential energy) and 0.5mv^2 (kinetic energy). The only difference is in the energy imparted into your reaction mass. A rocket's reaction mass is the fuel of the rocket - a space elevator's reaction mass is the entire Earth. Conservation of momentum linked with e=0.5mv^2 says that the larger reaction mass is far better.
GEO is 42,000 km radius. Earth's radius is 6400 km. Gravity acceleration goes down with the square of radius, so the acceleration due to gravity drops from 9.8 m/s to 0.2 m/s. Substantially weaker, but not zero - you still need a high velocity to stay in orbit.
In fact, normal low Earth orbital velocity is 7.8 km/s - GEO still needs 3 km/s.
My point on conversion energy is that an object place in GEO by rocket has exactly the same energy as that object delivered by the elevator - the only question is where the energy comes from.
A concern with talking about the efficiency of rockets is that you have to carefully define what you mean: normal chemical rockets have extremely high Carnot efficiencies, mediocre mass and energy efficiencies. A space elevator doesn't have a Carnot efficiency, has terrible mass efficiency, but extremely good energy efficiency. An ion thruster has no Carnot efficiency, has great mass efficiency, and terrible energy efficiency.
And these are all just from the basic physics of the devices - not a whole lot you can do about it really.
1) Space elevators do not lower the energy required - they just use the energy differently. 2) They do not take you to where the gravity is weak - they take you to the point where the force of gravity (which is essentially unchanged) is balanced by centripetal force (which, being linked to w^2r goes up linearly with distance). 3) Rockets typically take you to about 7.7 km/s (orbit), not 11.2km/s (escape). 4) The energy given to the satellite (assuming the same final orbit) is identical regardless of the launch vehicle/elevator used. What is different is the energy efficiency of the system in putting energy into the satellite:
A rocket sends lightweight propellant in the opposite direction very fast in order to transfer the energy. An elevator sends a huge mass (essentially the entire earth) very slowly in the opposite direction. Since momentum is conserved, the mass x velocity of both systems is the same - but since the Earth masses a lot more than most rockets, the Earth's relative velocity is far lower. This is where the e=0.5*m*v^2 comes in - the "wasted" energy is the energy provided to the Earth or propellant. Earth has a small v, big m - which works better than the rockets big v little m.
So you always have to give the satellite the same energy - there are just different efficiencies of giving it that energy. Space cannons have the problem of needing to give that energy extremely quickly... very difficult indeed.
Explain then why the masterminds of 9/11 decided to confess this morning, so that they could be executed? They say that they want to die as martyrs and get their virgins...
I can see the point, really. My point is that, while I agree that what Dijikstra does is not a software engineer, that does not mean that software engineers do not exist. Microsoft was the most successful software company in the world. Let me ask you - was their ability to write bug free code a factor?
No - software engineering is not about producing bug free code. Software engineering is about producing software. Balancing time, money, and equipment against production.
You may not do it, but that doesn't mean others don't.
(And, your right, the Arianne software error was a silly waste - and almost certainly should have been caught. But formal proofs would not have caught it... the implementation error would have required looking at things at the hardware level, not the algorithm level.)
The wraps should come off next March or so. I'd give you the company name, but honestly it will most likely change before then (engineers choose working names, marketing chooses launch names).
If it all works right, you shouldn't be able to miss it, though!
Slashdot Mirror
How funny! I guess since I was learning about it at the time, and the Cray wasn't RISC, Cray got no credit for it. All the credit went to the MIPS guys, and the 68000 and PowerPC guys.
And here Cray invented it!
Can't say I'm totally surprised though. I still remember being sad when I heard he had passed... he was a great engineer.
Cray's innovative design was ... Reduced Instruction Set
No - Cray didn't do RISC, he did Vector math. I don't have a cite - I did look - but it came up in my CPU design class way back when. (When I took the class, the Cray-1 was still pretty cool!)
If you honestly believe that "value brought to society" is the metric
I think that is the proper metric for basically any profession... and I do agree that software engineering as practiced today has nothing to do with computer science as taught today - and that this is a problem. I think where we disagree is that you seem to think that industry is wrong, and I think that academia is wrong.
To a certain extent, both are useful - I see the value in computer science in academia, with the trickle down effects. I just think that if the goal is to give students (that are going to earn a living creating code) a valuable education in exchange for their money, that goal is not achieved by moving education away from the practical and towards the theoretical. In other words, academia is being payed by students to prepare them for industry - the answer (in most cases) is not to say that industry is wrong.
(And just to clarify - I have written a LOT of code, used in a lot of different systems. I maintain it, and do virtually all the debugging. My code almost always works the first time, and I don't use formal proof methods. But I also take longer to examine procedures where my companies could lose money than to examine procedures that are less important - that sort of thing is the reason my salary is 3 standard deviations from the mean. I don't say this to brag, I say this to show the value in thinking about "maximizing value to society" rather than correctness.)
The tidal locking that makes one side of the moon face us is fairly well understood, I think. Essentially: you start with two large masses in orbit around each other. These masses are not perfectly uniformly distributed, and so "tides" appear - the Earth (and more visibly, the oceans) move "around" the equator following the moon. This slows the planet a tiny amount - and eventually the two objects always show the same side to each other.
Obviously, the moon completed this a long time ago - but the Earth will presumably tidally lock to the moon eventually.
Hmmm... thinking about this, what if there were 2 explosions - the first one separated the bodies, and the second one happened later, adding the orbital momentum to the moon?
Still, I think the two-body solution is the more likely one. But I recommend that we use this theory as an excuse to return to the moon and excavate the core! (Excavation will resolve all issues - proof left to the student as an exercise).
Well, when judging between two possible causes of an outcome, choosing the one that happens 99% of the time is smarter than choosing the one that is 1 in a million.
It is far more likely that an external object imparted angular momentum to the mass that became the moon than that the moon was ejected from the Earth and happened to hit a one in a million chance and get the angular momentum some other way.
Capture happens all the time. Explosions happen too - but as far as I know never have been seen to result in a stable, two body system.
Therefore, it seem much more likely that the moon was created from a two-body interaction (not capture - the moon was fried too thoroughly for that) than from a single body event.
The sun's gravity accelerates things near the Earth by 0.006 m/sec^2. Two objects close enough to be meaningfully gravitationally linked while orbiting the sun in the Earth's orbit will have a maximum differential acceleration of maybe a thousandth of that. So to get to 1000m/s takes 5 years.
So this essentially posits that an explosion had enough force to blown the planet apart, and send the pieces into space, but not to escape velocity (11.2 km/s) but instead to a velocity just short of that (11.19 km/s or so), so that the moon goes flying away for 2.5 years but 2.5 years later comes back and settles into a nice, circular orbit.
That would be hard to accomplish on purpose - saying an accident did it is beyond belief.
That would seem to require a fairly rapidly spinning ancient Earth, though. At least a possible explanation, but for momentum to be conserved the Earth would have to have been originally spinning at 2-3 RPM. That seems fast to me for a planet, but I guess larger things have been seen to spin faster (pulsars, etc).
True. I guess my real issue is that what students are taught in Computer Science degrees (mostly about the thing you mention - correctness proofs, computation theory, etc.) does not adequately prepare them for their real job (deciding when and what algorithms to use).
For example, saving a few bits of storage in the guidance system of a rocket is a bad optimization - as you say. But there are many times where incorrect optimizations are the correct way to solve a problem. For example, did you know that one of the early Cray computer hardware was designed (on purpose) using an algorithm (I think for division) that was very fast, but did not always produce the correct result. This was a good design - the incorrect result was a very rare case that could be checked for in software.
Often, the most important thing in software is not correctness - and the fact that you poo-poo Microsoft (and, yes, admittedly I poo-poo them too) really just seals my point. Microsoft was the best software developer on the planet - and it was not because of their ability to write good code. The metric of a software engineer is not "lines of bug free code", or something like that - it is "value brought to society".
So you have to balance many things - time, money, effort, and yes, correctness. But in our world 100% correctness is only marginally more valuable than 99% correctness - and the difference in resources required to achieve those standards is enormous.
I'm not sure I agree - The moon has an ungodly amount of angular momentum. I'm having trouble coming up with a method whereby a section of object a leaves object a, and then has enough thrust perpendicular to the direction of object a to get up to it's 1km/s orbital velocity.
Interestingly enough, you can just use the rocket exhaust to keep the chamber at vacuum using a properly shaped duct. Or just use really big pumps... but this kind of thing is done quite a bit.
The only way to get around (b) is to
But that is exactly the wrong way to think about things! If you only think about why it won't work, you'll never make anything new!
For example, you could say here: in order to get high efficiencies you would have to sit in a waveguide with nearly perfect reflectors on both ends. That could lead to buildings designed as such waveguides, etc. With new materials being developed with negative indexes of refraction at useful wavelengths, impossible waveguides are just more expensive.
not like to have my reproductive organs anywhere near such a device
Similarly, the obvious solution is to remove your reproductive organs at the door! Schnick!
Very hard to estimate, honestly. Most of the cost is not in the materials - it is in the manufacturing process. First, the Ariane 5 has two solid rocket motors, each 105 tons. You have to make a 105 ton solid object with no bubbles (the fuel is a lot like rubber - you pour it into a mold). You then have to inspect in, xray it, etc.
So what is "propellant" cost, and what is "structure" cost?
My rule of thumb is that solid propellant costs about $10/pound for the mix - but that is a lot more "estimaty" than the other rule of thumb. So that is about $300/pound delivered to orbit (propellant cost) if someone built as all-solid rocket.
I think the shuttle SRBs cost a few tens of millions. Ariane's SRBs are probably a few million dollars - but I don't know the Ariane's true pricing, of course.
If the elevator costs a lot and ends up only lasting 10 years, it'll be a lot more expensive than conventional rockets.
Just two things:
1) Comparing space elevators to "conventional rockets" is silly - carbon nanotubes used in space elevators will do wonders on rockets as well.
2) In another post, I share that a space elevator, due to it's length, can expect to be cut through by microscopic space junk about every ten days or so. So upkeep is a huge issue - probably swamps rocket costs over the long term.
I was assuming that the Ariane propellant costs would be similar.
Yes - very silly to launch a space station every time you want a satellite in orbit. For comparison, the Shuttle uses 630 tons of LOX and 100 tons of liquid hydrogen - about 4 times as much! Also, a lot of the propellant cost of the space shuttle is in the solid rockets - not only are solids more expensive, but they also require closer tolerances and inspections of the propellant. Solids are cheap to design and build, but relatively expensive to refuel - but as I say, people should ignore that, the cost of space access has nothing to do with fuel costs.
Maybe some day they'll figure out how to do a reusable launch vehicle that doesn't actually cost more than using an expendable launch vehicle.
I think that can be shortened to "maybe someday they'll make a reusable launch vehicle." The shuttle is basically rebuilt every launch... as you say.
(Of course, I must add the disclaimer that I am currently working on such a vehicle - which is, of course, way better that all the other vehicles with billion dollar budgets! Cuz weesums smart!)
But it can take 8 tonnes to geostationary, and the Shuttle can only take about 4 tonnes.
Well, the shuttle can't take anything to GEO - you would need a secondary booster of some kind (which they won't allow on the shuttle anymore, due to safety issues). I also don't know the numbers for GEO by heart - so here is the LEO analysis.
The shuttle can take 22,000 kg to LEO, the Ariane 5 can lift about the same, 16,000 kg. The Ariane 5 uses 130 tons of LOX, and 25 tons of liquid hydrogen. LOX is about $0.01 per pound in large quantities (remember, it is air!) - so that costs about $10,000! Liquid hydrogen costs about $10/lb - so that is $500K. Any solids used cost more, but this shows how ridiculously cheap propellant is for rockets!
Heh - I once proposed a rocket with mass approx 6*10^24 kg. Very efficient, amazing Isp to get to Leo, and very simple to build: you just take the parts of the Earth that remain after your rocket (only a few kg), and launch those using a rocket... (I don't think anyone got the joke, though)
Instead of taking a rocket to leave Earth, you send Earth away in a rocket... much more efficient!
1) Using a laser to provide power to an object 15,000 km away is tricky... possible, but very tricky.
2) Something like this would provide something (the wire would be turning with the Earth and thus the field, but the field strength would still be modulating). It would not provide much energy compared to a GW power plant, however.
3) Rocket fuel costs essentially nothing. The Space Shuttle costs half a billion per flight or so, the propellant costs are a few tens of millions. A good rule of thumb is that LEO is $20/kg in propellant.
4) Rockets and elevators would have very different advantages - they could easily work side by side. One caution - do not compare today's rockets to tomorrow's elevators. The technology that will enable space elevators will work wonders on rockets as well.
if people would just get up off their lazy butts and get working. ;-}
The real problems with space elevator-like designs are maintenance. You have a 35,000km long ribbon, in an environment where a 1 cm diameter cable will be cut on average once every 1000 km/years - so you would expect the ribbon to be cut every ten days or so. Even if you come up with a design that you can repair fast enough to keep up, most of the up-mass is repair stuff.
I agree, but I like to compare the Isp of elevators to rockets - I use milli-earths-seconds!
More seriously, another way to think about it (that looks at mass efficiency) is to look at the expected lifetime of the elevator (in kg delivered) divided by the mass of the elevator, and compare that to the expected lifetime of a similarly non-existing RLV made with nanotubes (in kg delivered) divided by the mass of the RLV.
That's what kills the space elevator, really. A nanotube RLV will be pretty darn impressive!
Look closely at what I wrote - I do not disagree with you. I was saying that the same object in the same orbit has the same energy no matter how it got there.
The percentage of energy expended that goes into the final object varies wildly - but the final object energy is the same.
So as to your statement:
reduce it to gravitational potential energy
note that no matter how you get to GEO, you have to expend the same mgh (gravitational potential energy) and 0.5mv^2 (kinetic energy). The only difference is in the energy imparted into your reaction mass. A rocket's reaction mass is the fuel of the rocket - a space elevator's reaction mass is the entire Earth. Conservation of momentum linked with e=0.5mv^2 says that the larger reaction mass is far better.
GEO is 42,000 km radius. Earth's radius is 6400 km. Gravity acceleration goes down with the square of radius, so the acceleration due to gravity drops from 9.8 m/s to 0.2 m/s. Substantially weaker, but not zero - you still need a high velocity to stay in orbit.
In fact, normal low Earth orbital velocity is 7.8 km/s - GEO still needs 3 km/s.
My point on conversion energy is that an object place in GEO by rocket has exactly the same energy as that object delivered by the elevator - the only question is where the energy comes from.
A concern with talking about the efficiency of rockets is that you have to carefully define what you mean: normal chemical rockets have extremely high Carnot efficiencies, mediocre mass and energy efficiencies. A space elevator doesn't have a Carnot efficiency, has terrible mass efficiency, but extremely good energy efficiency. An ion thruster has no Carnot efficiency, has great mass efficiency, and terrible energy efficiency.
And these are all just from the basic physics of the devices - not a whole lot you can do about it really.
Yes - turn the energy into photons, and point them out the back. The photons impart a tiny bit a momentum when they fly away.
Sort of like a solar sail, without the solar.
So much wrong, so little time...
Sorry, most of your post is factually challenged.
1) Space elevators do not lower the energy required - they just use the energy differently.
2) They do not take you to where the gravity is weak - they take you to the point where the force of gravity (which is essentially unchanged) is balanced by centripetal force (which, being linked to w^2r goes up linearly with distance).
3) Rockets typically take you to about 7.7 km/s (orbit), not 11.2km/s (escape).
4) The energy given to the satellite (assuming the same final orbit) is identical regardless of the launch vehicle/elevator used. What is different is the energy efficiency of the system in putting energy into the satellite:
A rocket sends lightweight propellant in the opposite direction very fast in order to transfer the energy. An elevator sends a huge mass (essentially the entire earth) very slowly in the opposite direction. Since momentum is conserved, the mass x velocity of both systems is the same - but since the Earth masses a lot more than most rockets, the Earth's relative velocity is far lower. This is where the e=0.5*m*v^2 comes in - the "wasted" energy is the energy provided to the Earth or propellant. Earth has a small v, big m - which works better than the rockets big v little m.
So you always have to give the satellite the same energy - there are just different efficiencies of giving it that energy. Space cannons have the problem of needing to give that energy extremely quickly... very difficult indeed.
Explain then why the masterminds of 9/11 decided to confess this morning, so that they could be executed? They say that they want to die as martyrs and get their virgins...
I can see the point, really. My point is that, while I agree that what Dijikstra does is not a software engineer, that does not mean that software engineers do not exist. Microsoft was the most successful software company in the world. Let me ask you - was their ability to write bug free code a factor?
No - software engineering is not about producing bug free code. Software engineering is about producing software. Balancing time, money, and equipment against production.
You may not do it, but that doesn't mean others don't.
(And, your right, the Arianne software error was a silly waste - and almost certainly should have been caught. But formal proofs would not have caught it... the implementation error would have required looking at things at the hardware level, not the algorithm level.)
Not yet - we are still very quiet.
The wraps should come off next March or so. I'd give you the company name, but honestly it will most likely change before then (engineers choose working names, marketing chooses launch names).
If it all works right, you shouldn't be able to miss it, though!