Then let anyone broadcast on any power/freqency, anything.
Thus ruining modern astronomy for everyone. Half the benefit of having such a tightly regulated spectrum is the fact that scientific observations can, and do, still go on.
You don't need to open up the full spectrum. It's actually nice having common bands that everyone uses, because the bands are all standardized, and the equipment is commodity.
I love hearing people say that the upper GHz band is ripe for the taking. Damnit. There's a lot of good science that can be done in the GHz band, because it's so quiet.
Isn't it bad enough that cities have massively light-polluted the sky simply because businesses are lazy and stupid? If you open up the spectrum completely, business will do the same. In order to avoid interference, slowly but surely the *entire* spectrum will be utilized. They'll say the links are directional, and for the most part, they will be - but with tons upon tons of transmitters, all the marginal side lobes will simply eventually make it impossible to do any science.
The optical portion of the spectrum is unregulated. What it's led to are insanely bright night skies that cause a lot of health problems for city dwellers. And why? No reason. Light pollution does not make cities safer (casts more shadows, so they are actually less safe than a dimmer city), and it's not even used by any of the people, as the spilling light heads upwards - the biggest light polluters are office windows at night, street lamps that aren't properly directionally downwards, and businesses like gas stations that feel they have to outshine the Sun to get customers. People can't self-regulate the *optical* band. What makes you think they could regulate the *radio* band, where most people can't even see the damage that's being done?
Granted, some people act like they are "ranking a decision tree" or "get stuck in decision deadlocks", but we don't all do that -- in fact, most of us don't. (Those would be the Asperger's Syndrome types and the catatonics respectively.)
Yes, but no matter what, humans act on input - they choose an action based on the input they have.
An intelligence could virtually be defined as a method for choosing between choices based upon the ability to predict the future, even crudely. Many humans suck at that, but they still do it. Objects without a nervous system don't do that, so it's a pretty good definition.
Now, when I say "pruning a decision tree", actually, what I mean is "pruning the number of possibilities that you have to decide from", and humans absolutely do that. In many cases, we're biologically designed to do that, by reducing the sensor inputs. Our brains do heavy pattern recognition on the visual input, and the result is very different than what we see. We do noise reduction on vocal conversations in a crowded room, as well.
That's exactly what a robot would need to do as well. You could presumedly also increase the "intelligence" of a robot by increasing its ability to predictively look farther and farther in the future.
This is also exactly akin to pruning the move tree in chess-solving programs: try to ignore unimportant data, and choose from the important alternatives.
It should also be noted that I'm not really talking about a procedural AI. Since an AI acts on its environment, and that environment acts as the function's input, which changes the AI (by adding to its experience history), it's necessarily a feedback loop, which is not precisely a procedural program. The program is only predictive with the same experience history and input - otherwise, different instances will act quite differently.
You don't know how to make an AI.
No, and no one else does either. You don't 'know" how to do that unless you actually do it, and as no one has actually built an AI, they don't know how to do it. The there are other people who know this stuff better than you... even though they haven't done it yet, either argument doesn't fly well.
Human version: 1) Be nice to each other. Take care of each other. 2) Obey, except if the order is 'unlawful', in that it would involve hurting others. 3) Take care of yourself.
No. Those are "higher-order" ethical laws - laws that society has developed in order for society to function. More importantly, they are absolutely not essential for a person to maintain sanity, as there are plenty of people who would never submit to being ordered around. There may be some basic psychological laws which most humans follow (like not killing your own kind) which prevents people from going nuts, but psychologists are absolutely not smart enough to say that concretely yet.
That's what I'm talking about - what is required to create a stable, sane intelligence? Certainly something, as there are definitely common features between people who have severe psychological problems. It may be that certain ethical laws are required, but they are certainly not restricted to the Three Laws. Now, the Three Laws may be required to allow those robots to function in human society, but that, again, is an area for sociologists to determine, not me.
The website deals with the mile wide gaps in these laws. Let's take it right from the top - Robots as functional as the ones in the film would be very good as soldiers, thus taking that first rule and chucking it right out. In fact, it's the defense industry that would most like robots like the ones in the film.
True, but the question is - can you actually make an artificial intelligence that's stable that are soldiers? It's notoriously hard to do it with natural intelligence, so it's not inconceivable that an AI would have significant stability issues being programmed as a self-sacrificial soldier.
The issue with AIs is that in order to be capable, they have to learn. Presumedly you would start the robot out with a corpus of knowledge, and maybe a physics engine, so that the robot could conceivably identify the future occurances of the world state around it. So, in order for it to remain sane, the world that it encounters has to be similar to the corpus that it's fed, otherwise it would presumedly enter a decision deadlock.
The military would still need robots that follow orders, after all. How would the robot decide whose order to follow, and whose not to follow? Plus what about any ethical decisions it needs to make during combat? The question is - would the robots be capable of pruning that decision tree as easily as they would if they had the Three Laws?
But let's stay on course, and assume these are robots meant as domestic servants. Does the robot take non-lethal contradictory rules and simply process them in order, taking the last order? Two children would amuse themselves for hours telling the robot "pick up that broom", "don't pick up that broom" and keeping the robot in limbo. The robot should tell the children to behave and go pick up their rooms. Directly violating rule 2.
Why would the robot do that? I don't see any problem with the robot being deadlocked receiving constant orders from humans. If a parent really wanted a robot not to get stuck like that, a parent could simply inform the robot that following such orders causes harm to the children (or the parent). Or a domestic servant could be fed a corpus of knowledge about child-raising, and would know that on its own.
How about the running into the burning building scenario? It's unclear that there is anybody in the building left alive to save, or if everyone has escaped or not. Does the robot violate Rule 3 in order to *possibly* meet Rule 1?
Depends how complicated the robot's programming is, and its corpus of knowledge. The robot simply needs to evaluate the situation and the possibilities, and take the path that is most likely to satisfy all Three Laws.
Anyhow, the website has more papers on the subject that examine the issue in a moral framework.
As far as I've found from that website, it's discussion is more on a 'moral' level, which I don't really understand. The point is that any AI is going to need a method to make decisions - that is, a method for evaluating which of the possibilities available to it the robot will take. One problem with any AI will be that the decision tree will be gigantic once the robot is complicated enough to plan even a minute or so in advance, and so it'll need some method for pruning the decision tree and locating viable options to iterate, and those methods are what humans call "ethics" - what choices are bad, and what choices are good? The Three Laws seem like a good starting point.
I've always been interested in creating a proof-of-concept Three Laws Safe 'robot'. Imagine a robot that's just a box with a tower with a retractable plate (to block falling objects), a motion sensor, and a few temperature sensors. The temperature sensors would determine whether or not a human was present, based on body heat. Might need another method for doing this, but face recognition is tough. The motion sensor would determine if something was falling towards the robot, and the robot would take comman
Yes, the biggest hole with the Three Laws is the assumption on which they are based: Somehow, these laws are so fundamental to the functioning of the robot's positronic brain that the robot would essentially have to destroy itself in order to get around the laws. They were "fundamental equations" -- I think that is a quote; certainly I'm paraphrasing a number of passages from the book.
This isn't necessarily crazy. It's unproven, and it's possible that it's untrue, but it's not currently crazy.
We don't know how to make an AI. But obviously an AI will have to be an algorithm that prunes and "ranks" a decision tree to locate what to do, presumedly based on either a physics engine or an experience database.
A learning AI would presumedly store the results of its decisions in its experience database. If its experience database grew far too conflicted and far too confused, the AI could conceivably be unable to do anything - stuck in a decision deadlock.
Moreover, the possibilities that can occur in reality are far too huge to compute every single possibility in any reasonable timeframe. It's entirely possible that you could develop laws that, were the robot to avoid them, it would be impossible for it to prune the decision tree enough for it to work at all. Those laws would then be, for all practical purposes, necessary for the robot to function, even if the laws themselves weren't the only ones that could prune the tree down. Obviously the Three Laws aren't the only way an AI can exist - humans are "biological AI", and we don't have those three laws.
However, one could build a design based around laws which would be fundamental to the design, by doing exactly what I said before.
but the fact that they are currently (and possibly inherently) impossible to implement.
That I significantly doubt. Part of the problem is that we don't know what hard AI is going to look like. That being said, we can guess what it's going to look like.
Imagine a "hard AI" robot as maintaining three states.
1) Personal state. This is the robot's idea of its own "health". Readouts from sensors - power level available, etc.
2) World state. This is the robot's idea of its surrounding environment. Probably a spatial map derived from sensors, external sensor status (temperature, pressure, etc.), and objects derived from spatial data interpretation. Deriving the world state from the sensor data is a very difficult problem, but for this gedankenexperiment, we can consider it solved.
3) Directive state. This is the robot's idea of what it's intending to do, and progress towards those goals.
You could imagine the Three Laws being developed by "seeding" a database, just like many "proto-AI" systems (like Cyc, for instance) try to do. Feed it huge amounts of sensor data and the forward/backward steps in time resulting from that sensor data. This is essentially similar to what a chess-solving algorithm would do!
The robot's "decision tree" could then be whittled down by applying the Three Laws - eliminate all decisions where a Human in the World state comes to harm, eliminate all decisions where the Self state declines, and eliminate all decisions that lower the completion of the active Directives state.
Obviously it would need to be more complicated than that, because of the "through inaction" stuff, though this is not that hard to fix - determine if an action is mandatory by iterating the World state forward in time, assuming no action, and see if a Human's health state declines. If yes, begin brute-forcing decisions to locate a future in which it does not, and add what's found to the Directives state. Same for the Self bit.
How the heck is a robot supposed to accurately judge that whether a random unique action in a unique situation will cause harm to a human or himself?
Doesn't need to do it accurately. It just needs to have a method of determining a likely outcome. Best way to do this is via an experience database, the same way that humans do it. This way the robot could actually deal with illogical actions by humans, by determining likelihoods.
Humans can't even do this.
True, but humans have "imperatives" and "desires" as well, and we don't get deadlocked by not being able to accurately predict the future. We just predict it as well as we can.
In the same way, a chess-playing program can't traverse the entire move tree, so it finds the best solution it can, and goes from there.
we will have the technology to produce usefull robots long before we have the technology to produce 3-Law abiding robots
What we consider robots is not what Asimov considered robots. An Asimov robot has AI, and it's very likely we will have to implement something like the 3 Laws to ensure that the decision tree that an AI traverses is human-safe.
Just going by the dashboard MPG readout, I definately get the best mileage in 6th (somewhere in the 1100-1200 rpm ballpark probabably, where on flat ground I'll get 40-45 MPG at 50 MPH).
Look up the torque curve for your car. Given that it's a Corvette, it most likely has a very flat and very low-onset torque curve. It might even start at 1000 RPM, though that seems really low.
Any car with a low 0-60 time will have a low-onset torque curve. Obviously. That's what allows you to accelerate fast. Different engine designs change the torque curve, and so you can have engines that are most efficient at higher RPMs, or lower RPMs.
That is, of course, the reason that the car is geared that way. On a car with a torque peak at higher RPMs, you'd probably wouldn't bother with more gears anyway.
Mileage even sucks around 3000 RPM. You definately want to be at the lowest RPM possible.
No. Look, gears don't generate additional power, so they can't by themselves lower fuel consumption. A car traveling at 50 mph in 4th gear might require 75 horsepower to maintain speed. It will still need 75 horsepower in 3rd gear, 2nd gear, or even 1st gear.
RPMs are not a measure of how much fuel you're using. They are a measure of how frequently the engine cycles are occuring. If you need 75 horsepower, and your engine is at 5000 rpm, then each cycle of the engine is delivering 250 uHP (microhorsepower). If your engine is at 2500 rpm, then each cycle of the engine is delivering 500 uHP.
In other words, at low RPMs, you're asking the engine to deliver more power per cycle than at higher RPMs. There is, of course, an optimal amount of power per cycle that an engine can deliver - it's the point at which the gas-air mixture is correct. Too high RPMs, and you run with too little gas, too much air - too low RPMs, and it's the opposite - too much gas, too little air. In either of those cases, the power that comes out of burning the gas is less than the power that would come out if you had burned it at the correct mixture.
(This is in "the world's simplest auto engine" - it doesn't include things like a throttle, or any other performance improvements that are designed to flatten the torque peak and smooth things out. But that's the basic reason why lower RPMs do not mean better fuel efficiency.)
It's important to remember that with lower RPMs, you're asking the engine to work harder, but less often. With higher RPMs, you're asking the engine to work less, but more often.
Anyway, if you really doubt it, continue your experiment a bit - push below 1000 rpm. Go crazy. Try 30 in 6th gear, and watch the gas mileage plummet.
there's no way a 4-cylinder 2 liter engine is going over 100!
Oh for crying out loud - stop replying with "well, my X can go faster than..." - I should've put "my" 4-cylinder, with 114 hp. Forgot to put the horsepower down. For a car with a listed maximum speed of ~100 on a flat road, it's very odd to gear the car like that. On the whole, all you do is lower the overall average gas mileage, because a huge portion of the power band of that last gear is unreachable.
The car is not geared to go 150mph, it is geared to have a lower rpm(2000rpm) at a normal speed (70mph)
But that's your highest gear out of 5, and presumedly 2000 rpm is a little after the start of the power band in your car. Presumedly the next-to-highest gear puts ~ 55-60 mph at 2000 rpm to optimize for that band.
The reason that I said it's an odd gear is because of the speeds that it optimizes for, and the speeds that it sacrifices. While I do get highest mileage at 70-75 mph, I get low mileage at 55 mph, because automatics shift too soon. (I get about 25 mpg at 55 mph, and about 35 mpg at 70 mph).
The problem with that is if they had geared the car tighter, then it would get better gas mileage at 65 mph, and not much worse at 70 mph, etc. The only thing it would cost you is lower top speed, but the top speed for this car is limited by the horsepower, so it wouldn't do anything except improve the gas mileage.
For a 5-speed automatic, I can understand it! There's another gear that optimizes for the 50-60 range, and then the highest gear for 60-70 range But for a 4-speed? There should be an extra gear in there. If you plot fuel efficiency vs. speed for this car, there's a noticeable dip around 55-65 mph where the transmission still wants to shift into 4th gear, but the engine isn't efficient enough at that point. The efficiency plot then recovers at about 70-75 before dropping off a little faster than 1/x due to aerodynamics.
It's odd, but I find it kind of amusing. Gives me justification for going 70-75 mph. Have to, for better gas mileage.:)
The engine efficiency drops massively as the RPMs increase to the torque peak
Woah, that was a typo! Meant to say it increases massively! And I keep using "torque peak" and "power band" interchangeably, which is wrong, of course. I mean power band.
specific fuel consumption is usually best around the torque peak in an internal combustion engine.
Barring all other improvements in engine design, yes. But those improvements usually serve to flatten out the torque peak, so most cars have a flat plateau in that torque peak. Since efficiency is pretty constant in that region, and aerodynamic losses are more than linear, you'll get best fuel efficiency somewhere near the lowest point in the power band.
This might have been valid 20 years ago, but it's hardly today. I've happily had my Honda S2000 (4 cyl, 2 liter normally aspriated, making 240HP) up to 140MPH
OK, there's no way my 2.0L is going over 100. It's not a racing car at all in its stock configuration (114 HP) so why they geared it to go over 150, I have no clue. Serves me right for trying to be dramatic, I should've just specified the horsepower.:)
... find out what the most efficiant speed is for my vehicle? I have a 2.7 L V6, so that sucks, but DOHC helps a little bit.
Experiment. There are far too many variables (gearing, aerodynamics, engine design, path of the road that you're taking) to account for.
Don't expect the fuel efficiency to drop off monotonically (that is, don't expect it to constantly drop and not increase again later), especially if you have an automatic.
Ok... then what's an overdrive for? I thought it was to get better mileage.
To go faster and reduce wear on the engine. This is America, after all.
Seriously, though, depending on the gearing, it may improve gas mileage at 55 mph if it's in the power band at 55 mph. However, that'd be a pretty low power band for a car for a normal overdrive - at 65 mph, you're usually close to the power band, and so at around 65 mph, it definitely improves gas mileage. (and be serious - who drives at 55?)
An automatic normally has its best gas mileage around 35-40 mph. Aerodynamics starts to take a lot of power out after that. But an automatic without overdrive, at 65, is usually pretty far up in the power band. The lower you are in the power band, the better your efficiency (slightly). So you'd be better off adding another gear, which is what overdrive is.
You can think of it like this: adding a gear adds a valley and a peak to a fuel efficiency curve. If you place the gearing tight enough, the valley disappears, and you just get a flat fuel efficiency curve (hence continuously variable transmissions). So it can improve gas mileage, but it can also greatly hurt it at some speeds, depending on the gearing.
(Why do I know this? Because the gearing on my Mazda is incredibly weird. The car would be gear-limited in 4th gear at 150 mph or so, but it's only a 2.0L engine! There's a wide gap between 3rd and 4th gear, and that gap corresponds to about 55 mph, where fuel efficiency drops to below 30 mpg, whereas at 70 it's 35 mpg.)
It can't possibly make sense to drive around at 5500 RPM all the time.
Sorry, that's the best idea. Why do you think Geo Metros got 40-50 mpg? Ever been in one? The engine is always in its torque peak. It can barely make it up mountains. Note I'm not suggesting a Metro - I'm just saying that's how it got the fuel efficiency.
(Your car likely has a broad power band, with a very slow torque peak. Probably something like 4000 rpm is where you'd get a peak in gas mileage)
And as I've pointed out elsewhere, this, of course, ages the engine faster. So you're screwed either in gas costs, or maintenance costs.
Yeah, doesn't make much sense to me either. My redline is 8000, torque peak is around 5000-6000, and I get noticeably better gas mileage if I do 55 vs. 65, even though I'm closer to the torque peak at 65.
RPMS are a function of gear, not just of speed. You get better gas mileage because aerodynamics hurt a lot. Change gears to a lower gear, and your gas mileage will go significantly up.
You'd probably experience another peak in gas mileage if you continued going faster (probably around 70). It might or might not improve your gas mileage more than what you were experiencing at 55 or 65 - it depends on the gearing (if you have an automatic).
Imagine a plot that started off very low, and increased (probably about as sqrt(X)), with several broad peaks (each corresponding to a gear) to a peak around 35 or 40 (which is where your car likely gets the best gas mileage), and then started falling (less fast than 1/x), but has a large peak at some point after that (probably in the neighborhood of 70 or 80).
That's the fuel efficiency curve. It's very complicated, and has a lot of features. And no, none of us are neglecting to take into account air resistance. I've had more than enough physics to remind me of that, thank you very much. If I see the Navier-Stokes equation again, I'll throw up.:)
The rolling-resistance loss goes pretty linearly with speed
Rolling resistance goes less than linearly with speed. On a (ideally) flat surface, a (flat) rolling object experiences little to no friction losses (as there's no movement at the contact point - at the bottom of the wheel, where it touches the surface, the wheel is not moving). Rolling resistance comes from imperfections in the surface and the rolling object.
You'd get the same mileage if the losses grew linearly with speed, but they dont.
Depends what you mean by losses. The engine efficiency drops massively as the RPMs increase to the torque peak - far faster than linear. This creates a peak in a fuel efficiency curve. Gearing means that for an automatic, there's a fuel efficiency peak for each gear, but since automatics change around that point, the maximum fuel efficient speed is usually the highest gear's maximum speed. The gearing in my car, for instance, is quite high, and maximizes fuel efficiency (at about 35 mpg) at about 70 mph. (Very strange for a 4-cylinder engine to have an overdrive that does 75 at 3000 RPM, and redlines at 6500! That would imply redlining at 150, and there's no way a 4-cylinder 2 liter engine is going over 100!)
I assume it was air resistance that was making me get poor mileage.
At 130 mph??? Holy bleep yes! At 130 mph probably something like 80-90% of your power output was going to fight the aerodynamics. You want to be at the point where 50% of the power is going to aerodynamics, 50% to rolling resistance. More or less, that's about good.
The point I'm trying to make is that if you then attempt to go 60 mph in 5th, you'll get lower gas mileage than if you go 60 mph in 4th (assuming that in 4th it's in the 3000 rpm range, and the engine is designed with the torque peak). Most people won't believe it, because the engine sounds like it's struggling and it sounds like you're using more fuel.
(Then again, you are wearing the engine faster, so one way or another, you're spending money. The maintenance cost of a car is usually ~ equivalent to the fuel cost, so you're screwed either way.:) )
My recollection is that maximal efficiency is roughly at torque peak (ignoring such things as aerodynamics and gearing), and that underpowering a car kills the mileage.
That's exactly what I said - though actually, efficiency is pretty constant in the power band, so maximum fuel efficiency is at the lowest point in the power band.
(Except for the last part, but that's addressed later...)
Case in point: a particular truck is offered in an economy V6 and a V8 trim. The V8 got better mileage because the V6 was always running full throttle (above the powerband).
Woah, woah - you're talking about two different situations here. Most cars are way overpowered for going at the speed where aerodynamic losses equal total residual losses - this is about 35 or 40 mph for most cars. So when I said underpower the engine, I meant underpower it compared to most cars, not underpower it compared to its needs.
You're exactly correct that a car that's running full throttle will have crap efficiency, but that's because it's past it's torque peak. You want to be at the torque peak, not above it (full throttle) or below it (going slow).
In your case, the aerodynamics and rolling resistance are so high because the weight is so high that the car is now not overpowered to go the speed that's efficient for aerodynamics. The V6 would get better gas mileage than the V8 if it went slower.
Your Geo may get good mileage, but it's crap, and I won't drive one. I have an MR2 that gets 30 MPG and handles nicely, so I don't have to.
I don't own a Geo. It is however a good example of a car that uses standard design principles to get high gas mileage. Small engine, light weight.
It's important to note that MPG has a lot to do with driving style. While my car cannot get 1700 MPG, a bit of predictive driving (i.e. know when to start slowing down, when to build up momentum) will greatly increase the MPG.
Absolutely. My wife gets noticeably lower gas mileage than I do while driving, mainly because of less consistent speeds while driving. (Not a big deal, as she's actually a safer driver than I am anyway, so it all evens out.)
One thing that's a problem is that a lot of people don't actually understand the ways to drive to maximize MPG - drive smoothly, try not to brake as much as possible (just let up off the gas!), and if you've got a manual, make sure you're in the right gear. Most cars are not efficient at low RPMs - cars get best gas efficiency when they're in their power band. So you should strive to make sure that a car's tachometer stays quite high (probably 3000 or higher - try to find the car's torque peak and go a little lower for your own). A car sputtering around town at 500-1000 rpm will get a small fraction of the gas mileage it gets on the highway.
The best way to maximize gas mileage is to pay attention. Experiment - cars are very different. Alter the average speed at which you drive, and record the mileage. Best way to do it.
Going at 15 mph, there's not much safety equipment required.
Fuel efficiency is a difficult thing to deal with - engines have the highest efficiency (power out/fuel in) basically at the minimum point in the power band. Yes: this means that a common engine is getting terrible gas mileage if you're moving along at ~15 mph normally. This is why a car's maximum fuel efficient speed is complicated (and is rarely 55 mph, regardless of what hundreds of websites with terrible math will tell you!) and depends very strongly on the car's gearing. Many cars with overdrive will actually have a "two hump" fuel efficiency curve - that is, they'll be most efficient at about 30 mph or so if you're in 3rd gear, but also have another efficiency peak at 65-70 mph that's lower than the first (but still higher than going 55 mph in the overdrive gear).
The way to get good fuel efficiency with a standard design engine is twofold - make the car light, make the engine underpowered, and go slow. If the engine is always struggling, it's always in the power band, and always efficient. Hence the reason that a Geo Metro gets great gas efficiency.
Note the details of these cars - slow speed (15 mph), massively underpowered engine (3-4 hp), and very light chassis.
(As an aside, most websites are crap at explaning this. See here, where they state that going from 100 kph to 120 kph increases the fuel consumption by 20%. Since you're moving 20% faster, a 20% increased fuel consumption means exactly the same gas mileage.)
Re:Author has "no idea what was responsible for na
on
The History Of Pentium
·
· Score: 1
Author also seems to believe that the P1 went up to 300Mhz
It did, just only in its mobile incarnation. But it was the same core (with MMX stuck on).
but it takes 20 years (and change) to the guy on earth, the guy on earth sees the guy it the spaceship going slower and slower, not faster and faster.
Well, this is true.
Think, McFly, think. It is the guy on Earth living at a hummingbirds pace.
But this is not. To the person on the spaceship, the person on Earth is living extremely extremely slowly as well.
The hummingbird's pace doesn't happen until the person on the spacecraft slows down, turns around, and returns to Earth, or just until he slows down, but even then it's a little more complicated because the two people aren't at the same location.
Gah, Google math sucks. I have no idea what caused it to screw that up. I thought the orbital speed seemed too low... OK, fixing:
Payloads at Earth orbit are moving at 463 m/s. Payloads at GEO are moving at 873 m/s. Delta-V is 410 m/s. Delta-T is 604800 seconds.
To move an 18,143 kg mass from 463 to 873 m/s in 604800 seconds takes 12.29 newtons. In pounds, that's 2.76 pounds of force. Which I could generate by leaning against the cable.
That's a very interesting use of the word "instant." I believe It would take hundreds of years for it to solidify again after becoming a ball of molten rock. Two planets colliding cause a little bit of friction...
Millions of years, more likely!
But as we've got hundreds of millions of years left, as long as we get to it in the next few million years, we should be fine.:)
(I thought it was a more interesting use of the word "Poof".)
Re:Mars GPS = APS, Earth GPS = TPS ?!!
on
GPS on Mars?
·
· Score: 2, Insightful
Now, the parent refers to Mars GPS as Ares positioning System. Does that mean we should rename Earth GPS, TPS (Terra Positioning System)?
GPS could easily be renamed "geographical positioning system", instead of "global". APS should then be "areographical positioning system", using the correct Latin prefixes.
Sadly NASA will probably pick something stupid like MGPS or MPS, because no one likes Latin anymore.
Re:Who cares what a reader asked?
on
GPS on Mars?
·
· Score: 2, Insightful
did he thing that some/.er would be able to glean answers from a crystal ball or tea leaves?
Considering that several people who read Slashdot work for JPL and NASA, I think they might be able to glean the answers from their own brain.
Just might, though. We are talking about JPL here.
Slashdot Mirror
Then let anyone broadcast on any power/freqency, anything.
Thus ruining modern astronomy for everyone. Half the benefit of having such a tightly regulated spectrum is the fact that scientific observations can, and do, still go on.
You don't need to open up the full spectrum. It's actually nice having common bands that everyone uses, because the bands are all standardized, and the equipment is commodity.
I love hearing people say that the upper GHz band is ripe for the taking. Damnit. There's a lot of good science that can be done in the GHz band, because it's so quiet.
Isn't it bad enough that cities have massively light-polluted the sky simply because businesses are lazy and stupid? If you open up the spectrum completely, business will do the same. In order to avoid interference, slowly but surely the *entire* spectrum will be utilized. They'll say the links are directional, and for the most part, they will be - but with tons upon tons of transmitters, all the marginal side lobes will simply eventually make it impossible to do any science.
The optical portion of the spectrum is unregulated. What it's led to are insanely bright night skies that cause a lot of health problems for city dwellers. And why? No reason. Light pollution does not make cities safer (casts more shadows, so they are actually less safe than a dimmer city), and it's not even used by any of the people, as the spilling light heads upwards - the biggest light polluters are office windows at night, street lamps that aren't properly directionally downwards, and businesses like gas stations that feel they have to outshine the Sun to get customers. People can't self-regulate the *optical* band. What makes you think they could regulate the *radio* band, where most people can't even see the damage that's being done?
Granted, some people act like they are "ranking a decision tree" or "get stuck in decision deadlocks", but we don't all do that -- in fact, most of us don't. (Those would be the Asperger's Syndrome types and the catatonics respectively.)
Yes, but no matter what, humans act on input - they choose an action based on the input they have.
An intelligence could virtually be defined as a method for choosing between choices based upon the ability to predict the future, even crudely. Many humans suck at that, but they still do it. Objects without a nervous system don't do that, so it's a pretty good definition.
Now, when I say "pruning a decision tree", actually, what I mean is "pruning the number of possibilities that you have to decide from", and humans absolutely do that. In many cases, we're biologically designed to do that, by reducing the sensor inputs. Our brains do heavy pattern recognition on the visual input, and the result is very different than what we see. We do noise reduction on vocal conversations in a crowded room, as well.
That's exactly what a robot would need to do as well. You could presumedly also increase the "intelligence" of a robot by increasing its ability to predictively look farther and farther in the future.
This is also exactly akin to pruning the move tree in chess-solving programs: try to ignore unimportant data, and choose from the important alternatives.
It should also be noted that I'm not really talking about a procedural AI. Since an AI acts on its environment, and that environment acts as the function's input, which changes the AI (by adding to its experience history), it's necessarily a feedback loop, which is not precisely a procedural program. The program is only predictive with the same experience history and input - otherwise, different instances will act quite differently.
You don't know how to make an AI.
No, and no one else does either. You don't 'know" how to do that unless you actually do it, and as no one has actually built an AI, they don't know how to do it. The there are other people who know this stuff better than you... even though they haven't done it yet, either argument doesn't fly well.
Human version:
1) Be nice to each other. Take care of each other.
2) Obey, except if the order is 'unlawful', in that it would involve hurting others.
3) Take care of yourself.
No. Those are "higher-order" ethical laws - laws that society has developed in order for society to function. More importantly, they are absolutely not essential for a person to maintain sanity, as there are plenty of people who would never submit to being ordered around. There may be some basic psychological laws which most humans follow (like not killing your own kind) which prevents people from going nuts, but psychologists are absolutely not smart enough to say that concretely yet.
That's what I'm talking about - what is required to create a stable, sane intelligence? Certainly something, as there are definitely common features between people who have severe psychological problems. It may be that certain ethical laws are required, but they are certainly not restricted to the Three Laws. Now, the Three Laws may be required to allow those robots to function in human society, but that, again, is an area for sociologists to determine, not me.
The website deals with the mile wide gaps in these laws. Let's take it right from the top - Robots as functional as the ones in the film would be very good as soldiers, thus taking that first rule and chucking it right out. In fact, it's the defense industry that would most like robots like the ones in the film.
True, but the question is - can you actually make an artificial intelligence that's stable that are soldiers? It's notoriously hard to do it with natural intelligence, so it's not inconceivable that an AI would have significant stability issues being programmed as a self-sacrificial soldier.
The issue with AIs is that in order to be capable, they have to learn. Presumedly you would start the robot out with a corpus of knowledge, and maybe a physics engine, so that the robot could conceivably identify the future occurances of the world state around it. So, in order for it to remain sane, the world that it encounters has to be similar to the corpus that it's fed, otherwise it would presumedly enter a decision deadlock.
The military would still need robots that follow orders, after all. How would the robot decide whose order to follow, and whose not to follow? Plus what about any ethical decisions it needs to make during combat? The question is - would the robots be capable of pruning that decision tree as easily as they would if they had the Three Laws?
But let's stay on course, and assume these are robots meant as domestic servants. Does the robot take non-lethal contradictory rules and simply process them in order, taking the last order? Two children would amuse themselves for hours telling the robot "pick up that broom", "don't pick up that broom" and keeping the robot in limbo. The robot should tell the children to behave and go pick up their rooms. Directly violating rule 2.
Why would the robot do that? I don't see any problem with the robot being deadlocked receiving constant orders from humans. If a parent really wanted a robot not to get stuck like that, a parent could simply inform the robot that following such orders causes harm to the children (or the parent). Or a domestic servant could be fed a corpus of knowledge about child-raising, and would know that on its own.
How about the running into the burning building scenario? It's unclear that there is anybody in the building left alive to save, or if everyone has escaped or not. Does the robot violate Rule 3 in order to *possibly* meet Rule 1?
Depends how complicated the robot's programming is, and its corpus of knowledge. The robot simply needs to evaluate the situation and the possibilities, and take the path that is most likely to satisfy all Three Laws.
Anyhow, the website has more papers on the subject that examine the issue in a moral framework.
As far as I've found from that website, it's discussion is more on a 'moral' level, which I don't really understand. The point is that any AI is going to need a method to make decisions - that is, a method for evaluating which of the possibilities available to it the robot will take. One problem with any AI will be that the decision tree will be gigantic once the robot is complicated enough to plan even a minute or so in advance, and so it'll need some method for pruning the decision tree and locating viable options to iterate, and those methods are what humans call "ethics" - what choices are bad, and what choices are good? The Three Laws seem like a good starting point.
I've always been interested in creating a proof-of-concept Three Laws Safe 'robot'. Imagine a robot that's just a box with a tower with a retractable plate (to block falling objects), a motion sensor, and a few temperature sensors. The temperature sensors would determine whether or not a human was present, based on body heat. Might need another method for doing this, but face recognition is tough. The motion sensor would determine if something was falling towards the robot, and the robot would take comman
Yes, the biggest hole with the Three Laws is the assumption on which they are based: Somehow, these laws are so fundamental to the functioning of the robot's positronic brain that the robot would essentially have to destroy itself in order to get around the laws. They were "fundamental equations" -- I think that is a quote; certainly I'm paraphrasing a number of passages from the book.
This isn't necessarily crazy. It's unproven, and it's possible that it's untrue, but it's not currently crazy.
We don't know how to make an AI. But obviously an AI will have to be an algorithm that prunes and "ranks" a decision tree to locate what to do, presumedly based on either a physics engine or an experience database.
A learning AI would presumedly store the results of its decisions in its experience database. If its experience database grew far too conflicted and far too confused, the AI could conceivably be unable to do anything - stuck in a decision deadlock.
Moreover, the possibilities that can occur in reality are far too huge to compute every single possibility in any reasonable timeframe. It's entirely possible that you could develop laws that, were the robot to avoid them, it would be impossible for it to prune the decision tree enough for it to work at all. Those laws would then be, for all practical purposes, necessary for the robot to function, even if the laws themselves weren't the only ones that could prune the tree down. Obviously the Three Laws aren't the only way an AI can exist - humans are "biological AI", and we don't have those three laws.
However, one could build a design based around laws which would be fundamental to the design, by doing exactly what I said before.
but the fact that they are currently (and possibly inherently) impossible to implement.
That I significantly doubt. Part of the problem is that we don't know what hard AI is going to look like. That being said, we can guess what it's going to look like.
Imagine a "hard AI" robot as maintaining three states.
1) Personal state. This is the robot's idea of its own "health". Readouts from sensors - power level available, etc.
2) World state. This is the robot's idea of its surrounding environment. Probably a spatial map derived from sensors, external sensor status (temperature, pressure, etc.), and objects derived from spatial data interpretation. Deriving the world state from the sensor data is a very difficult problem, but for this gedankenexperiment, we can consider it solved.
3) Directive state. This is the robot's idea of what it's intending to do, and progress towards those goals.
You could imagine the Three Laws being developed by "seeding" a database, just like many "proto-AI" systems (like Cyc, for instance) try to do. Feed it huge amounts of sensor data and the forward/backward steps in time resulting from that sensor data. This is essentially similar to what a chess-solving algorithm would do!
The robot's "decision tree" could then be whittled down by applying the Three Laws - eliminate all decisions where a Human in the World state comes to harm, eliminate all decisions where the Self state declines, and eliminate all decisions that lower the completion of the active Directives state.
Obviously it would need to be more complicated than that, because of the "through inaction" stuff, though this is not that hard to fix - determine if an action is mandatory by iterating the World state forward in time, assuming no action, and see if a Human's health state declines. If yes, begin brute-forcing decisions to locate a future in which it does not, and add what's found to the Directives state. Same for the Self bit.
How the heck is a robot supposed to accurately judge that whether a random unique action in a unique situation will cause harm to a human or himself?
Doesn't need to do it accurately. It just needs to have a method of determining a likely outcome. Best way to do this is via an experience database, the same way that humans do it. This way the robot could actually deal with illogical actions by humans, by determining likelihoods.
Humans can't even do this.
True, but humans have "imperatives" and "desires" as well, and we don't get deadlocked by not being able to accurately predict the future. We just predict it as well as we can.
In the same way, a chess-playing program can't traverse the entire move tree, so it finds the best solution it can, and goes from there.
we will have the technology to produce usefull robots long before we have the technology to produce 3-Law abiding robots
What we consider robots is not what Asimov considered robots. An Asimov robot has AI, and it's very likely we will have to implement something like the 3 Laws to ensure that the decision tree that an AI traverses is human-safe.
Just going by the dashboard MPG readout, I definately get the best mileage in 6th (somewhere in the 1100-1200 rpm ballpark probabably, where on flat ground I'll get 40-45 MPG at 50 MPH).
Look up the torque curve for your car. Given that it's a Corvette, it most likely has a very flat and very low-onset torque curve. It might even start at 1000 RPM, though that seems really low.
Any car with a low 0-60 time will have a low-onset torque curve. Obviously. That's what allows you to accelerate fast. Different engine designs change the torque curve, and so you can have engines that are most efficient at higher RPMs, or lower RPMs.
That is, of course, the reason that the car is geared that way. On a car with a torque peak at higher RPMs, you'd probably wouldn't bother with more gears anyway.
Mileage even sucks around 3000 RPM. You definately want to be at the lowest RPM possible.
No. Look, gears don't generate additional power, so they can't by themselves lower fuel consumption. A car traveling at 50 mph in 4th gear might require 75 horsepower to maintain speed. It will still need 75 horsepower in 3rd gear, 2nd gear, or even 1st gear.
RPMs are not a measure of how much fuel you're using. They are a measure of how frequently the engine cycles are occuring. If you need 75 horsepower, and your engine is at 5000 rpm, then each cycle of the engine is delivering 250 uHP (microhorsepower). If your engine is at 2500 rpm, then each cycle of the engine is delivering 500 uHP.
In other words, at low RPMs, you're asking the engine to deliver more power per cycle than at higher RPMs. There is, of course, an optimal amount of power per cycle that an engine can deliver - it's the point at which the gas-air mixture is correct. Too high RPMs, and you run with too little gas, too much air - too low RPMs, and it's the opposite - too much gas, too little air. In either of those cases, the power that comes out of burning the gas is less than the power that would come out if you had burned it at the correct mixture.
(This is in "the world's simplest auto engine" - it doesn't include things like a throttle, or any other performance improvements that are designed to flatten the torque peak and smooth things out. But that's the basic reason why lower RPMs do not mean better fuel efficiency.)
It's important to remember that with lower RPMs, you're asking the engine to work harder, but less often. With higher RPMs, you're asking the engine to work less, but more often.
Anyway, if you really doubt it, continue your experiment a bit - push below 1000 rpm. Go crazy. Try 30 in 6th gear, and watch the gas mileage plummet.
there's no way a 4-cylinder 2 liter engine is going over 100!
Oh for crying out loud - stop replying with "well, my X can go faster than..." - I should've put "my" 4-cylinder, with 114 hp. Forgot to put the horsepower down. For a car with a listed maximum speed of ~100 on a flat road, it's very odd to gear the car like that. On the whole, all you do is lower the overall average gas mileage, because a huge portion of the power band of that last gear is unreachable.
The car is not geared to go 150mph, it is geared to have a lower rpm(2000rpm) at a normal speed (70mph)
:)
But that's your highest gear out of 5, and presumedly 2000 rpm is a little after the start of the power band in your car. Presumedly the next-to-highest gear puts ~ 55-60 mph at 2000 rpm to optimize for that band.
The reason that I said it's an odd gear is because of the speeds that it optimizes for, and the speeds that it sacrifices. While I do get highest mileage at 70-75 mph, I get low mileage at 55 mph, because automatics shift too soon. (I get about 25 mpg at 55 mph, and about 35 mpg at 70 mph).
The problem with that is if they had geared the car tighter, then it would get better gas mileage at 65 mph, and not much worse at 70 mph, etc. The only thing it would cost you is lower top speed, but the top speed for this car is limited by the horsepower, so it wouldn't do anything except improve the gas mileage.
For a 5-speed automatic, I can understand it! There's another gear that optimizes for the 50-60 range, and then the highest gear for 60-70 range But for a 4-speed? There should be an extra gear in there. If you plot fuel efficiency vs. speed for this car, there's a noticeable dip around 55-65 mph where the transmission still wants to shift into 4th gear, but the engine isn't efficient enough at that point. The efficiency plot then recovers at about 70-75 before dropping off a little faster than 1/x due to aerodynamics.
It's odd, but I find it kind of amusing. Gives me justification for going 70-75 mph. Have to, for better gas mileage.
The engine efficiency drops massively as the RPMs increase to the torque peak
:)
Woah, that was a typo! Meant to say it increases massively! And I keep using "torque peak" and "power band" interchangeably, which is wrong, of course. I mean power band.
specific fuel consumption is usually best around the torque peak in an internal combustion engine.
Barring all other improvements in engine design, yes. But those improvements usually serve to flatten out the torque peak, so most cars have a flat plateau in that torque peak. Since efficiency is pretty constant in that region, and aerodynamic losses are more than linear, you'll get best fuel efficiency somewhere near the lowest point in the power band.
This might have been valid 20 years ago, but it's hardly today. I've happily had my Honda S2000 (4 cyl, 2 liter normally aspriated, making 240HP) up to 140MPH
OK, there's no way my 2.0L is going over 100. It's not a racing car at all in its stock configuration (114 HP) so why they geared it to go over 150, I have no clue. Serves me right for trying to be dramatic, I should've just specified the horsepower.
I beg to differ. I regularly drove my 1.7 litre Volvo at around 110mph.
:)
Oh, for crying out loud. OK, I meant my 4-cylinder 2 liter engine is never going over 100. The car's too heavy.
In any case, that car's not going 150, so why in hell it's geared to redline at 150, I have no clue.
I need a sig that says "Any and all mistakes in the above post are for dramatic effect only."
... find out what the most efficiant speed is for my vehicle? I have a 2.7 L V6, so that sucks, but DOHC helps a little bit.
Experiment. There are far too many variables (gearing, aerodynamics, engine design, path of the road that you're taking) to account for.
Don't expect the fuel efficiency to drop off monotonically (that is, don't expect it to constantly drop and not increase again later), especially if you have an automatic.
Ok... then what's an overdrive for? I thought it was to get better mileage.
To go faster and reduce wear on the engine. This is America, after all.
Seriously, though, depending on the gearing, it may improve gas mileage at 55 mph if it's in the power band at 55 mph. However, that'd be a pretty low power band for a car for a normal overdrive - at 65 mph, you're usually close to the power band, and so at around 65 mph, it definitely improves gas mileage. (and be serious - who drives at 55?)
An automatic normally has its best gas mileage around 35-40 mph. Aerodynamics starts to take a lot of power out after that. But an automatic without overdrive, at 65, is usually pretty far up in the power band. The lower you are in the power band, the better your efficiency (slightly). So you'd be better off adding another gear, which is what overdrive is.
You can think of it like this: adding a gear adds a valley and a peak to a fuel efficiency curve. If you place the gearing tight enough, the valley disappears, and you just get a flat fuel efficiency curve (hence continuously variable transmissions). So it can improve gas mileage, but it can also greatly hurt it at some speeds, depending on the gearing.
(Why do I know this? Because the gearing on my Mazda is incredibly weird. The car would be gear-limited in 4th gear at 150 mph or so, but it's only a 2.0L engine! There's a wide gap between 3rd and 4th gear, and that gap corresponds to about 55 mph, where fuel efficiency drops to below 30 mpg, whereas at 70 it's 35 mpg.)
It can't possibly make sense to drive around at 5500 RPM all the time.
Sorry, that's the best idea. Why do you think Geo Metros got 40-50 mpg? Ever been in one? The engine is always in its torque peak. It can barely make it up mountains. Note I'm not suggesting a Metro - I'm just saying that's how it got the fuel efficiency.
(Your car likely has a broad power band, with a very slow torque peak. Probably something like 4000 rpm is where you'd get a peak in gas mileage)
And as I've pointed out elsewhere, this, of course, ages the engine faster. So you're screwed either in gas costs, or maintenance costs.
Yeah, doesn't make much sense to me either. My redline is 8000, torque peak is around 5000-6000, and I get noticeably better gas mileage if I do 55 vs. 65, even though I'm closer to the torque peak at 65.
RPMS are a function of gear, not just of speed. You get better gas mileage because aerodynamics hurt a lot. Change gears to a lower gear, and your gas mileage will go significantly up.
You'd probably experience another peak in gas mileage if you continued going faster (probably around 70). It might or might not improve your gas mileage more than what you were experiencing at 55 or 65 - it depends on the gearing (if you have an automatic).
Imagine a plot that started off very low, and increased (probably about as sqrt(X)), with several broad peaks (each corresponding to a gear) to a peak around 35 or 40 (which is where your car likely gets the best gas mileage), and then started falling (less fast than 1/x), but has a large peak at some point after that (probably in the neighborhood of 70 or 80).
That's the fuel efficiency curve. It's very complicated, and has a lot of features. And no, none of us are neglecting to take into account air resistance. I've had more than enough physics to remind me of that, thank you very much. If I see the Navier-Stokes equation again, I'll throw up.
The rolling-resistance loss goes pretty linearly with speed
Rolling resistance goes less than linearly with speed. On a (ideally) flat surface, a (flat) rolling object experiences little to no friction losses (as there's no movement at the contact point - at the bottom of the wheel, where it touches the surface, the wheel is not moving). Rolling resistance comes from imperfections in the surface and the rolling object.
You'd get the same mileage if the losses grew linearly with speed, but they dont.
Depends what you mean by losses. The engine efficiency drops massively as the RPMs increase to the torque peak - far faster than linear. This creates a peak in a fuel efficiency curve. Gearing means that for an automatic, there's a fuel efficiency peak for each gear, but since automatics change around that point, the maximum fuel efficient speed is usually the highest gear's maximum speed. The gearing in my car, for instance, is quite high, and maximizes fuel efficiency (at about 35 mpg) at about 70 mph. (Very strange for a 4-cylinder engine to have an overdrive that does 75 at 3000 RPM, and redlines at 6500! That would imply redlining at 150, and there's no way a 4-cylinder 2 liter engine is going over 100!)
I assume it was air resistance that was making me get poor mileage.
:) )
At 130 mph??? Holy bleep yes! At 130 mph probably something like 80-90% of your power output was going to fight the aerodynamics. You want to be at the point where 50% of the power is going to aerodynamics, 50% to rolling resistance. More or less, that's about good.
The point I'm trying to make is that if you then attempt to go 60 mph in 5th, you'll get lower gas mileage than if you go 60 mph in 4th (assuming that in 4th it's in the 3000 rpm range, and the engine is designed with the torque peak). Most people won't believe it, because the engine sounds like it's struggling and it sounds like you're using more fuel.
(Then again, you are wearing the engine faster, so one way or another, you're spending money. The maintenance cost of a car is usually ~ equivalent to the fuel cost, so you're screwed either way.
My recollection is that maximal efficiency is roughly at torque peak (ignoring such things as aerodynamics and gearing), and that underpowering a car kills the mileage.
That's exactly what I said - though actually, efficiency is pretty constant in the power band, so maximum fuel efficiency is at the lowest point in the power band.
(Except for the last part, but that's addressed later...)
Case in point: a particular truck is offered in an economy V6 and a V8 trim. The V8 got better mileage because the V6 was always running full throttle (above the powerband).
Woah, woah - you're talking about two different situations here. Most cars are way overpowered for going at the speed where aerodynamic losses equal total residual losses - this is about 35 or 40 mph for most cars. So when I said underpower the engine, I meant underpower it compared to most cars, not underpower it compared to its needs.
You're exactly correct that a car that's running full throttle will have crap efficiency, but that's because it's past it's torque peak. You want to be at the torque peak, not above it (full throttle) or below it (going slow).
In your case, the aerodynamics and rolling resistance are so high because the weight is so high that the car is now not overpowered to go the speed that's efficient for aerodynamics. The V6 would get better gas mileage than the V8 if it went slower.
Your Geo may get good mileage, but it's crap, and I won't drive one. I have an MR2 that gets 30 MPG and handles nicely, so I don't have to.
I don't own a Geo. It is however a good example of a car that uses standard design principles to get high gas mileage. Small engine, light weight.
It's important to note that MPG has a lot to do with driving style. While my car cannot get 1700 MPG, a bit of predictive driving (i.e. know when to start slowing down, when to build up momentum) will greatly increase the MPG.
Absolutely. My wife gets noticeably lower gas mileage than I do while driving, mainly because of less consistent speeds while driving. (Not a big deal, as she's actually a safer driver than I am anyway, so it all evens out.)
One thing that's a problem is that a lot of people don't actually understand the ways to drive to maximize MPG - drive smoothly, try not to brake as much as possible (just let up off the gas!), and if you've got a manual, make sure you're in the right gear. Most cars are not efficient at low RPMs - cars get best gas efficiency when they're in their power band. So you should strive to make sure that a car's tachometer stays quite high (probably 3000 or higher - try to find the car's torque peak and go a little lower for your own). A car sputtering around town at 500-1000 rpm will get a small fraction of the gas mileage it gets on the highway.
The best way to maximize gas mileage is to pay attention. Experiment - cars are very different. Alter the average speed at which you drive, and record the mileage. Best way to do it.
Going at 15 mph, there's not much safety equipment required.
Fuel efficiency is a difficult thing to deal with - engines have the highest efficiency (power out/fuel in) basically at the minimum point in the power band. Yes: this means that a common engine is getting terrible gas mileage if you're moving along at ~15 mph normally. This is why a car's maximum fuel efficient speed is complicated (and is rarely 55 mph, regardless of what hundreds of websites with terrible math will tell you!) and depends very strongly on the car's gearing. Many cars with overdrive will actually have a "two hump" fuel efficiency curve - that is, they'll be most efficient at about 30 mph or so if you're in 3rd gear, but also have another efficiency peak at 65-70 mph that's lower than the first (but still higher than going 55 mph in the overdrive gear).
The way to get good fuel efficiency with a standard design engine is twofold - make the car light, make the engine underpowered, and go slow. If the engine is always struggling, it's always in the power band, and always efficient. Hence the reason that a Geo Metro gets great gas efficiency.
Note the details of these cars - slow speed (15 mph), massively underpowered engine (3-4 hp), and very light chassis.
Here is a very good explanation.
(As an aside, most websites are crap at explaning this. See here, where they state that going from 100 kph to 120 kph increases the fuel consumption by 20%. Since you're moving 20% faster, a 20% increased fuel consumption means exactly the same gas mileage.)
Author also seems to believe that the P1 went up to 300Mhz
It did, just only in its mobile incarnation. But it was the same core (with MMX stuck on).
but it takes 20 years (and change) to the guy on earth, the guy on earth sees the guy it the spaceship going slower and slower, not faster and faster.
Well, this is true.
Think, McFly, think. It is the guy on Earth living at a hummingbirds pace.
But this is not. To the person on the spaceship, the person on Earth is living extremely extremely slowly as well.
The hummingbird's pace doesn't happen until the person on the spacecraft slows down, turns around, and returns to Earth, or just until he slows down, but even then it's a little more complicated because the two people aren't at the same location.
(Hence the twin paradox.)
Gah, Google math sucks. I have no idea what caused it to screw that up. I thought the orbital speed seemed too low... OK, fixing:
Payloads at Earth orbit are moving at 463 m/s.
Payloads at GEO are moving at 873 m/s.
Delta-V is 410 m/s.
Delta-T is 604800 seconds.
To move an 18,143 kg mass from 463 to 873 m/s in 604800 seconds takes 12.29 newtons. In pounds, that's 2.76 pounds of force. Which I could generate by leaning against the cable.
That's a very interesting use of the word "instant." I believe It would take hundreds of years for it to solidify again after becoming a ball of molten rock. Two planets colliding cause a little bit of friction...
Millions of years, more likely!
But as we've got hundreds of millions of years left, as long as we get to it in the next few million years, we should be fine.
(I thought it was a more interesting use of the word "Poof".)
Now, the parent refers to Mars GPS as Ares positioning System. Does that mean we should rename Earth GPS, TPS (Terra Positioning System)?
GPS could easily be renamed "geographical positioning system", instead of "global". APS should then be "areographical positioning system", using the correct Latin prefixes.
Sadly NASA will probably pick something stupid like MGPS or MPS, because no one likes Latin anymore.
did he thing that some /.er would be able to glean answers from a crystal ball or tea leaves?
Considering that several people who read Slashdot work for JPL and NASA, I think they might be able to glean the answers from their own brain.
Just might, though. We are talking about JPL here.
(Kidding!)