We know the theory has flaws. If we don't find the flaws, we can't fix the theory. We are NOT in the situation where we think our theory perfectly describes the physical world and are just looking for proof - we know it breaks down. We just don't know why.
But we know we aren't right. We cannot correct our flawed models of the natural world until we find the flaws in them. We know that our models are wrong, but we don't know why. The whole point of building equipment like this is to find out where our models break down so that we can build better models. If we spend billions of dollars only to learn that we were right (up to the resolution of the instrument), then we wasted billions of dollars and need to build a better instrument.
The best GPS on the market cant even figure out the best way around traffic conditions when given a link to my cities live traffic monitoring system.
Try Waze.
Further more, do you know how much work would go into developing a map that would have potential alternate landing zones
I'm not seeing that as anywhere near the level of effort it would take to make the plane fly itself. Basically have a bunch of people looking at pictures classifying things as farmland, rivers, etc. Census maps with population densities, existing road maps, etc. Not cheap, but probably a lot cheaper than paying 2 highly-skilled guys to sit in a cockpit.
I don't think anyone is suggesting that you could just stick a computer in an existing A320 and fly it as-is. There would quite obviously need to be more sensory equipment.
No, certainly you are not THAT smart if you think that this is the only needed input and that you can set its value well in advance.
Apparently I'm in good company. I never said it would be the only input, only that it is a necessary one.
What if the promised 70% automation error ends up being 85% (oh, who could have expected that!) but the companies still save the half a million in salaries and the extra cargo space?
Are you asking for my opinion? I'd force them to be as safe as prior to the change. Obviously safety is a relative measure. We already make compromises for the sake of costs, so I don't really see this as a new issue.
No, the one that needed to be smart enough is the programer,
I'm not THAT smart, but even I know that the auto-flown plane would need to be programmed to ditch into the least populated areas possible. It's one thing to kill everyone on the plane, quite another to fly into a populated area.
Taking off the pilot from the cockpit is as of now not in the slightest about increasing security but about how higher a dead toll looks acceptable for better financial results.
Without some data to consider, it could very well be both. If the 80% pilot error can be brought down to a 70% automation error, that's still a safety improvement. It does not need to be perfect - just better than humans. If it also lets you save half a million in salaries and use more of the plane for paying passengers, then yay.
If they don't understand what Snowden is telling them well... their fault.
Their fault, but my problem. It's in my self-interest to make sure people are not ignorant, because those ignorant jackasses have a vote that is equal to mine. The only way to recruit people away from "Team R" or "Team D" is to make them into educated voters.
Heh, I appreciate the snark, but I really do think they would have to program ditch sites along the route. At the very least, the plane needs to crash into places that will cause the fewest ground fatalities. It's not like this data is hard to gather compared to the task of building an automatic airplane.
Still, I'd rather have those decisions carefully weighted ahead of time by people who understand the system inside and out rather than depending on a single human being to "roll the dice". Even if Sullenberger was right in all of his decisions (and who could argue with him?), not every pilot is going to be capable of such rational thought under duress. Every computer is going to do exactly the same thing.
I understand it would be a hard problem - otherwise we'd all be automated already. It just seems possible to me. I mean, we're talking about a "when all else fails" scenario. You certainly aren't going to program it to maintain current course - you are going to want to crash the plane away from inhabited areas if possible. Better to risk a collision with a ship than to plow into the most densely populated part of the US.
Again, not my field so I'm coming at this very naively.
Wouldn't an automated system have "ditch" sites programmed in? It wouldn't need to be smart enough to identify a river - just know that the river is there.
And while I must confess to not knowing what is on the checklist, couldn't the computer run through it in a few seconds? Or more probably, run things in parallel that a human can only do in serial? Is there something about the situation that would have prevented the computer from firing up the APU?
Why wouldn't it go: "no runways in range, consult chart for alternate landing zones"? I'm thinking that the Hudson (or any other large body of water) would be on the alternates chart.
Wouldn't an automated plane have a different sensor suite? I don't think we are talking about inflating the automatic pilot from Airplane! and calling it a day.
But there are standard procedures to follow when the sensors give bad results, which were apparently not followed by the pilots on 447. I disagree that these would be more confusing to the remote pilot. The autopilot was smart enough to recognize a problem and disconnect. If there was nothing but an autopilot, it could be set to a fail safe mode. A typical smartphone is probably better equipped than a human when it comes to flying without visual reference points.
Resources may not make enough jobs to keep you happy, but it does provide jobs - and good ones at that. Look at the boon to the Texas economy from fracking.
Lenovo may not be the best example for you to use, since Thinkpads were already manufactured overseas. But I take your point - sometimes a purchase is just the brand name, and the jobs go up in a puff of smoke. But to be honest, those cases are usually fire sales where the company is on its last legs anyway. Think Palm or AMC.
I agree that my posts are radically simplified. They spent over an hour talking about this on a recent NPR show, and they barely scratched the surface. Nevertheless, I stand by my narrative that owing the Chinese government over a trillion dollars is not as incredibly horrible as it would seem at first blush.
Resources == Jobs. Companies are nothing but collections of people. Even real estate is an almost infinite resource in the US. I'm not going to say that running a constant deficit is a good thing, but we could do far worse than to have another country lend us a bunch of our own currency at a rate below inflation.
Yes, exactly. And on top of it all, ultimately dollars are only good for purchasing US goods and services. So at some point - admittedly it could be far in the future - those dollars will presumably help support American jobs.
I don't follow. How would a built-in HTML library have taken away from Netscape as a standalone web browser? I could see it taking away from some theoretical Netscape libraries meant for application development, but if such a product existed it was not very prevalent.
Or are you saying that competing web browsers were using the IE libraries? That may have been the case, but my recollection of the time was that everyone was using home-grown code.
That's a semantic argument. There are applications that depend on the libraries used by IE in order to render HTML. But nothing about Windows requires IE be your web browser. Most OSes have a built-in HTML rendering library these days.
To put it another way, a built-in library to render HTML from within applications would not have killed Netscape.
Funny that you mention "repo men" since the car in the summary, down to the license plate, is from the opening scene in Repo Man. Don't look in the trunk.
Slashdot Mirror
We know the theory has flaws. If we don't find the flaws, we can't fix the theory. We are NOT in the situation where we think our theory perfectly describes the physical world and are just looking for proof - we know it breaks down. We just don't know why.
But we know we aren't right. We cannot correct our flawed models of the natural world until we find the flaws in them. We know that our models are wrong, but we don't know why. The whole point of building equipment like this is to find out where our models break down so that we can build better models. If we spend billions of dollars only to learn that we were right (up to the resolution of the instrument), then we wasted billions of dollars and need to build a better instrument.
The best GPS on the market cant even figure out the best way around traffic conditions when given a link to my cities live traffic monitoring system.
Try Waze.
Further more, do you know how much work would go into developing a map that would have potential alternate landing zones
I'm not seeing that as anywhere near the level of effort it would take to make the plane fly itself. Basically have a bunch of people looking at pictures classifying things as farmland, rivers, etc. Census maps with population densities, existing road maps, etc. Not cheap, but probably a lot cheaper than paying 2 highly-skilled guys to sit in a cockpit.
I don't think anyone is suggesting that you could just stick a computer in an existing A320 and fly it as-is. There would quite obviously need to be more sensory equipment.
No, certainly you are not THAT smart if you think that this is the only needed input and that you can set its value well in advance.
Apparently I'm in good company. I never said it would be the only input, only that it is a necessary one.
What if the promised 70% automation error ends up being 85% (oh, who could have expected that!) but the companies still save the half a million in salaries and the extra cargo space?
Are you asking for my opinion? I'd force them to be as safe as prior to the change. Obviously safety is a relative measure. We already make compromises for the sake of costs, so I don't really see this as a new issue.
No, the one that needed to be smart enough is the programer,
I'm not THAT smart, but even I know that the auto-flown plane would need to be programmed to ditch into the least populated areas possible. It's one thing to kill everyone on the plane, quite another to fly into a populated area.
Taking off the pilot from the cockpit is as of now not in the slightest about increasing security but about how higher a dead toll looks acceptable for better financial results.
Without some data to consider, it could very well be both. If the 80% pilot error can be brought down to a 70% automation error, that's still a safety improvement. It does not need to be perfect - just better than humans. If it also lets you save half a million in salaries and use more of the plane for paying passengers, then yay.
If they don't understand what Snowden is telling them well... their fault.
Their fault, but my problem. It's in my self-interest to make sure people are not ignorant, because those ignorant jackasses have a vote that is equal to mine. The only way to recruit people away from "Team R" or "Team D" is to make them into educated voters.
Heh, I appreciate the snark, but I really do think they would have to program ditch sites along the route. At the very least, the plane needs to crash into places that will cause the fewest ground fatalities. It's not like this data is hard to gather compared to the task of building an automatic airplane.
Still, I'd rather have those decisions carefully weighted ahead of time by people who understand the system inside and out rather than depending on a single human being to "roll the dice". Even if Sullenberger was right in all of his decisions (and who could argue with him?), not every pilot is going to be capable of such rational thought under duress. Every computer is going to do exactly the same thing.
I understand it would be a hard problem - otherwise we'd all be automated already. It just seems possible to me. I mean, we're talking about a "when all else fails" scenario. You certainly aren't going to program it to maintain current course - you are going to want to crash the plane away from inhabited areas if possible. Better to risk a collision with a ship than to plow into the most densely populated part of the US.
Presumably - without pilots - this would have to change for even routine runway operations.
Again, not my field so I'm coming at this very naively.
Wouldn't an automated system have "ditch" sites programmed in? It wouldn't need to be smart enough to identify a river - just know that the river is there.
And while I must confess to not knowing what is on the checklist, couldn't the computer run through it in a few seconds? Or more probably, run things in parallel that a human can only do in serial? Is there something about the situation that would have prevented the computer from firing up the APU?
Why wouldn't it go: "no runways in range, consult chart for alternate landing zones"? I'm thinking that the Hudson (or any other large body of water) would be on the alternates chart.
Wouldn't an automated plane have a different sensor suite? I don't think we are talking about inflating the automatic pilot from Airplane! and calling it a day.
Is a glide path into a river something beyond the capability of automation? I'm skeptical, but this isn't my field.
But there are standard procedures to follow when the sensors give bad results, which were apparently not followed by the pilots on 447. I disagree that these would be more confusing to the remote pilot. The autopilot was smart enough to recognize a problem and disconnect. If there was nothing but an autopilot, it could be set to a fail safe mode. A typical smartphone is probably better equipped than a human when it comes to flying without visual reference points.
"Store it"? You are doing it wrong.
Resources may not make enough jobs to keep you happy, but it does provide jobs - and good ones at that. Look at the boon to the Texas economy from fracking.
Lenovo may not be the best example for you to use, since Thinkpads were already manufactured overseas. But I take your point - sometimes a purchase is just the brand name, and the jobs go up in a puff of smoke. But to be honest, those cases are usually fire sales where the company is on its last legs anyway. Think Palm or AMC.
I agree that my posts are radically simplified. They spent over an hour talking about this on a recent NPR show, and they barely scratched the surface. Nevertheless, I stand by my narrative that owing the Chinese government over a trillion dollars is not as incredibly horrible as it would seem at first blush.
The VS system didn't come out until after the NES (well, the Famikon version from Japan). You have the timeline backwards.
Many of the cabinets on VS machines are older because they were converted Donkey Kong or Mario Bros cabinets.
Resources == Jobs. Companies are nothing but collections of people. Even real estate is an almost infinite resource in the US. I'm not going to say that running a constant deficit is a good thing, but we could do far worse than to have another country lend us a bunch of our own currency at a rate below inflation.
Yes, exactly. And on top of it all, ultimately dollars are only good for purchasing US goods and services. So at some point - admittedly it could be far in the future - those dollars will presumably help support American jobs.
I don't follow. How would a built-in HTML library have taken away from Netscape as a standalone web browser? I could see it taking away from some theoretical Netscape libraries meant for application development, but if such a product existed it was not very prevalent.
Or are you saying that competing web browsers were using the IE libraries? That may have been the case, but my recollection of the time was that everyone was using home-grown code.
That's a semantic argument. There are applications that depend on the libraries used by IE in order to render HTML. But nothing about Windows requires IE be your web browser. Most OSes have a built-in HTML rendering library these days.
To put it another way, a built-in library to render HTML from within applications would not have killed Netscape.
Obtaining actual physical goods for IOUs is a pretty good deal IMHO.
Funny that you mention "repo men" since the car in the summary, down to the license plate, is from the opening scene in Repo Man. Don't look in the trunk.