You miss the point. In fact you agree with my point, and Google's own concerns. I don't assume anything about the biggest cause of rear end highway crashes. The NHTSA does. Based on their data, just as you said the driver behind the driver that gets cut off doesn't react quick enough. Just as you said the AI car will react quicker than a human driven car, meaning there is even less time for the human driven car behind the AI car to react. So yes, as you said, you get rear ended because the person behind you did the same thing as the AI car but didn't quite make it.
Io is the most volcanic body in the solar system. By definition, wouldn't that mean it has a volcanic sub-surface? Last time I checked, magma comes up from underground to the surface.
And yet, when those uses are pointed out to you, you protest. There are really a lot of reasons for a car to be without a driver at all, and even more for an inattentive driver. Drive at night, get some sleep! Do the crossword, catch up on mail.
Look, if you think it is a great idea to pay $50,000 or so for a car so you can text or sleep or whatever, great, go do it. But, unless you are doing that now while your are driving, please don't try and say it will be safer. But remember that not everybody involved with an accident with a drunk driver is drunk, nor everybody in an accident because a driver fell asleep is also sleeping. So, sleep away in the vehicle while it is going down the road at night. Tell yourself it is safe and you don't have a worry in the world. But don't be surprised if you are still in an accident because not everybody purchased a $50,000 car like yours.
Airliners have crash avoidance technology, too. They can even automatically land and take off (or at least be programmed to do so). But for some reason, even with all of this technology, planes without pilots does not seem in the best interest of public safety. Why would cars without drivers be different?
According to the NHTS board, most multi car accidents occur on highways or intersections. That's not hard to believe, is it? Of the highway accidents, most of those accidents occur on multi lane highways (more than two lanes in one direction) and they occur at times of most congestion. Again, that's not hard to believe is it? That is when the most drivers are on the highway and there is the least amount of room to manuever your vehicle because of the other drivers (NHTS explanation, not mine).
So, the most accident prone areas drivers face are intersections and heavily congested multi lane highways. Not my opinion, but actual research data. Now, from Google's own briefs on the their new vehicle, these are the two areas that are most difficult to provide for. Why? Again, according to google, once the vehicle enters the intersection, there is little the car can do to detect and avoid another driver entering the intersection from perpendicular directions (my paraphrase, not theirs). As for congested highways, they admit that their vehicle is best at accident avoidance when it has ample room to maneuver around an obstacle (strange, that is when I am best at accident avoidance, too).
So, while the google car will not enter an intersection if it is blocked, it does not prevent others from entering and hitting the car. Likewise, in heavy traffic, with all lanes occupied, the avoidance abilities are severely restricted (ie. pretty much left to braking). The major types of accidents the google vehicle will prevent are those where you as the driver would have caused it because of your inattentiveness.
Don't get me wrong, Google has come up with some impressive technology. But it cannot and never will be able to protect you driving your car from something I may do in my non-robotic car. That is why, it doesn't truly improve highway safety until there are sufficient numbers of them to minimize the risk in the majority of where accidents occur.
Now, please roll your eyes back so that you may see the road the next time you are driving.
You know... I've finally adopted the 'use linux as much as possible' mindset... settled on a distro and now they are trying to fuck it up as much as they can.
Bravo.
-AI
Not sure what you are referring to. But, if the distro you settled on is Ubuntu and the desktop is gnome. The old classic gnome is still there and you can login and use it. Gnome itself is changing and the old desktop is going away, but that is happening with or without Ubuntu's involvement. Changing to lightdm from gdm shouldn't make a difference to your use. It is simple the portion where you type your user id and password.
Almost gnome free. Xubuntu has a number of apps that still have gnome-dependencies, like the administrative tools, the package manager (both their own and synaptic), evince (although a gnome free version is in the repository), etc. That said, these would only run a gnome session when actually used, whereas gdm runs all the time. LightDM doesn't have a configuration tool, but it is easy to configure. I would expect themes to become available without much delay. The lack of GDM customization is not Ubuntu's fault. The Gnome developers took that ability away and nobody has taken it upon themself to write a new app to do it.
To summarise, their argument is that LightDM is light on code because it can't do as much as GDM and the others, and if you removed those features from the others they would be light as well.
If that's true and that is the main difference, maybe it'd be easier to strip out, or turn off, parts of GDM if Canonical wants to dispose of certain features to achieve a faster boot time.
11.04 is SO SLOW to boot in comparison to 10.10.
I don't think stripping out parts from GDM would be a good idea, no matter how much it needs to be done. They've already caught a lot of flack about stripping out parts from gnome3 (by using Unity instead of gnome-shell).
As for 11.04 being slow to boot. I believe that is a kernel regression. A lot of other distros are having similar issues with boot time and also ath5k and 9k wireless. Once it is fixed upstream it will resolve itself in all the distros.
Webkit is most definitely not required for lightdm. It is just one of several presentation layers. However, if used on something like a netbook or tablet, it would make a lot of sense, since the browser would be one of the first things opened on the device. As for exploits, again, if the user is going to be doing pretty much everything in the cloud, which seems to be Ubuntu's opinion (and many others), then a webkit exploit is just as exploitive at the login screen vs the actual desktop. However, since access can be restricted at the login screen, the danger should be minimized.
You are missing the point. While I do not care for the Unity interface, you are missing the point. Unity was rushed, because gnome3 and gnome-shell were so, so late. Then, because of the delay, Ubuntu, look at what they found problematic with gnome-shell (both functional and political) and made their own shell. Again, while I don't care for Unity, it is pretty usable and was made so in a relatively short period of time. From that perspective, the developers should get recognition. Again, if gnome3 had shipped on time, there probably wouldn't be unity.
As for Unity not being ready. I think most people, including Ubuntu would admit that, or at least that it isn't completed yet. For new users, they usually direct them to their LTS releases which Unity is not even an issue. For advanced users, they can easily switch to the old gnome interface. I do think, the old interface should have been the default with Unity being optional, but they didn't ask me. What having Unity in 11.04 provides Ubuntu a very large usability and testing population to perfect it for 11.10.
As for moving to lightdm, they aren't the only ones. While KDE has KDM, all the other environments have to use XDM or GDM. XDM works, but it isn't very pretty. GDM makes a lot of gnome load and run in the background, even though you aren't running gnome for a desktop. Switching to lightdm will mean that if you want to run XFCE or openbox or E17 or whatever, you don't have to have a gnome session running just to provide login/logout capabilities.
A very easy way to reduce highway accidents and fatalities would be to lower speed limits.
And yet, amazingly, when the raised the speed limit from 55 to 65 nationwide there weren't any more fatalities. Probably because all it meant was that people went from driving 67mph to driving 69mph on average.
So basically you're full of shit and have no idea what you're talking about.
Actually, highway accidents increased and have stayed at the higher level since they raised the speed limit. I would have to research fatalities, but it could be attributed to things like airbags, etc.
there is no real safety increase unless most drivers are driving the new vehicle.
Because you're off course an ubermensch who never gets distracted, tired or loses concentration while driving. Never, ever. And you have perfect reactions and situational awareness at all times while driving. Plus the training to apply them.
your "new" car doesn't have many options other than to brake.
It can react/brake more quickly than a human and can probably notice the driver changing lanes before a human could (since it's "eyes" aren't limited to one direction at a time). Since it also has perfect situational awareness it can simply changes lanes if there is room and if the driver behind is too close.
I'm confused. You seem to be arguing my point. I am the one saying that because of other's reaction time, the new car is not as safe as it might seem. What good does it do to have a vehicle that avoids hitting the car in front of it with it's faster than human reaction time when it causes the car behind it to plow into it, because of its (the second car) human reaction time?
As for drivers being distracted. How will your safe new robotic car prevent that in the others who don't have it? It can't. Again, this means your car will be reacting to these distracted drivers and trying to avoid them. That is good. However, if you are in the middle lane with cars on both sides, where exactly will your car go? The answer is straight ahead, it is the only option. It will try and brake, again quicker than humanly possible, so as to try and avoid an accident, but that is what causes the driver behind to hit it. In the end, without the new safe car, in such a situation, you would have been in an accident by hitting the person who cut you off or by sideswiping another vehicle. With the new car, you are still in an accident.
Given real driving situations, on congested highways, there is not much that can be done to avoid a collision, since it is out of your control what is around you. On less congested roads, that is different. But, then, your new car may help you by letting you do things while riding as a passenger in it, like texting, talking on the phone, reading, putting on makeup, having sex, whatever, that would distract you if you were doing any of that and trying to drive. But, it can only protect you from what you were doing, it cannot keep me or anybody else from being distracted and killing you. That is why there is no real safety increase, unless everybody or most everybody drives such a vehicle.
Carfax reports all accidents. There is a difference from being in an accident that you cause or someone else causes. Then there are those things outside of anyone's control -- black ice, tree limbs, hail, etc. The 100K number would be accidents caused by the car malfunctioning itself. The robot cars can't do very little about acts of god or other idiots or pedestrians and animals. If, they are in areas with little congestion, it will seem that way, but that is because percentage wise, there is little to interact negatively with the vehicle.
There is a commercial on TV for Mercedes Benz, I think, where the woman states how she was drifting off and the car alerted her that she was drifting into the other lane. That's fine on the deserted night road they showed in the commercial. I-95 during rush hour would be a totally different story as the time for the car to alert you and for you to take corrective action would exceed the next car over from hitting you.
Now, this google car would do great in that situation, it won't drift over. Where the question arises, however, would be as to what happens when the human driven vehicle ahead of you one lane over swerves into your lane? In traffic, there is little room to swerve to miss another vehicle, at least without hitting one beside you. The real options are to brake. But, again, the car applies the brake quiker than a person can realize what is happening and slows down to avoid the accident. Unfortunately, the SUV behind you doesn't and runs over the top of your car.
Most accidents are not head on collisions, which are where a robotic car's collision avoidance would be of most advantage. Most are side and rear collisions. It is pretty difficult for a car, no matter how sophisticated, to avoid these. At least to be able to avoid them at g-forces the human body can withstand. Now, if every vehicle,or even the vast majority of vehicles were this new car type, then maybe those other collisions would be avoided, too. Unfortunately, even if these cars were available everywhere tommorrow, there would be 125 million vehicles needing replacement. That will not happen for decades.
Meanwhile, during those decades, these cars will be in numerous accidents and blamed for all kinds of things that wouldn't have happened if there had been a real driver. Look at seatbelts, you never hear about how many lives are saved by seatbelts, only those that survived because they weren't wearing one. The same will happen with these cars. You won't hear about the 100 crashes that were avoided, only the one, where the kid was killed because the car malfunctioned.
But even assuming that the vehicle can survive all of the negative publicity, the cost, and whatever and is finally excepted by the public. In 30 years there are enough on the road to make it safe or at least safer than it is now. Of course, that only works if the technology freezes. But I would hope that in 30 years there are much better avenues for safe high speed transit. Most likely it won't involve individual passenger vehicles which regardless of their propulsion are the least efficient means of transportation (but are usually the most convenient). Long before these vehicles reach critical mass to make good on the safety projections, they will already be obsolete and again wont be what is accepted.
Google, if they truly were visionary, should be looking at what would really revolutionize transportation, versus incremental evolutionary changes.
That's not the point. If there were 400 or 4,000,000 buggy whip makers, the profession would still be obsolete and banning cars wouldn't have been a reasonable action just to protect those jobs. Yes, it's possible that self driving vehicles could reduce the call for taxis and delivery people, but that doesn't mean that we should avoid developing the technology, especially if it does it better than those people do it now. Robots have displaced millions of factory workers, computers in general have displaced millions more. Do you think we should be seriously thinking that we shouldn't have developed computers or robots though because of that?
The fact is that as technology progresses, the need for human labor generally trends down, and we haven't come up with a really good way to deal with what to do with that excess labor force. I'm pretty sure the answer isn't to stop developing new technologies.
Don't put words in my mouth (or on my keyboard). I am all for technology that improves the overall human condition. I don't understand how a driverless car will do that, unless being able to do other things while commuting is important. And if that is the case, then is an individual vehicle the best means of obtaining that goal? Replacing the commuter train from NJ to NY with a bunch of cars that can drive themself does not seem the most efficient use or resources.
Don't get me wrong, I think these cars are a bad idea, but not because of the reasons I've posted in this particular thread. I was just commenting on somebody else's post, not the original article. I would however ask for some type of proof of your last point in referring to human labor generally going down. Studies show that labor does not decrease, but only is transformed to different types of labor, from farm to factory, from factory to office, etc. In addition, studies have shown for office workers, technology has actually increased the amount of labor of an individual if measured by time. Instead of all of this technology freeing us up so we have more free time, instead it turns out that it actually enables/causes us to work longer hours. At least that is what the studies show.
And all of those taxi and delivery drivers who will be out of work can become the workforce that builds the new cars, right? Of course, they'll probably have to move to Japan.
And what about all those buggy whip makers who will go out of business because no one requires their long leather whips with fine, knurled wooden handles chased with silver filigree!?!?
In the US, anyway, there were never more than 400 buggy whip manufactures. That's not companies, but those employed in the industry as actual buggy whip makers. There are more taxi cab drivers sitting at any of our busiest airports right now than that.
But the only way for such a system to actually lower accident/injury/death rate would be for the majority of vehicles to be of the new system. As long as the other 125 million vehicles on the road aren't the new system, then any safety improvements can't occur.
Facts not in evidence. If an automatic car can be made to drive more safely than the average driver, would it not lower deaths with even partial implementation?
If you are in one of these "smart" cars, but have vehicles on both sides, there are not many options to avoiding an accident other than braking.
Which the automated car can do *much* more quickly. Reaction times matter enormously in that sort of situation, and that's a contest the machine wins every time.
You have a solution for the problem google already identified, you should contact them. One of their areas of concern is specifically heavy traffic on an interstate where not all cars are their new vehicle. Evidently, their own engineers are worried about the vehicle causing accidents because the human driven vehicles won't react in time to the quick reaction of their car.
They specifically picked Nevada because of low density traffic patterns to test the vehicle with the general public. But, again, if you solved the concerns of their own engineers, you should be contacting them.
Don't forget the economics. Assuming 1/2 of the 125 million cars on the road need to be this new vehicle to gain the safety numbers needed for public acceptance. Also, figure int he price, say what $40,000 as the Chevy Volt? That is priced out of the range were the majority of Americans can afford, so the government steps in an subsidzes the cost with a $10,000 tax credit. That at least brings it down to $30,000, which half still can't afford, but ignoring that, it costs the taxpayers $625B. Where will that money come from?
They wanted to automate aircraft and railroads a long time ago. They had/have the technology to do it. However, the unpredictability of the operating environment shelved the idea. Why would we expect a car to behave any differently? The notion that a driver could take over if the system experienced difficulty only works if a) there is time and b) they are alert enough to do it. On an aircraft which takes a minute to fall from the sky, that is one thing. On a car that takes seconds to hit another car or person, it is unlikely. If we had that good reaction time, we would need robotic cars int he first place.
The issue is not that current cars are not safe. It is that current drivers are doing all sorts of things besides driving and they don't pay attention.
They would likely gain full public acceptance after a period of time where the general public has been free to utilize them and among the people who are using them, there is a demonstrably lower accident rate.
The general public's ability to utilize them will be strongly influenced by the price. The Chevy Volt is $40,000. How much will this cost? If the price is above what most people can afford, particularly young people -- who cause most accidents, then how will that lower accident rate ever be realized. The only way to get a sufficient number of these vehicles on the road at an affordable price would be to subsidize them with tax dollars. Assuming a $10,000 subsidy to replace just half of the 125 million vehicles on the road would cost the taxpayers $625B. Where will that money come from? Even still, over half the population probably still wouldn't be able to afford it.
A predicted crash every 100k miles is doing pretty well. I probably have that rate. So far one wreck with failed ABS, it decided to not apply the brakes even though the wheels were not locked. The other due to black ice. Both were technically my fault.
Okay, you have had one mechanical failure and one inclement weather accident. Neither of those are preventable by the new vehicle and are just as likely to occur.
But, the error rate is per 100,000 miles driven. Would you actually purchase and drive a vehicle that states at least once in the lifetime of the vehicle, you are expected to get into a serious accident because of a software glitch?
Automation will bring prices down so much that you hardly have to work to buy what you need. I can find computers in the trash that would be considered Super Computers 20 years ago.
Yes, we've seen that with current car technology. In 1970 an average vehicle cost 20% of average wages. Today it is close to 50%. And yet, in 1970, most vehicles were assembled by hand. Today, most are assembled through automation. Music CDs are infinitely cheaper to produce than cassette tapes, but even after taking into account inflation, they haven't come down in price. While it is true somethings do get cheaper, it tends to do more with supply and demand than cost of production. A business will charge whatever it can on a product, regardless of the cost to manufacture. If there isn't enough of a difference between the selling price and the cost to produce, they will quit producing it.
Automation only reduces production costs, it doesn't impact selling price. Only competition can do that. But notice, even though there are numerous car manufacturers, they all sell in the same price range. Why? Because that is what the market will bear.
Society is going to be the problem here anyway. People are going to freak out at cases where the driving AI is responsible for a fatal accident. A quick search shows that 33808 people died in road accidents in the US, in 2009. And that's apparently a 60-year low. This still translates to some 92 traffic fatalities per day. But society accepts that... whereas I'm sure they would freak out if a full transition to self-driving cars happened, with the driver AI being responsible for 1 fatality per day. Fatality numbers could go down by almost two orders of magnitude, but people would feel less safe on the road because of "killer cars" out there.
I feel this is a big problem overall - people are willing to accept human controlled systems where the human factor regularly leads to accidents/injuries/deaths, but if that system can be automated with a much lower accident/injury/death rate, the society would not feel it's safe.
But the only way for such a system to actually lower accident/injury/death rate would be for the majority of vehicles to be of the new system. As long as the other 125 million vehicles on the road aren't the new system, then any safety improvements can't occur. Most fatalities occur on multi-lane roads (highways). If you are in one of these "smart" cars, but have vehicles on both sides, there are not many options to avoiding an accident other than braking. Then you better hope the SUV behind you stops in time. Otherwise, even though it isn't your fault or the fault of the "smart" vehicle, accident/injury/death will occur.
A very easy way to reduce highway accidents and fatalities would be to lower speed limits. It also would lessen dependance on foreign oil. That's something that could be done today.
Society doesn't need "smart" cars, we need smart people.
Slashdot Mirror
Too bad the called Ubuntu a communist state. That plays right into the hands of Microsoft and proponents of closed source programs
You miss the point. In fact you agree with my point, and Google's own concerns. I don't assume anything about the biggest cause of rear end highway crashes. The NHTSA does. Based on their data, just as you said the driver behind the driver that gets cut off doesn't react quick enough. Just as you said the AI car will react quicker than a human driven car, meaning there is even less time for the human driven car behind the AI car to react. So yes, as you said, you get rear ended because the person behind you did the same thing as the AI car but didn't quite make it.
From the summary:
The global magma ocean about 30 to 50 kilometers (20 to 30 miles) beneath Io's surface explains the moon's activity.
All these years, I thought it was physics that explained the moons activity.
Io is the most volcanic body in the solar system. By definition, wouldn't that mean it has a volcanic sub-surface? Last time I checked, magma comes up from underground to the surface.
And yet, when those uses are pointed out to you, you protest. There are really a lot of reasons for a car to be without a driver at all, and even more for an inattentive driver. Drive at night, get some sleep! Do the crossword, catch up on mail.
Look, if you think it is a great idea to pay $50,000 or so for a car so you can text or sleep or whatever, great, go do it. But, unless you are doing that now while your are driving, please don't try and say it will be safer. But remember that not everybody involved with an accident with a drunk driver is drunk, nor everybody in an accident because a driver fell asleep is also sleeping. So, sleep away in the vehicle while it is going down the road at night. Tell yourself it is safe and you don't have a worry in the world. But don't be surprised if you are still in an accident because not everybody purchased a $50,000 car like yours.
Airliners have crash avoidance technology, too. They can even automatically land and take off (or at least be programmed to do so). But for some reason, even with all of this technology, planes without pilots does not seem in the best interest of public safety. Why would cars without drivers be different?
Last post on this topic.
According to the NHTS board, most multi car accidents occur on highways or intersections. That's not hard to believe, is it? Of the highway accidents, most of those accidents occur on multi lane highways (more than two lanes in one direction) and they occur at times of most congestion. Again, that's not hard to believe is it? That is when the most drivers are on the highway and there is the least amount of room to manuever your vehicle because of the other drivers (NHTS explanation, not mine).
So, the most accident prone areas drivers face are intersections and heavily congested multi lane highways. Not my opinion, but actual research data. Now, from Google's own briefs on the their new vehicle, these are the two areas that are most difficult to provide for. Why? Again, according to google, once the vehicle enters the intersection, there is little the car can do to detect and avoid another driver entering the intersection from perpendicular directions (my paraphrase, not theirs). As for congested highways, they admit that their vehicle is best at accident avoidance when it has ample room to maneuver around an obstacle (strange, that is when I am best at accident avoidance, too).
So, while the google car will not enter an intersection if it is blocked, it does not prevent others from entering and hitting the car. Likewise, in heavy traffic, with all lanes occupied, the avoidance abilities are severely restricted (ie. pretty much left to braking). The major types of accidents the google vehicle will prevent are those where you as the driver would have caused it because of your inattentiveness.
Don't get me wrong, Google has come up with some impressive technology. But it cannot and never will be able to protect you driving your car from something I may do in my non-robotic car. That is why, it doesn't truly improve highway safety until there are sufficient numbers of them to minimize the risk in the majority of where accidents occur.
Now, please roll your eyes back so that you may see the road the next time you are driving.
You know... I've finally adopted the 'use linux as much
as possible' mindset... settled on a distro and now they
are trying to fuck it up as much as they can.
Bravo.
-AI
Not sure what you are referring to. But, if the distro you settled on is Ubuntu and the desktop is gnome. The old classic gnome is still there and you can login and use it. Gnome itself is changing and the old desktop is going away, but that is happening with or without Ubuntu's involvement. Changing to lightdm from gdm shouldn't make a difference to your use. It is simple the portion where you type your user id and password.
Almost gnome free. Xubuntu has a number of apps that still have gnome-dependencies, like the administrative tools, the package manager (both their own and synaptic), evince (although a gnome free version is in the repository), etc. That said, these would only run a gnome session when actually used, whereas gdm runs all the time. LightDM doesn't have a configuration tool, but it is easy to configure. I would expect themes to become available without much delay. The lack of GDM customization is not Ubuntu's fault. The Gnome developers took that ability away and nobody has taken it upon themself to write a new app to do it.
http://www.advogato.org/person/mjg59/diary.html?start=296
To summarise, their argument is that LightDM is light on code because it can't do as much as GDM and the others, and if you removed those features from the others they would be light as well.
If that's true and that is the main difference, maybe it'd be easier to strip out, or turn off, parts of GDM if Canonical wants to dispose of certain features to achieve a faster boot time.
11.04 is SO SLOW to boot in comparison to 10.10.
I don't think stripping out parts from GDM would be a good idea, no matter how much it needs to be done. They've already caught a lot of flack about stripping out parts from gnome3 (by using Unity instead of gnome-shell).
As for 11.04 being slow to boot. I believe that is a kernel regression. A lot of other distros are having similar issues with boot time and also ath5k and 9k wireless. Once it is fixed upstream it will resolve itself in all the distros.
It's a normal X login manager without all the extra crap from gdm and kdm, and since it's in the Ubuntu repos already, all it needs is a good theme.
I agree 100%. The needing a good theme, that is. The default webkit and gtk themes really suck.
Webkit is most definitely not required for lightdm. It is just one of several presentation layers. However, if used on something like a netbook or tablet, it would make a lot of sense, since the browser would be one of the first things opened on the device. As for exploits, again, if the user is going to be doing pretty much everything in the cloud, which seems to be Ubuntu's opinion (and many others), then a webkit exploit is just as exploitive at the login screen vs the actual desktop. However, since access can be restricted at the login screen, the danger should be minimized.
You are missing the point. While I do not care for the Unity interface, you are missing the point. Unity was rushed, because gnome3 and gnome-shell were so, so late. Then, because of the delay, Ubuntu, look at what they found problematic with gnome-shell (both functional and political) and made their own shell. Again, while I don't care for Unity, it is pretty usable and was made so in a relatively short period of time. From that perspective, the developers should get recognition. Again, if gnome3 had shipped on time, there probably wouldn't be unity.
As for Unity not being ready. I think most people, including Ubuntu would admit that, or at least that it isn't completed yet. For new users, they usually direct them to their LTS releases which Unity is not even an issue. For advanced users, they can easily switch to the old gnome interface. I do think, the old interface should have been the default with Unity being optional, but they didn't ask me. What having Unity in 11.04 provides Ubuntu a very large usability and testing population to perfect it for 11.10.
As for moving to lightdm, they aren't the only ones. While KDE has KDM, all the other environments have to use XDM or GDM. XDM works, but it isn't very pretty. GDM makes a lot of gnome load and run in the background, even though you aren't running gnome for a desktop. Switching to lightdm will mean that if you want to run XFCE or openbox or E17 or whatever, you don't have to have a gnome session running just to provide login/logout capabilities.
A very easy way to reduce highway accidents and fatalities would be to lower speed limits.
And yet, amazingly, when the raised the speed limit from 55 to 65 nationwide there weren't any more fatalities. Probably because all it meant was that people went from driving 67mph to driving 69mph on average.
So basically you're full of shit and have no idea what you're talking about.
Actually, highway accidents increased and have stayed at the higher level since they raised the speed limit. I would have to research fatalities, but it could be attributed to things like airbags, etc.
there is no real safety increase unless most drivers are driving the new vehicle.
Because you're off course an ubermensch who never gets distracted, tired or loses concentration while driving. Never, ever. And you have perfect reactions and situational awareness at all times while driving. Plus the training to apply them.
your "new" car doesn't have many options other than to brake.
It can react/brake more quickly than a human and can probably notice the driver changing lanes before a human could (since it's "eyes" aren't limited to one direction at a time). Since it also has perfect situational awareness it can simply changes lanes if there is room and if the driver behind is too close.
I'm confused. You seem to be arguing my point. I am the one saying that because of other's reaction time, the new car is not as safe as it might seem. What good does it do to have a vehicle that avoids hitting the car in front of it with it's faster than human reaction time when it causes the car behind it to plow into it, because of its (the second car) human reaction time?
As for drivers being distracted. How will your safe new robotic car prevent that in the others who don't have it? It can't. Again, this means your car will be reacting to these distracted drivers and trying to avoid them. That is good. However, if you are in the middle lane with cars on both sides, where exactly will your car go? The answer is straight ahead, it is the only option. It will try and brake, again quicker than humanly possible, so as to try and avoid an accident, but that is what causes the driver behind to hit it. In the end, without the new safe car, in such a situation, you would have been in an accident by hitting the person who cut you off or by sideswiping another vehicle. With the new car, you are still in an accident.
Given real driving situations, on congested highways, there is not much that can be done to avoid a collision, since it is out of your control what is around you. On less congested roads, that is different. But, then, your new car may help you by letting you do things while riding as a passenger in it, like texting, talking on the phone, reading, putting on makeup, having sex, whatever, that would distract you if you were doing any of that and trying to drive. But, it can only protect you from what you were doing, it cannot keep me or anybody else from being distracted and killing you. That is why there is no real safety increase, unless everybody or most everybody drives such a vehicle.
Carfax reports all accidents. There is a difference from being in an accident that you cause or someone else causes. Then there are those things outside of anyone's control -- black ice, tree limbs, hail, etc. The 100K number would be accidents caused by the car malfunctioning itself. The robot cars can't do very little about acts of god or other idiots or pedestrians and animals. If, they are in areas with little congestion, it will seem that way, but that is because percentage wise, there is little to interact negatively with the vehicle.
There is a commercial on TV for Mercedes Benz, I think, where the woman states how she was drifting off and the car alerted her that she was drifting into the other lane. That's fine on the deserted night road they showed in the commercial. I-95 during rush hour would be a totally different story as the time for the car to alert you and for you to take corrective action would exceed the next car over from hitting you.
Now, this google car would do great in that situation, it won't drift over. Where the question arises, however, would be as to what happens when the human driven vehicle ahead of you one lane over swerves into your lane? In traffic, there is little room to swerve to miss another vehicle, at least without hitting one beside you. The real options are to brake. But, again, the car applies the brake quiker than a person can realize what is happening and slows down to avoid the accident. Unfortunately, the SUV behind you doesn't and runs over the top of your car.
Most accidents are not head on collisions, which are where a robotic car's collision avoidance would be of most advantage. Most are side and rear collisions. It is pretty difficult for a car, no matter how sophisticated, to avoid these. At least to be able to avoid them at g-forces the human body can withstand. Now, if every vehicle,or even the vast majority of vehicles were this new car type, then maybe those other collisions would be avoided, too. Unfortunately, even if these cars were available everywhere tommorrow, there would be 125 million vehicles needing replacement. That will not happen for decades.
Meanwhile, during those decades, these cars will be in numerous accidents and blamed for all kinds of things that wouldn't have happened if there had been a real driver. Look at seatbelts, you never hear about how many lives are saved by seatbelts, only those that survived because they weren't wearing one. The same will happen with these cars. You won't hear about the 100 crashes that were avoided, only the one, where the kid was killed because the car malfunctioned.
But even assuming that the vehicle can survive all of the negative publicity, the cost, and whatever and is finally excepted by the public. In 30 years there are enough on the road to make it safe or at least safer than it is now. Of course, that only works if the technology freezes. But I would hope that in 30 years there are much better avenues for safe high speed transit. Most likely it won't involve individual passenger vehicles which regardless of their propulsion are the least efficient means of transportation (but are usually the most convenient). Long before these vehicles reach critical mass to make good on the safety projections, they will already be obsolete and again wont be what is accepted.
Google, if they truly were visionary, should be looking at what would really revolutionize transportation, versus incremental evolutionary changes.
That's not the point. If there were 400 or 4,000,000 buggy whip makers, the profession would still be obsolete and banning cars wouldn't have been a reasonable action just to protect those jobs. Yes, it's possible that self driving vehicles could reduce the call for taxis and delivery people, but that doesn't mean that we should avoid developing the technology, especially if it does it better than those people do it now. Robots have displaced millions of factory workers, computers in general have displaced millions more. Do you think we should be seriously thinking that we shouldn't have developed computers or robots though because of that?
The fact is that as technology progresses, the need for human labor generally trends down, and we haven't come up with a really good way to deal with what to do with that excess labor force. I'm pretty sure the answer isn't to stop developing new technologies.
Don't put words in my mouth (or on my keyboard). I am all for technology that improves the overall human condition. I don't understand how a driverless car will do that, unless being able to do other things while commuting is important. And if that is the case, then is an individual vehicle the best means of obtaining that goal? Replacing the commuter train from NJ to NY with a bunch of cars that can drive themself does not seem the most efficient use or resources.
Don't get me wrong, I think these cars are a bad idea, but not because of the reasons I've posted in this particular thread. I was just commenting on somebody else's post, not the original article. I would however ask for some type of proof of your last point in referring to human labor generally going down. Studies show that labor does not decrease, but only is transformed to different types of labor, from farm to factory, from factory to office, etc. In addition, studies have shown for office workers, technology has actually increased the amount of labor of an individual if measured by time. Instead of all of this technology freeing us up so we have more free time, instead it turns out that it actually enables/causes us to work longer hours. At least that is what the studies show.
And all of those taxi and delivery drivers who will be out of work can become the workforce that builds the new cars, right? Of course, they'll probably have to move to Japan.
And what about all those buggy whip makers who will go out of business because no one requires their long leather whips with fine, knurled wooden handles chased with silver filigree!?!?
In the US, anyway, there were never more than 400 buggy whip manufactures. That's not companies, but those employed in the industry as actual buggy whip makers. There are more taxi cab drivers sitting at any of our busiest airports right now than that.
Facts not in evidence. If an automatic car can be made to drive more safely than the average driver, would it not lower deaths with even partial implementation?
Which the automated car can do *much* more quickly. Reaction times matter enormously in that sort of situation, and that's a contest the machine wins every time.
You have a solution for the problem google already identified, you should contact them. One of their areas of concern is specifically heavy traffic on an interstate where not all cars are their new vehicle. Evidently, their own engineers are worried about the vehicle causing accidents because the human driven vehicles won't react in time to the quick reaction of their car.
They specifically picked Nevada because of low density traffic patterns to test the vehicle with the general public. But, again, if you solved the concerns of their own engineers, you should be contacting them.
Don't forget the economics. Assuming 1/2 of the 125 million cars on the road need to be this new vehicle to gain the safety numbers needed for public acceptance. Also, figure int he price, say what $40,000 as the Chevy Volt? That is priced out of the range were the majority of Americans can afford, so the government steps in an subsidzes the cost with a $10,000 tax credit. That at least brings it down to $30,000, which half still can't afford, but ignoring that, it costs the taxpayers $625B. Where will that money come from?
They wanted to automate aircraft and railroads a long time ago. They had/have the technology to do it. However, the unpredictability of the operating environment shelved the idea. Why would we expect a car to behave any differently? The notion that a driver could take over if the system experienced difficulty only works if a) there is time and b) they are alert enough to do it. On an aircraft which takes a minute to fall from the sky, that is one thing. On a car that takes seconds to hit another car or person, it is unlikely. If we had that good reaction time, we would need robotic cars int he first place.
The issue is not that current cars are not safe. It is that current drivers are doing all sorts of things besides driving and they don't pay attention.
They would likely gain full public acceptance after a period of time where the general public has been free to utilize them and among the people who are using them, there is a demonstrably lower accident rate.
The general public's ability to utilize them will be strongly influenced by the price. The Chevy Volt is $40,000. How much will this cost? If the price is above what most people can afford, particularly young people -- who cause most accidents, then how will that lower accident rate ever be realized. The only way to get a sufficient number of these vehicles on the road at an affordable price would be to subsidize them with tax dollars. Assuming a $10,000 subsidy to replace just half of the 125 million vehicles on the road would cost the taxpayers $625B. Where will that money come from? Even still, over half the population probably still wouldn't be able to afford it.
A predicted crash every 100k miles is doing pretty well. I probably have that rate. So far one wreck with failed ABS, it decided to not apply the brakes even though the wheels were not locked. The other due to black ice. Both were technically my fault.
Okay, you have had one mechanical failure and one inclement weather accident. Neither of those are preventable by the new vehicle and are just as likely to occur.
But, the error rate is per 100,000 miles driven. Would you actually purchase and drive a vehicle that states at least once in the lifetime of the vehicle, you are expected to get into a serious accident because of a software glitch?
If you drove a billion miles, you would have lived far longer than any other human being... and older than most trees, in fact.
There are 125 million vehicles in the US. If each of those vehicles drives 1000 miles this year, then that is 125 billion miles.
Automation will bring prices down so much that you hardly have to work to buy what you need. I can find computers in the trash that would be considered Super Computers 20 years ago.
Yes, we've seen that with current car technology. In 1970 an average vehicle cost 20% of average wages. Today it is close to 50%. And yet, in 1970, most vehicles were assembled by hand. Today, most are assembled through automation. Music CDs are infinitely cheaper to produce than cassette tapes, but even after taking into account inflation, they haven't come down in price. While it is true somethings do get cheaper, it tends to do more with supply and demand than cost of production. A business will charge whatever it can on a product, regardless of the cost to manufacture. If there isn't enough of a difference between the selling price and the cost to produce, they will quit producing it.
Automation only reduces production costs, it doesn't impact selling price. Only competition can do that. But notice, even though there are numerous car manufacturers, they all sell in the same price range. Why? Because that is what the market will bear.
Society is going to be the problem here anyway. People are going to freak out at cases where the driving AI is responsible for a fatal accident. A quick search shows that 33808 people died in road accidents in the US, in 2009. And that's apparently a 60-year low. This still translates to some 92 traffic fatalities per day. But society accepts that... whereas I'm sure they would freak out if a full transition to self-driving cars happened, with the driver AI being responsible for 1 fatality per day. Fatality numbers could go down by almost two orders of magnitude, but people would feel less safe on the road because of "killer cars" out there.
I feel this is a big problem overall - people are willing to accept human controlled systems where the human factor regularly leads to accidents/injuries/deaths, but if that system can be automated with a much lower accident/injury/death rate, the society would not feel it's safe.
But the only way for such a system to actually lower accident/injury/death rate would be for the majority of vehicles to be of the new system. As long as the other 125 million vehicles on the road aren't the new system, then any safety improvements can't occur. Most fatalities occur on multi-lane roads (highways). If you are in one of these "smart" cars, but have vehicles on both sides, there are not many options to avoiding an accident other than braking. Then you better hope the SUV behind you stops in time. Otherwise, even though it isn't your fault or the fault of the "smart" vehicle, accident/injury/death will occur.
A very easy way to reduce highway accidents and fatalities would be to lower speed limits. It also would lessen dependance on foreign oil. That's something that could be done today.
Society doesn't need "smart" cars, we need smart people.