Toasters and baths have already been solved. It is code in most areas that all bathroom outlets have ground-fault interrupters, and these are routinely tested any time a house is bought and sold. A toaster dropped into a bathtub in such a house shouldn't have any noticable effect other than the click of the breaker.
Explosion upon dropping in water is not a desirable failure mode for a consumer device.
I'm all for not protecting people from themselves, but simple standards that add little cost while adding significant benefit are perfectly reasonable.
If you want to see a green mars in your lifetime, forget it. If every human on earth were to work at it full time 100 hours per week you'd still never see it.
On the other hand, we could send up a rocket to intercept a comet and cause a course correction to cause it to hit mars in 10,000 years to introduce more water to the atmosphere (do this 1000 times probably and you're getting there). We could also send up rockets with bacteria designed to start a useful ecosystem. All this we could probably do today, and in 10,000 years Mars might be habitable.
Terraforming is probably the most economical way to colonize another world. It just is also the slowest...
Sounds like when the Pope settled a huge national dispute by drawing a line on the globe and making the one half Spanish, and other half portrugese. That really worked well, didn't it...
Ultimately, the laws on mars will have to suit the colonists. If they are reasonable then people will follow them. If they're written by people who don't know the first thing about living on Mars, then everybody up there will just violate them, including the Martian police...
If you are sued over something in particular, then yes, you may have to suspend retention policies.
Usually such suspension is very specific.
Suppose you are sued by somebody claiming that you incorrectly block them from network access with invalid password errors, or something like that. Then you set up a rule to grep logs for records pertaining to login acceptance/denials in reference to that user only. You could add more if it would aid your defence (such as all users statistical denial rate - to show that they are just a bad password typer). However, unless specifically subpoenaed for a piece of info, you don't need to retain it.
Which sounds easier:
1. Writing a grep rule to retain a few key records and getting the other side to OK it and running it on all new data.
2. Taking 100GB of log data from the last 10 years, recorded in 12 different formats by 15 different authentication servers over the years, and having to promise that you thorougly searched it all to meet some demand.
If the log isn't useful to you anymore, then don't retain it...
Think of it like an apartment building - you go to the bank, get a loan, and open up an apartment building in your town. Now, you have a few paying renters, and a few people who just move in and refuse to pay rent. Do you ignore the people who aren't paying, and just focus on those that are? No! You need to rent out as many apartments as possible in order to pay back the bank loan. So you call the police and kick out the non-payers.
Bad analogy. Imagine an apartment building with no upkeep costs (just initial construction), and which has an infinite number of rooms.
Under these circumstances, they should indeed just ignore those who refuse to pay rent - as long as there is significant benefit to paying rent people will continue to do so.
The reason that deadbeats are a problem with physical property is that it is a limited resource. A deadbeat ties up a room and consumes resources without generating return. In the software world, a pirate consumes no resources at all, and doesn't tie up anything.
The fact is that Valve needs to sell x copies of HL2 to justify their investment. I doubt they'll sell many additional copies as a result of the copy protection. If anything, by making pirated copies more convenient to use, they might reduce sales...
Ok, so suppose that only 20,000 people log into the game server, and every one of those is accounted for in sales. Have they made any more money? Is it suddenly more profitable as a result?
Now, online play actually does have some cause to exclude pirated copies - after all, they incur expense in running the servers and don't get any money from the game sales. If somebody copies a game and plays it offline, the publisher doens't lose a dime (they just don't make a dime either).
Fixing the online play issue is simple - just charge 50 cents/month to use their servers, and possibly allow others to set up their own free servers (which cost Valve nothing). Then whether using a pirated copy or not, the company doesn't lose on bandwidth costs.
I doubt that copy protection leads to increased sales. I favor OSS to avoid pirating software. I do buy some softare as well. I used to shell out money every year for Turbotax. I do not shell out money for it anymore - because of their activation approach and the denial of the first-sale doctrine (it is perfectly legal to prepare a single tax return and then sell the software to a friend for half-price when you're done if you delete your install).
In any case, I'm sure somebody will have HL2 flawlessly cracked in a few weeks (if not already), and in the end all this onerous copy-protection will just end up inconveniencing the people who paid for the software. I doubt it will affect sales much at all...
If a corporation owned by 100 people equally, then the owners are pretty-much immune to any liability the company faces.
If a coproration owned by a single person is sued, in some cases the owner is fully liable. The whole idea is that in a group it is questionable whether the owners could have had any personal responsibility for what happened. If there is only one owner, then there is only one person to point a finger at.
Google for "piercing the corporate veil" for info on this legal concept. Corporations cannot be used as mere fronts for irresponsible activity by a single person. However, they do raise some barriers - as long as you don't do anything too stupid it will help.
I remember that back when I actually used RH/Mandrake it wasn't all that easy to auto-download packages either. Security updates were highly manual without a paid subscription to a service.
A possible alterior motive for this is CD sales. When software is updated, RH doesn't want you to do an emerge -uD world, they want you to go to the store and spend $30 on a CD. The business model has changed a little since then - now they want you to spend $1000 on a CD. Either way, they don't want you to just upgrade in place without spending money...
However, if I want to download a windows 3D scientific calculator that was built using VB6 and which uses.Net and DirectX 17.5, then the way it is usually distriubted is as a 100MB download which includes installers for all its dependencies. Never mind that 99% of the time most of that download is just deleted since it is already installed.
A typical linux package just contains the calculator, and indicates that it needs a VB runtime, a.Net install, and DirectX v17.5 or whatever. As a result, the linux distro is 100k - the size of the calculator.
Look at open source software for windows - you'll find dependency issues unless you download a big unified installer that installs all the necessary libraries.
Oh, and remember when that vulnerability came out which necessitated scanning your windows system for hundreds of possible vulnerable DLLs all over the filesystem? That is because windows apps tend to be self-contained. On linux there would be a single so file used by all the apps on the system, and only one file to update.
Uh, how do they get around needing four tuners? I guess if they are all claimed as being in the same home it would work, but if you asked for 10 tuners they would probably investigate.
You don't even need to run cables between houses to have that kind of a setup - or even be in the same city. Just have one person pay the bill, receive the checks, and mail the access cards around. The only downside is that you'd probably want to avoid hooking up the phone lines to get PPV - otherwise they might look at the area codes the calls are coming from and figure it out.
In any case, only one person needs to even pay the $15 - after that anybody can mirror it freely on their website - the GPL, after all, allows free redistribution under the GPL.
The GPL doesn't prevent you from selling a GPL'ed program for $50,000/license. However, what it does do is allow the first person who buys it from you to sell 10 copies at only $10,000 apiece (thereby recovering his initial investment), and the next guy from selling them for $1000, and so on until the price has dropped to free.
Uh, GPS doesn't involve pings, and GPS receivers don't transmit at all.
A GPS satellite broadcasts a continuous stream of time data. A GPS receiver looks at the time several satellites are reporting at a given instant and calcuates the pairwise difference between them all. Each generates a hyperbola on the surface of the earth, and you are located at the intersection of all.
GPS can work arbitrarily fast (up to the data rate of the timestream - perhaps not GHz, but certainly in the KHz range) - it all depends on the hardware in the receiver.
The more conventional landing system is ILS - which does not involve sonar. It is also radio-based. It involves a vertical (glidescope) and horizontal (localizer) transmitter. I'm not intimately familier with the details, but both are transmitted from directly before the runway.
In any case, GPS may or may not be suitable for landing a plane, but not for the reasons you suggest...
Well, in the past they used to have a metal bar that defined the meter. The new standard was designed to be the equivalent of the old.
I guess we could just define it as the distance light travels in 1 ms or something like that, but then we'd have to rewrite every document known to man to fit the new standard...
Indeed - from everything I've read it seems like Excel is not really the ideal tool for this stuff most of the time either. I'd either use an Access-like tool for quick-and-dirty work, or a real RDBMS and a query tool.
Most of the examples in this discussion can be readily applied against millions of data points in only a few seconds with a line of SQL that would take a query tool all of 20 seconds to design. I wouldn't recommend trying that in Excel.
I'm not saying that pivot tables don't have their uses, but rather that the people who use them might find that they like Access-like tools better once they learn how to use them...
I have to ditto that. I have coworkers (in borderline IT no less) who use Excel for almost everything. Once or twice I was handed tasks that they basically expected involved printing out spreadsheets and doing what was essentially an outer join by hand. And when they keep databases in Excel they scratch their heads and wonder why they have what amount to normalization issues with their data...
So, I just copied and pasted the whole mess into Access and did what would be a two day collation mess in about 20 minutes. (Sure, Oracle is a much nicer long-term tool, but Access is great for quick-and-dirty (emphasis on dirty).)
Some types of financial figures are fairly easy to work with in a pivot tool (there are a bunch on the market). However, if you really want to understand your data, learning SQL is invaluable!
Time spent commuting is on the WORKER'S hours, not the businesses expense.
Sort-of. If it takes 2 hours to drive to work, however, the employer is forced to raise wages to attract talent. If you pay the same rate as the small business outside the city, people will just switch jobs. Nobody leaves a job for a 10-minute-shorter commute, but when you get into the 60+ minute range, you have to pay a premium to retain talent.
Of course, business-cycle downturns which cause hundreds of thousands of workers to change employment locations tend to restart the cycle.
Except that everybody who lost their job can no longer afford their massively-inflated mortgage. If they are forced to sell they take a huge loss and end up having to declare bankruptcy.
The issue is definitely a complex one. Everybody wants to be close to work, but everybody also wants their own back yard. The two don't mix well...
You can get around the Comcast issue by just leasing a T1 line. Shouldn't cost any more than a car. On the other hand, you can save a bundle and share the line.
No different with the future car described here - why pay $30k to own a car when you can pay $1000/year to have one at your door in under a minute. (If this service were common, this could be a realistic service level.)
Also - note that your employer would probably start charging for use of the parking lot - why would he want to pay for a lot of land that nobody in the office uses except you?
People often share planes if they are recreational users - they only occupy them 1% of the time, so why pay 100% of the cost?
As long as the cars are kept clean and have standard levels of luxury it would be very attractive. I picture standard-sized bags that people can use to tote their stuff around in. Also, your car would have an mp3 library in it that knows your favorite songs when you get in, and the seats and climate control would already be set to your favorite settings before the car even arrives to pick you up. Seats would probably swivel to face each other - why face the front the whole time? Rear-facing is actually safer anyway...
1. Space isn't really the issue - if cars maintained steady speeds we wouldn't need to pack them bumper to bumpter.
2. Even if you put the mini and the truck bumper-to-bumper, the answer is that both cars swerve and brake to the best of their ability. There are a few possibilites here: a. The mini can't swerve around the kid - the kid dies. Of course, if the robot can't swerve around the kid, neither could a human driver, so that kid was toast in any case. b. The mini swerves, but the truck can't clear the kid. This is no worse than the kid just jumping in front of the truck without the mini present - again not all that different from the manual situation. c. The mini swerves, and the truck evades the kid. All other oncoming traffic stops and blasts horns so that the kid's mom comes along screaming and gets the kid off the road.
The truck would not hit the mini in any case - the mini would not brake faster than the truck could since adding to the carnage doesn't help anything.
3. If there were no robot, what would probably happen is that the mini would just run over the kid since the driver was talking on the cell phone. Alternatively the mini slams the brake and the truck jacknifes and takes out 5 cars. Even if only the kid dies, you now have lawsuits on top of lawsuits to deal with. If everything were automated, there would be no liability except for the parent who let their kids wander onto the street.
4. All of this assumes that we don't put safety perimters around the street - if we had sensors near the curb then the baby couldn't sneak up on traffic in the first place...
Well, the price of public transit wouldn't be a problem if it weren't for the 100 other drawbacks. Think about the wear on your car. If you didn't drive to work, you'd get reduced insurance rates. If you didn't drive at all you'd save a HUGE bundle on your car.
In most areas it isn't $2 per ride for regular commuters. Many cities sell monthly passes for under $100.
The cost would of course drop if anybody bothered to ride the bus.
Of course, the real problem with the bus is that it only leaves once an hour, generally requires 2-3 transfers that take you 10 miles out of your way, and generally take an hour to drive a distance that a car covers in 15 minutes...
The problem is that a huge waste stockpile probalby only produces a couple of hundred watts of heat. Ok, a really huge one might make a few kilowatts, but still, nothing to write home about. A lake of waste might be enough to heat your home - but you'd have to store it in your basement to not lose all the electricity in transmission. Volunteers?
If they want to spy on the world - they should put spy satellites in polar orbit.
If they just want to spy on their own people - they should use balloons or automated high-endurance aircraft (say 1 month aloft time - solar powered - like that NASA thing).
One satellite for the whole country would be useless for spying - too much space to monitor if you want to be really intrusive. They'd need dozens, or 100 even.
Instead, you just float a balloon for a month at a time over each populated area. You can get better resolution than the US satellites get for less cost simply because you are FAR closer to the ground, and yet you could watch the entire city from a high-enough altitude. No contending with solar radiation, and it is easy to do repairs.
Something like this was talked about to replace cell towers in low-usage areas.
The only limitation of this plan is it is only good for domestic spying. However, it would make sense to use the expensive satellites to spy on other countries, and cheaper technology to spy on yourself...
Why not just use GNUnet? It already supports this (as well as a freenet-style insert system).
A problem with this approach is that it is less anonymous. A Freenet node will tend to lose keys over time. A node which is hosting off the hard drive will not. So, all you need to do is connect to lots of nodes and send requests and get a feeling for whether these nodes have the file stored locally. If you do this over time, you can figure out who is hosting the file.
Inserting is slow and painful, but it is the strongest way to protect anonymity.
Keep in mind the goal of Freenet isn't to be the most practical anonymous network out there. It is designed to be the most anonymous network out there...
I'd be curious as to a good debate over the file hosting issue, and how easy it is to attack. Why would you want to use a P2P network where people could track your activities if they were likely to bring wrongful persecution?
This is a big deal in pharma as well. If somebody sues you for getting sick from a pill, you really want to be able to find out whether it was made by you in the first place.
I think that chemical tracers are often used for this (trace amounts of certain elements in certain ratios, or isotopes, or little (smaller than bacteria) pieces of multi-colored plastic, or whatever).
Slashdot Mirror
Toasters and baths have already been solved. It is code in most areas that all bathroom outlets have ground-fault interrupters, and these are routinely tested any time a house is bought and sold. A toaster dropped into a bathtub in such a house shouldn't have any noticable effect other than the click of the breaker.
Explosion upon dropping in water is not a desirable failure mode for a consumer device.
I'm all for not protecting people from themselves, but simple standards that add little cost while adding significant benefit are perfectly reasonable.
The expense depends on the timeframe most likely.
If you want to see a green mars in your lifetime, forget it. If every human on earth were to work at it full time 100 hours per week you'd still never see it.
On the other hand, we could send up a rocket to intercept a comet and cause a course correction to cause it to hit mars in 10,000 years to introduce more water to the atmosphere (do this 1000 times probably and you're getting there). We could also send up rockets with bacteria designed to start a useful ecosystem. All this we could probably do today, and in 10,000 years Mars might be habitable.
Terraforming is probably the most economical way to colonize another world. It just is also the slowest...
Sounds like when the Pope settled a huge national dispute by drawing a line on the globe and making the one half Spanish, and other half portrugese. That really worked well, didn't it...
Ultimately, the laws on mars will have to suit the colonists. If they are reasonable then people will follow them. If they're written by people who don't know the first thing about living on Mars, then everybody up there will just violate them, including the Martian police...
If you are sued over something in particular, then yes, you may have to suspend retention policies.
Usually such suspension is very specific.
Suppose you are sued by somebody claiming that you incorrectly block them from network access with invalid password errors, or something like that. Then you set up a rule to grep logs for records pertaining to login acceptance/denials in reference to that user only. You could add more if it would aid your defence (such as all users statistical denial rate - to show that they are just a bad password typer). However, unless specifically subpoenaed for a piece of info, you don't need to retain it.
Which sounds easier:
1. Writing a grep rule to retain a few key records and getting the other side to OK it and running it on all new data.
2. Taking 100GB of log data from the last 10 years, recorded in 12 different formats by 15 different authentication servers over the years, and having to promise that you thorougly searched it all to meet some demand.
If the log isn't useful to you anymore, then don't retain it...
And it's illegal to let a friend borrow it.
Hardly - it is only illegal to have it installed on two computers at once without specific license.
If I install at home, remove from home, install next door, remove next door, and reinstall at home, I have broken no law.
Think of it like an apartment building - you go to the bank, get a loan, and open up an apartment building in your town. Now, you have a few paying renters, and a few people who just move in and refuse to pay rent. Do you ignore the people who aren't paying, and just focus on those that are? No! You need to rent out as many apartments as possible in order to pay back the bank loan. So you call the police and kick out the non-payers.
Bad analogy. Imagine an apartment building with no upkeep costs (just initial construction), and which has an infinite number of rooms.
Under these circumstances, they should indeed just ignore those who refuse to pay rent - as long as there is significant benefit to paying rent people will continue to do so.
The reason that deadbeats are a problem with physical property is that it is a limited resource. A deadbeat ties up a room and consumes resources without generating return. In the software world, a pirate consumes no resources at all, and doesn't tie up anything.
The fact is that Valve needs to sell x copies of HL2 to justify their investment. I doubt they'll sell many additional copies as a result of the copy protection. If anything, by making pirated copies more convenient to use, they might reduce sales...
Ok, so suppose that only 20,000 people log into the game server, and every one of those is accounted for in sales. Have they made any more money? Is it suddenly more profitable as a result?
Now, online play actually does have some cause to exclude pirated copies - after all, they incur expense in running the servers and don't get any money from the game sales. If somebody copies a game and plays it offline, the publisher doens't lose a dime (they just don't make a dime either).
Fixing the online play issue is simple - just charge 50 cents/month to use their servers, and possibly allow others to set up their own free servers (which cost Valve nothing). Then whether using a pirated copy or not, the company doesn't lose on bandwidth costs.
I doubt that copy protection leads to increased sales. I favor OSS to avoid pirating software. I do buy some softare as well. I used to shell out money every year for Turbotax. I do not shell out money for it anymore - because of their activation approach and the denial of the first-sale doctrine (it is perfectly legal to prepare a single tax return and then sell the software to a friend for half-price when you're done if you delete your install).
In any case, I'm sure somebody will have HL2 flawlessly cracked in a few weeks (if not already), and in the end all this onerous copy-protection will just end up inconveniencing the people who paid for the software. I doubt it will affect sales much at all...
Even this isn't straightforward.
If a corporation owned by 100 people equally, then the owners are pretty-much immune to any liability the company faces.
If a coproration owned by a single person is sued, in some cases the owner is fully liable. The whole idea is that in a group it is questionable whether the owners could have had any personal responsibility for what happened. If there is only one owner, then there is only one person to point a finger at.
Google for "piercing the corporate veil" for info on this legal concept. Corporations cannot be used as mere fronts for irresponsible activity by a single person. However, they do raise some barriers - as long as you don't do anything too stupid it will help.
I remember that back when I actually used RH/Mandrake it wasn't all that easy to auto-download packages either. Security updates were highly manual without a paid subscription to a service.
A possible alterior motive for this is CD sales. When software is updated, RH doesn't want you to do an emerge -uD world, they want you to go to the store and spend $30 on a CD. The business model has changed a little since then - now they want you to spend $1000 on a CD. Either way, they don't want you to just upgrade in place without spending money...
However, if I want to download a windows 3D scientific calculator that was built using VB6 and which uses .Net and DirectX 17.5, then the way it is usually distriubted is as a 100MB download which includes installers for all its dependencies. Never mind that 99% of the time most of that download is just deleted since it is already installed.
.Net install, and DirectX v17.5 or whatever. As a result, the linux distro is 100k - the size of the calculator.
A typical linux package just contains the calculator, and indicates that it needs a VB runtime, a
Look at open source software for windows - you'll find dependency issues unless you download a big unified installer that installs all the necessary libraries.
Oh, and remember when that vulnerability came out which necessitated scanning your windows system for hundreds of possible vulnerable DLLs all over the filesystem? That is because windows apps tend to be self-contained. On linux there would be a single so file used by all the apps on the system, and only one file to update.
Uh, how do they get around needing four tuners? I guess if they are all claimed as being in the same home it would work, but if you asked for 10 tuners they would probably investigate.
You don't even need to run cables between houses to have that kind of a setup - or even be in the same city. Just have one person pay the bill, receive the checks, and mail the access cards around. The only downside is that you'd probably want to avoid hooking up the phone lines to get PPV - otherwise they might look at the area codes the calls are coming from and figure it out.
In any case, only one person needs to even pay the $15 - after that anybody can mirror it freely on their website - the GPL, after all, allows free redistribution under the GPL.
The GPL doesn't prevent you from selling a GPL'ed program for $50,000/license. However, what it does do is allow the first person who buys it from you to sell 10 copies at only $10,000 apiece (thereby recovering his initial investment), and the next guy from selling them for $1000, and so on until the price has dropped to free.
However, if he does change the license, then he cannot use any GPL code copyrighted by others in his project, which was his point.
Uh, GPS doesn't involve pings, and GPS receivers don't transmit at all.
A GPS satellite broadcasts a continuous stream of time data. A GPS receiver looks at the time several satellites are reporting at a given instant and calcuates the pairwise difference between them all. Each generates a hyperbola on the surface of the earth, and you are located at the intersection of all.
GPS can work arbitrarily fast (up to the data rate of the timestream - perhaps not GHz, but certainly in the KHz range) - it all depends on the hardware in the receiver.
The more conventional landing system is ILS - which does not involve sonar. It is also radio-based. It involves a vertical (glidescope) and horizontal (localizer) transmitter. I'm not intimately familier with the details, but both are transmitted from directly before the runway.
In any case, GPS may or may not be suitable for landing a plane, but not for the reasons you suggest...
Well, in the past they used to have a metal bar that defined the meter. The new standard was designed to be the equivalent of the old.
I guess we could just define it as the distance light travels in 1 ms or something like that, but then we'd have to rewrite every document known to man to fit the new standard...
Don't let zealotry dictate what tools you use.
Indeed - from everything I've read it seems like Excel is not really the ideal tool for this stuff most of the time either. I'd either use an Access-like tool for quick-and-dirty work, or a real RDBMS and a query tool.
Most of the examples in this discussion can be readily applied against millions of data points in only a few seconds with a line of SQL that would take a query tool all of 20 seconds to design. I wouldn't recommend trying that in Excel.
I'm not saying that pivot tables don't have their uses, but rather that the people who use them might find that they like Access-like tools better once they learn how to use them...
I have to ditto that. I have coworkers (in borderline IT no less) who use Excel for almost everything. Once or twice I was handed tasks that they basically expected involved printing out spreadsheets and doing what was essentially an outer join by hand. And when they keep databases in Excel they scratch their heads and wonder why they have what amount to normalization issues with their data...
So, I just copied and pasted the whole mess into Access and did what would be a two day collation mess in about 20 minutes. (Sure, Oracle is a much nicer long-term tool, but Access is great for quick-and-dirty (emphasis on dirty).)
Some types of financial figures are fairly easy to work with in a pivot tool (there are a bunch on the market). However, if you really want to understand your data, learning SQL is invaluable!
Time spent commuting is on the WORKER'S hours, not the businesses expense.
Sort-of. If it takes 2 hours to drive to work, however, the employer is forced to raise wages to attract talent. If you pay the same rate as the small business outside the city, people will just switch jobs. Nobody leaves a job for a 10-minute-shorter commute, but when you get into the 60+ minute range, you have to pay a premium to retain talent.
Of course, business-cycle downturns which cause hundreds of thousands of workers to change employment locations tend to restart the cycle.
Except that everybody who lost their job can no longer afford their massively-inflated mortgage. If they are forced to sell they take a huge loss and end up having to declare bankruptcy.
The issue is definitely a complex one. Everybody wants to be close to work, but everybody also wants their own back yard. The two don't mix well...
You can get around the Comcast issue by just leasing a T1 line. Shouldn't cost any more than a car. On the other hand, you can save a bundle and share the line.
No different with the future car described here - why pay $30k to own a car when you can pay $1000/year to have one at your door in under a minute. (If this service were common, this could be a realistic service level.)
Also - note that your employer would probably start charging for use of the parking lot - why would he want to pay for a lot of land that nobody in the office uses except you?
People often share planes if they are recreational users - they only occupy them 1% of the time, so why pay 100% of the cost?
As long as the cars are kept clean and have standard levels of luxury it would be very attractive. I picture standard-sized bags that people can use to tote their stuff around in. Also, your car would have an mp3 library in it that knows your favorite songs when you get in, and the seats and climate control would already be set to your favorite settings before the car even arrives to pick you up. Seats would probably swivel to face each other - why face the front the whole time? Rear-facing is actually safer anyway...
This system has a LOT of merit.
A couple of things:
1. Space isn't really the issue - if cars maintained steady speeds we wouldn't need to pack them bumper to bumpter.
2. Even if you put the mini and the truck bumper-to-bumper, the answer is that both cars swerve and brake to the best of their ability. There are a few possibilites here:
a. The mini can't swerve around the kid - the kid dies. Of course, if the robot can't swerve around the kid, neither could a human driver, so that kid was toast in any case.
b. The mini swerves, but the truck can't clear the kid. This is no worse than the kid just jumping in front of the truck without the mini present - again not all that different from the manual situation.
c. The mini swerves, and the truck evades the kid. All other oncoming traffic stops and blasts horns so that the kid's mom comes along screaming and gets the kid off the road.
The truck would not hit the mini in any case - the mini would not brake faster than the truck could since adding to the carnage doesn't help anything.
3. If there were no robot, what would probably happen is that the mini would just run over the kid since the driver was talking on the cell phone. Alternatively the mini slams the brake and the truck jacknifes and takes out 5 cars. Even if only the kid dies, you now have lawsuits on top of lawsuits to deal with. If everything were automated, there would be no liability except for the parent who let their kids wander onto the street.
4. All of this assumes that we don't put safety perimters around the street - if we had sensors near the curb then the baby couldn't sneak up on traffic in the first place...
Well, the price of public transit wouldn't be a problem if it weren't for the 100 other drawbacks. Think about the wear on your car. If you didn't drive to work, you'd get reduced insurance rates. If you didn't drive at all you'd save a HUGE bundle on your car.
In most areas it isn't $2 per ride for regular commuters. Many cities sell monthly passes for under $100.
The cost would of course drop if anybody bothered to ride the bus.
Of course, the real problem with the bus is that it only leaves once an hour, generally requires 2-3 transfers that take you 10 miles out of your way, and generally take an hour to drive a distance that a car covers in 15 minutes...
The problem is that a huge waste stockpile probalby only produces a couple of hundred watts of heat. Ok, a really huge one might make a few kilowatts, but still, nothing to write home about. A lake of waste might be enough to heat your home - but you'd have to store it in your basement to not lose all the electricity in transmission. Volunteers?
If they want to spy on the world - they should put spy satellites in polar orbit.
If they just want to spy on their own people - they should use balloons or automated high-endurance aircraft (say 1 month aloft time - solar powered - like that NASA thing).
One satellite for the whole country would be useless for spying - too much space to monitor if you want to be really intrusive. They'd need dozens, or 100 even.
Instead, you just float a balloon for a month at a time over each populated area. You can get better resolution than the US satellites get for less cost simply because you are FAR closer to the ground, and yet you could watch the entire city from a high-enough altitude. No contending with solar radiation, and it is easy to do repairs.
Something like this was talked about to replace cell towers in low-usage areas.
The only limitation of this plan is it is only good for domestic spying. However, it would make sense to use the expensive satellites to spy on other countries, and cheaper technology to spy on yourself...
Why not just use GNUnet? It already supports this (as well as a freenet-style insert system).
A problem with this approach is that it is less anonymous. A Freenet node will tend to lose keys over time. A node which is hosting off the hard drive will not. So, all you need to do is connect to lots of nodes and send requests and get a feeling for whether these nodes have the file stored locally. If you do this over time, you can figure out who is hosting the file.
Inserting is slow and painful, but it is the strongest way to protect anonymity.
Keep in mind the goal of Freenet isn't to be the most practical anonymous network out there. It is designed to be the most anonymous network out there...
I'd be curious as to a good debate over the file hosting issue, and how easy it is to attack. Why would you want to use a P2P network where people could track your activities if they were likely to bring wrongful persecution?
This is a big deal in pharma as well. If somebody sues you for getting sick from a pill, you really want to be able to find out whether it was made by you in the first place.
I think that chemical tracers are often used for this (trace amounts of certain elements in certain ratios, or isotopes, or little (smaller than bacteria) pieces of multi-colored plastic, or whatever).