Why is it a bad thing to *limit* the number of countries which have the ability to cause such destruction? Especially in the case where the major countries that *have* such weapons have shown great restraint for nearly as long as the weapons have existed.
When a country has nuclear weapons, the US stops meddling in its internal affairs and begins to treat it as an equal.
The process of deriving the ideas used to build programs is science. Applying those ideas to actually construct programs is engineering. The programs are technology.
science |sns|
noun
the intellectual and practical activity encompassing the systematic study of the structure and behavior of the physical and natural world through observation and experiment
Information and computation are part of the physical and natural world. As we delve into the nature of subatomic particles, many are beginning to theorize that, in fact, reality is constructed of patterns, of data. Computer science, at heart, is the study of data and data transformation.
Also, one branch of computer science, artificial intelligence, is working to understand the structure and behavior of the most amazing part of nature... the brain. How does thinking work? And what is its essence, decoupled from the physical structure in which we see it? We don't know the answers to those questions, but it is computer science that will provide them.
Disclaimer: I'm a physicist, which makes me a REAL scientist.
I doubt that. I know a lot of physicists, and none of them are as blind and narrow-minded as you appear to be.
I'm not keen on dying in a car accident as the result of a software glitch--because no manual override was included, because ON AVERAGE the software does much better than humans.
But you are keen on dying in a car accident as a result of driver fatigue, distraction, lack of skill or physical glitch (say, the guy in the oncoming lane has a heart attack), because ON AVERAGE human drivers do much worse than the software?
It's all a question of odds, and not choosing the option that maximizes the odds is stupid.
By "keep going straight", I meant more "stay on the road in the current lane" than "generally go straight ahead." The latter we have today (assuming your car is properly aligned and you don't veer onto the sidewalk). The former would be a step towards self-driving cars.
There are several cars on the market that already do lane following. Have been for a few years now. So far they all try to require you to at least keep your hands on the wheel, in an effort to try to make you pay attention, but it turns out that at least some are pretty easy to fool: http://www.roadandtrack.com/ca...
So, yes, we already have this, though we try not to.
And without calling Sir Issac Newton a liar, a bullet imparts significantly more energy onto the recipient than the shooter.
Nope, to say that you have to call Sir Isaac Newton a liar. A bullet imparts significantly less energy onto the recipient than the shooter. Recoil springs and slide rails don't absorb any of that energy, they just spread it over a longer period of time. Large muzzle brakes actually can some of the energy in the direction, but those only exist on very large-caliber weapons (mostly.50 BMG).
Number three,and to a lesser extent two, are what people are referring to with the term "Stopping power". If a bullet is said to have more stopping power, they usually mean hydroshock temporarily interrupting nervous system function. It isn't just for the central nervous system however, it works everywhere. Think of it this way, have you ever been struck so hard or hit something so hard that part of your body went numb? Imagine that feeling applied with an order of magnitude more force through a bullet hit.
Actually, a bullet strike generally carries much less energy than many other forms of impact that you might receive, and be stunned by.
By your own admission stopping power isn't a myth, just firearms jargon you were not fully aware of.
It's a myth in handguns. And actually pretty rare even in rifles.
The only practical solution currently is rubber bullets. The cops get to keep the ease of use and most of the stopping power of a gun but the lethality levels go way down.
"Stopping power" is a myth -- a rather obvious one if you think about the physics. The bullet can't carry any more energy than is imparted on the shooter, and actually carries less.
People who are shot stop for one of four reasons.
1. People stop because they know they're supposed to fall down when they get shot. That is, the bullet doesn't actually do any incapacitating damage, but they fall down anyway. Rubber bullets might be able to do this, but it doesn't work on everyone, and once everyone knows the police are carrying rubber bullets, it will work on even fewer people.
2. People stop because the bullet did structural damage that prevents them from being able to move. Mostly this means broken bones in strategic places. For example, if a bullet shatters an ankle or a knee, you're going to have a hard time walking. If a bullet shatters your pelvis, you will be completely unable to even stand. Rubber bullets can't do this reliably, and might not be able to do it at all.
3. People stop because the bullet severely traumatized their central nervous system. Shoot someone in the head and they'll (usually) switch off like a light. Rubber bullets might be able to do this, sometimes (e.g. penetrating through an eye socket, or through a thinned area of the skull), but not non-lethally. Very high-powered rounds can also achieve the same instant lights-out effect through hydrostatic shock. A large-caliber rifle round to the upper chest, for example, might not do lethal damage (assuming treatment is quickly available), but might generate a hydrostatic shock wave that slams into brain and/or brain stem with enough force to temporarily disable the target. Handguns cannot do this, it requires enough energy that it's really only feasible to get from a weapon with considerable recoil, enough that you almost certainly need a stock to transmit the recoil to the shooter's torso. Rubber bullets carrying that much energy would probably penetrate, assuming it was even feasible for police to regularly carry high-powered hunting rifles (note that mid-energy weapons like AR-15s can't do it).
4. People stop because they black out from blood loss. This is the primary goal of shooting someone with a handgun, to create a hole (or, more likely, holes) that open up large blood vessels, causing the target's blood pressure to drop dramatically, reducing blood flow to the brain and causing a blackout. It's most reliably achieved by several bullets into center mass, into the large mass of organs and blood vessels in the torso. Rubber bullets can't do this. And if they could they'd be no less lethal than lead bullets.
So, no, rubber bullets do not provide "stopping power". They're useful for harassing people who are willing to run away when faced with painful bruises, and they have the advantage (to the police) that the bruises can be delivered from a distance. But against someone you really want to stop? No way.
EVs won't discharge power through their charging ports. The charging ports aren't just dumb connections.
So you are saying that this requires modifications to the vehicle itself? Because that makes it sound like the car requires no modifications at all.
That requires cryptographic authentication from the interface. I suppose if you could extract the key from one of those, you might be able to make a fake charger. Or you could buy one of those units and put a fake front on it.
In this instance all the profit went to the employees, thus instead of being taxed at the corporate level it was taxed at the (higher!) individual level. So what's the difference?
The difference is that the taxes were collected directly from voters who were able to see exactly how much they paid, rather than from a non-voting entity who would pass the tax to voters invisibly. Corporate taxation is all about hiding taxes from the people who pay them, so it's a problem when the taxes don't get appropriately hidden. In order to keep attention from being paid to the actual taxes collected, therefore, those who wish to hide taxes from taxpayers moralize about how corporations aren't paying their fair share, thus encouraging the voters to push for better hiding of the taxes from themselves.
It's a pretty incredible setup when you realize what's going on. People aggressively demand to pay more taxes with less visibility and accountability.
An electric with a 1000 mil range could be charged entirely at home
Sure it could. But if you charge at home, you're paying for the electricity. Charging at work is free (for the car owner, that is). If you have a choice of paying to charge your car or charging it at no cost to you, which are you going to choose?
Meh. Electricity is cheap. Sure it's better to charge on someone else's dime if you can, but it costs far less per mile than gasoline so it's really not that important an issue.
Wire them up so the "FREE" charger discharges the battery of anyone who plugs into it while feeding the power to the "Out of Order" charger your own electric car is plugged into.
EVs won't discharge power through their charging ports. The charging ports aren't just dumb connections.
Now I feel like installing fake charging ports just to fuck with assholes like you.
FYI, in order to make that work your fake charging station would have to actually charge cars. Otherwise car owners would realize as soon as they connected that it wasn't working. Even if you charged their cars for a few minutes, then shut off the juice, you still wouldn't fool them for long. EV owners use smartphone apps to exchange information about the chargers around, and yours would quickly get flagged as broken.
Regarding the AC's comment, it is pretty rude for ICEVs to park in charging spots. Not that keying their cars is at all an appropriate response. I usually just leave a note on their windshield, pointing out (nicely) that they're just like the guy who parks in front of the gas pump at a busy station, and then goes into the convenience store to shop and eat lunch -- but worse, because odds are there are enough gas pumps and stations so that you can actually get to one. Charging stations tend to be much harder to find in many areas.
Those spots are a convenience for employees who don't want to schedule a stop at a charging station going to/from work. And as is pretty common, that type of thing goes from being a perk to an Entitlement pretty quickly.
You don't drive an EV, I see.
Scheduling a stop at a charging station going to/from work is impractical. Charging an EV isn't like filling a gasoline tank; it's not a five-minute operation. You really wouldn't want to make an hour-long stop on your way to or from the office every day, would you? This isn't an issue of entitlement, it's a question of whether or not an EV with a given range will even work for your commute.
EVs need to be charged in places where the car is going to be parked for a substantial period of time anyway. This means home and/or office. If you live close enough, or have an EV with enough range, to make the round trip without charging at the office, then you don't need to charge at the office (and any desire to do so does fall into the perk/entitlement category). If you can't make the round trip, then you need one of three things: (1) the ability to charge at the office, (2) an EV with more range or (3) an ICEV. If you buy a short-range EV, counting on charging at the office, then it's not a perk or an entitlement issue; you actually need to charge at work. Or accept that the short-range EV doesn't work for your commute and get something else.
Where it gets ugly is in the gray area between "Can always charge at work" and "Can't charge at work", because it becomes "Can sometimes get home in a reasonable amount of time". Or, if the area doesn't have any other charging infrastructure, "Can sometimes get home".
Despite appearances, a charging station isn't a parking spot with a plug for your car. It's a spot at a gas pump that takes half an hour to use. And that's the real challenge with electric cars...not range, not cost.
No, the need for charging at work is driven by the range/cost issue. Make EVs with substantial range affordable and the buyers won't need to charge at the office.
When I worked in an office that had charging stations, the employees with LEAFs and iMievs used them. The employees with Teslas did not, because they could get home without it.
Also, note that recharging doesn't take a half hour, it takes multiple hours. Typical workplace, etc., startions are 220W, 30A level 2 ports. My LEAF takes just shy of four hours to recharge from empty on a level 2 station. Of course most people don't arrive to work with their battery flat empty, so it's going to be less than that, but figure 1-2 hours. That's with a car with a 6kW charger in it; some cars have 3kW chargers, so double it.
What I want in my future hypothetical EV is a port which will lock itself until I have enough charge to get home and/or do whatever else I've programmed the car to think I'll be doing for the rest of the day, then unlock itself so that if someone else really needs a charge they can get one.
My 2013 LEAF has part of that. It will lock the charging port until the battery is full, where "full" can be either 80% or 100%.
Or another alternate headline: "Rich people fight over free lunches"
It's not so much about the free electricity. I think most would be fine with paying for the juice. It's about being able to get home.
When I worked in an office with chargers, though, it really wasn't a problem. We just set up a mailing list that everyone with a short-range EV subscribed to, and used it to communicate about vehicle swapping on the chargers. Unplugging someone else's car would have been considered very uncool, and no one ever did that. Those who lived close or had longer-ranged cars (i.e. Teslas), just didn't charge at the office, with the exception of one Tesla owner who didn't have a place to charge at home.
I notice that my 2013 LEAF (unlikely my 2011 LEAF) has a charging port lock, which can be set to lock the charger in until the car is fully charged. I've never had the need to use it, but I can see the purpose. What would be even better is if you could set it to unlock at a specific charge level. Then people could set it to unlock as soon as they have enough juice to get home.
So now master is the development branch! master is the release branch! THAT is terrifying. Although they do tag releases, but still.
Developing on and releasing from master has its risks, but given appropriate QA, including code reviews, extensive automated unit, functional and integration tests, and extensive release tests, it can work very well. That's what Google does. 25,000 engineers, one source repository, 45,000 commits per day, developing on and releasing from HEAD.
Well, almost. Developers create local branches for their work and don't commit into master until code review is complete -- including of automated tests. The actual commit into master doesn't go in unless the commit and everything else that could possibly depend on it builds and passes all of the tests (the build/test/submit cycle is automated; engineers kick it off and then get informed of the results). Releases are branched off to freeze them while release testing is done, and sometimes a few commits are cherry-picked into a release to fix issues, but mostly the release either passes the tests and goes out, or fails the tests and is abandoned. Most projects operate on a weekly release cycle, so the impact of abandoning a release is small. As long as it doesn't happen too often.
Note that I'm speaking of the web properties; search, Gmail, etc. Obviously other groups have different approaches. For example, I currently work on Android, which has a roughly annual release cycle. That drives a very different strategy. One with lots of branching, actually.
Also note that I'm not claiming that this is a good strategy for every team or company. I'm just pointing out that it can work, if you manage it well. Of course, the same is true of virtually every development process, though different processes are better suited to different contexts.
Slashdot Mirror
Why is it a bad thing to *limit* the number of countries which have the ability to cause such destruction? Especially in the case where the major countries that *have* such weapons have shown great restraint for nearly as long as the weapons have existed.
When a country has nuclear weapons, the US stops meddling in its internal affairs and begins to treat it as an equal.
There are counterexamples both directions.
But are computer programs science, or technology?
The process of deriving the ideas used to build programs is science. Applying those ideas to actually construct programs is engineering. The programs are technology.
science |sns| noun the intellectual and practical activity encompassing the systematic study of the structure and behavior of the physical and natural world through observation and experiment
Information and computation are part of the physical and natural world. As we delve into the nature of subatomic particles, many are beginning to theorize that, in fact, reality is constructed of patterns, of data. Computer science, at heart, is the study of data and data transformation.
Also, one branch of computer science, artificial intelligence, is working to understand the structure and behavior of the most amazing part of nature... the brain. How does thinking work? And what is its essence, decoupled from the physical structure in which we see it? We don't know the answers to those questions, but it is computer science that will provide them.
Disclaimer: I'm a physicist, which makes me a REAL scientist.
I doubt that. I know a lot of physicists, and none of them are as blind and narrow-minded as you appear to be.
I'm not keen on dying in a car accident as the result of a software glitch--because no manual override was included, because ON AVERAGE the software does much better than humans.
But you are keen on dying in a car accident as a result of driver fatigue, distraction, lack of skill or physical glitch (say, the guy in the oncoming lane has a heart attack), because ON AVERAGE human drivers do much worse than the software?
It's all a question of odds, and not choosing the option that maximizes the odds is stupid.
By "keep going straight", I meant more "stay on the road in the current lane" than "generally go straight ahead." The latter we have today (assuming your car is properly aligned and you don't veer onto the sidewalk). The former would be a step towards self-driving cars.
There are several cars on the market that already do lane following. Have been for a few years now. So far they all try to require you to at least keep your hands on the wheel, in an effort to try to make you pay attention, but it turns out that at least some are pretty easy to fool: http://www.roadandtrack.com/ca...
So, yes, we already have this, though we try not to.
Guns loaded with rubber bullets accomplish those objectives as has been proven countless times during riot duty.
Rifle-fired volleys of rubber bullets, yes. Good luck convincing any police officer to load his personal defense handgun with them.
You seem to have trouble with reading comprehension. Either that or you like making strawman arguments.
You seem to be arguing against a point I never made. Please re-read the thread.
Nothing free can be thought of as 'reliable'. It only takes one more cow orker to buy an EV any reliable isn't.
In my experience -- of actually doing this -- that's not true.
What if I started keying EVs because they weren't paying their fair share of road tax?
I'm not advocating keying anyone's vehicle. As for the taxes, if you don't like it, talk to your representatives.
5. Adequate pain to convince them to stop what they are otherwise trying to do.
That's the same as #1, more or less. It also doesn't really work because traumatic injury is generally not painful right away.
And without calling Sir Issac Newton a liar, a bullet imparts significantly more energy onto the recipient than the shooter.
Nope, to say that you have to call Sir Isaac Newton a liar. A bullet imparts significantly less energy onto the recipient than the shooter. Recoil springs and slide rails don't absorb any of that energy, they just spread it over a longer period of time. Large muzzle brakes actually can some of the energy in the direction, but those only exist on very large-caliber weapons (mostly .50 BMG).
Number three,and to a lesser extent two, are what people are referring to with the term "Stopping power". If a bullet is said to have more stopping power, they usually mean hydroshock temporarily interrupting nervous system function. It isn't just for the central nervous system however, it works everywhere. Think of it this way, have you ever been struck so hard or hit something so hard that part of your body went numb? Imagine that feeling applied with an order of magnitude more force through a bullet hit.
Actually, a bullet strike generally carries much less energy than many other forms of impact that you might receive, and be stunned by.
By your own admission stopping power isn't a myth, just firearms jargon you were not fully aware of.
It's a myth in handguns. And actually pretty rare even in rifles.
If someone can't drive to and from work with a single charge in their EV, then isn't an EV a rather terrible choice of vehicle for them?
Not if they can reliably charge their EV at work.
The only practical solution currently is rubber bullets. The cops get to keep the ease of use and most of the stopping power of a gun but the lethality levels go way down.
"Stopping power" is a myth -- a rather obvious one if you think about the physics. The bullet can't carry any more energy than is imparted on the shooter, and actually carries less.
People who are shot stop for one of four reasons.
1. People stop because they know they're supposed to fall down when they get shot. That is, the bullet doesn't actually do any incapacitating damage, but they fall down anyway. Rubber bullets might be able to do this, but it doesn't work on everyone, and once everyone knows the police are carrying rubber bullets, it will work on even fewer people.
2. People stop because the bullet did structural damage that prevents them from being able to move. Mostly this means broken bones in strategic places. For example, if a bullet shatters an ankle or a knee, you're going to have a hard time walking. If a bullet shatters your pelvis, you will be completely unable to even stand. Rubber bullets can't do this reliably, and might not be able to do it at all.
3. People stop because the bullet severely traumatized their central nervous system. Shoot someone in the head and they'll (usually) switch off like a light. Rubber bullets might be able to do this, sometimes (e.g. penetrating through an eye socket, or through a thinned area of the skull), but not non-lethally. Very high-powered rounds can also achieve the same instant lights-out effect through hydrostatic shock. A large-caliber rifle round to the upper chest, for example, might not do lethal damage (assuming treatment is quickly available), but might generate a hydrostatic shock wave that slams into brain and/or brain stem with enough force to temporarily disable the target. Handguns cannot do this, it requires enough energy that it's really only feasible to get from a weapon with considerable recoil, enough that you almost certainly need a stock to transmit the recoil to the shooter's torso. Rubber bullets carrying that much energy would probably penetrate, assuming it was even feasible for police to regularly carry high-powered hunting rifles (note that mid-energy weapons like AR-15s can't do it).
4. People stop because they black out from blood loss. This is the primary goal of shooting someone with a handgun, to create a hole (or, more likely, holes) that open up large blood vessels, causing the target's blood pressure to drop dramatically, reducing blood flow to the brain and causing a blackout. It's most reliably achieved by several bullets into center mass, into the large mass of organs and blood vessels in the torso. Rubber bullets can't do this. And if they could they'd be no less lethal than lead bullets.
So, no, rubber bullets do not provide "stopping power". They're useful for harassing people who are willing to run away when faced with painful bruises, and they have the advantage (to the police) that the bruises can be delivered from a distance. But against someone you really want to stop? No way.
EVs won't discharge power through their charging ports. The charging ports aren't just dumb connections.
So you are saying that this requires modifications to the vehicle itself? Because that makes it sound like the car requires no modifications at all.
That requires cryptographic authentication from the interface. I suppose if you could extract the key from one of those, you might be able to make a fake charger. Or you could buy one of those units and put a fake front on it.
In this instance all the profit went to the employees, thus instead of being taxed at the corporate level it was taxed at the (higher!) individual level. So what's the difference?
The difference is that the taxes were collected directly from voters who were able to see exactly how much they paid, rather than from a non-voting entity who would pass the tax to voters invisibly. Corporate taxation is all about hiding taxes from the people who pay them, so it's a problem when the taxes don't get appropriately hidden. In order to keep attention from being paid to the actual taxes collected, therefore, those who wish to hide taxes from taxpayers moralize about how corporations aren't paying their fair share, thus encouraging the voters to push for better hiding of the taxes from themselves.
It's a pretty incredible setup when you realize what's going on. People aggressively demand to pay more taxes with less visibility and accountability.
An electric with a 1000 mil range could be charged entirely at home
Sure it could. But if you charge at home, you're paying for the electricity. Charging at work is free (for the car owner, that is). If you have a choice of paying to charge your car or charging it at no cost to you, which are you going to choose?
Meh. Electricity is cheap. Sure it's better to charge on someone else's dime if you can, but it costs far less per mile than gasoline so it's really not that important an issue.
Wire them up so the "FREE" charger discharges the battery of anyone who plugs into it while feeding the power to the "Out of Order" charger your own electric car is plugged into.
EVs won't discharge power through their charging ports. The charging ports aren't just dumb connections.
Now I feel like installing fake charging ports just to fuck with assholes like you.
FYI, in order to make that work your fake charging station would have to actually charge cars. Otherwise car owners would realize as soon as they connected that it wasn't working. Even if you charged their cars for a few minutes, then shut off the juice, you still wouldn't fool them for long. EV owners use smartphone apps to exchange information about the chargers around, and yours would quickly get flagged as broken.
Regarding the AC's comment, it is pretty rude for ICEVs to park in charging spots. Not that keying their cars is at all an appropriate response. I usually just leave a note on their windshield, pointing out (nicely) that they're just like the guy who parks in front of the gas pump at a busy station, and then goes into the convenience store to shop and eat lunch -- but worse, because odds are there are enough gas pumps and stations so that you can actually get to one. Charging stations tend to be much harder to find in many areas.
Those spots are a convenience for employees who don't want to schedule a stop at a charging station going to/from work. And as is pretty common, that type of thing goes from being a perk to an Entitlement pretty quickly.
You don't drive an EV, I see.
Scheduling a stop at a charging station going to/from work is impractical. Charging an EV isn't like filling a gasoline tank; it's not a five-minute operation. You really wouldn't want to make an hour-long stop on your way to or from the office every day, would you? This isn't an issue of entitlement, it's a question of whether or not an EV with a given range will even work for your commute.
EVs need to be charged in places where the car is going to be parked for a substantial period of time anyway. This means home and/or office. If you live close enough, or have an EV with enough range, to make the round trip without charging at the office, then you don't need to charge at the office (and any desire to do so does fall into the perk/entitlement category). If you can't make the round trip, then you need one of three things: (1) the ability to charge at the office, (2) an EV with more range or (3) an ICEV. If you buy a short-range EV, counting on charging at the office, then it's not a perk or an entitlement issue; you actually need to charge at work. Or accept that the short-range EV doesn't work for your commute and get something else.
Where it gets ugly is in the gray area between "Can always charge at work" and "Can't charge at work", because it becomes "Can sometimes get home in a reasonable amount of time". Or, if the area doesn't have any other charging infrastructure, "Can sometimes get home".
Despite appearances, a charging station isn't a parking spot with a plug for your car. It's a spot at a gas pump that takes half an hour to use. And that's the real challenge with electric cars...not range, not cost.
No, the need for charging at work is driven by the range/cost issue. Make EVs with substantial range affordable and the buyers won't need to charge at the office.
When I worked in an office that had charging stations, the employees with LEAFs and iMievs used them. The employees with Teslas did not, because they could get home without it.
Also, note that recharging doesn't take a half hour, it takes multiple hours. Typical workplace, etc., startions are 220W, 30A level 2 ports. My LEAF takes just shy of four hours to recharge from empty on a level 2 station. Of course most people don't arrive to work with their battery flat empty, so it's going to be less than that, but figure 1-2 hours. That's with a car with a 6kW charger in it; some cars have 3kW chargers, so double it.
What I want in my future hypothetical EV is a port which will lock itself until I have enough charge to get home and/or do whatever else I've programmed the car to think I'll be doing for the rest of the day, then unlock itself so that if someone else really needs a charge they can get one.
My 2013 LEAF has part of that. It will lock the charging port until the battery is full, where "full" can be either 80% or 100%.
Or another alternate headline: "Rich people fight over free lunches"
It's not so much about the free electricity. I think most would be fine with paying for the juice. It's about being able to get home.
When I worked in an office with chargers, though, it really wasn't a problem. We just set up a mailing list that everyone with a short-range EV subscribed to, and used it to communicate about vehicle swapping on the chargers. Unplugging someone else's car would have been considered very uncool, and no one ever did that. Those who lived close or had longer-ranged cars (i.e. Teslas), just didn't charge at the office, with the exception of one Tesla owner who didn't have a place to charge at home.
I notice that my 2013 LEAF (unlikely my 2011 LEAF) has a charging port lock, which can be set to lock the charger in until the car is fully charged. I've never had the need to use it, but I can see the purpose. What would be even better is if you could set it to unlock at a specific charge level. Then people could set it to unlock as soon as they have enough juice to get home.
So now master is the development branch! master is the release branch! THAT is terrifying. Although they do tag releases, but still.
Developing on and releasing from master has its risks, but given appropriate QA, including code reviews, extensive automated unit, functional and integration tests, and extensive release tests, it can work very well. That's what Google does. 25,000 engineers, one source repository, 45,000 commits per day, developing on and releasing from HEAD.
Well, almost. Developers create local branches for their work and don't commit into master until code review is complete -- including of automated tests. The actual commit into master doesn't go in unless the commit and everything else that could possibly depend on it builds and passes all of the tests (the build/test/submit cycle is automated; engineers kick it off and then get informed of the results). Releases are branched off to freeze them while release testing is done, and sometimes a few commits are cherry-picked into a release to fix issues, but mostly the release either passes the tests and goes out, or fails the tests and is abandoned. Most projects operate on a weekly release cycle, so the impact of abandoning a release is small. As long as it doesn't happen too often.
Note that I'm speaking of the web properties; search, Gmail, etc. Obviously other groups have different approaches. For example, I currently work on Android, which has a roughly annual release cycle. That drives a very different strategy. One with lots of branching, actually.
Also note that I'm not claiming that this is a good strategy for every team or company. I'm just pointing out that it can work, if you manage it well. Of course, the same is true of virtually every development process, though different processes are better suited to different contexts.