The right thing is to gracefully introduce the new capability. TSYNC isn't at all a bad thing, that's why Linus let it go in. The correct thing in userspace is to use the feature if it is there and do it the old way if it is not. Then, later on when the new kernel feature becomes universal through natural upgrade cycles you can remove the fallback.
The correct solution is the one that best solves your use case. If Google's use case for chromium isn't supporting Debian, then there is really nothing wrong with their approach. Presumably their main linux target for chromium is chromiumos, and I'm sure it works just fine there.
It only hit the kernel a few months ago, don't be surprised if every single process on Linux is using it.
Any other applications that adopt it will likely do so in a much more sane manner, though: use it if it is present and function without it if it isn't. Having a hard dependency on an optional feature that's only a few months old is kind of an insane approach. Who programs like that?
Somebody whose main linux distro target is one they publish themselves?
Sure, but that's not really the argument, is it? I thought it was. Because with a tracking/aiming system every weapon can hit anything which is not fast enough to escape.
If the target is maneuvering, the speed of the projectile makes a HUGE difference. If your target is 20 meters away, it is much easier to shoot it with a gun than to hit it with a water balloon, even if the balloon is thrown perfectly. Lasers travel at the speed of light, making target maneuvers almost negligible. Granted, at a distance like my geosync orbit example, a maneuvering target could evade attack unless you can direct enough energy in its general vicinity to make maneuvers unimportant (ie fire many beams over an area of space).
Lasers basically push the potential engagement ranges MUCH further out.
The application dropped support for production kernels... because it wants a patch that isn't yet in production kernels.
The feature is in several stable kernel branches. Your distro might just not support them, so either don't use Chrome, or don't use that distro, or figure out how to use a newer kernel on your distro.:)
So "The Google Way" is the correct way, and every other application in Linux is insecure?
Security isn't a binary thing. Google's way is more secure, and I'm sure other applications will start using it. It only hit the kernel a few months ago, don't be surprised if every single process on Linux is using it.
2. Vendor doesn't seem to care about kernel/userspace boundaries and very happily writes code on both sides to an interface they've designed themselves, for themselves.
So, nothing ends up in Linux unless Linus lets it in, and it isn't like this was snuck in.
Really, Google is doing the right thing here. Instead of trying to hack inter-process security in on the userspace side, they're extending the kernel to improve things. Inter-process communications/relationships/etc is one of the things the kernel is supposed to do.
This is no more a violation of boundaries than putting modesetting in the kernel instead of running X11 as root and having it set device registers and such.
That's what airburst grenades are for.:) I'm not saying the laser is the solution to EVERY problem. I don't really see them as being replacements for hand-held rifles. This is something that would be combined with an automated control system.
Also, when somebody else throws a grenade at your wall, then your laser turret can blast it out of the air.
They should do some tests to see how many seconds a 120mm shell fired from an M1A1 at ~1700m/s needs to be in contact with the truck to disable it.
Perhaps a more important question is how long the laser has to be pointed at the 120mm shell in-flight from the M1A1 at 1700m/s before the shell does nothing on contact with the target.
Blowing up targets sitting still on a range has been a solved problem for centuries. That doesn't mean that an Aegis air defense system has no more value than a medieval cannon.
0.1% of the kill rate of an A10 Warthog, for only 1000 times the cost. But hey, somebody's making a lot of money here, so we can't complain.
Speed-of-light weapons that have virtually linear trajectories certainly have the potential to change warfare though.
Put it this way - if you pointed that A10 cannon upwards at an aircraft at 70k feet, you'd be hard-pressed to hit it at all. On the other hand, a laser would have relatively little difficulty hitting the aircraft even if it were in geosync orbit, or even on the surface of the moon.
Stick something like this on a plane and you could use it to shoot down incoming missiles, shoot artillery shells in mid-flight, shoot aircraft, and so on.
Sure, the technology is immature, but it certainly is a capability that is valuable for a military to posses.
Politicians don't listen to grassroots activists, they listen to $100,000 contributors.
Yup. This is largely due to the fact that voters don't listen to grassroots activists either - they do whatever the ads on TV tell them to do, maybe augmented by whatever the union boss or preacher down the street tells them.
Of course, a fair and impartial trial will also require him to accept a very high likeliness of losing the case, based on the current evidence against him.
Yup. When the law is immoral, only the immoral are innocent of breaking the law.
Whether Android does this I have no idea, but the device could be configured to power off if the wrong screen PIN was entered too many times. A FDE password has to withstand offline attack, which means unlimited attempts at a high rate.
It is completely appropriate to use a different level of security for each.
It's depressing that morons are willing to do so much damage for so little gain.
Here is the thing - they're not morons - they're sociopaths. The morons are the rest of us for not making simple legal changes to prevent this nonsense, largely because the beneficiaries of this theft want it to continue.
The answer to fixing this problem is to require scrap metal dealers to be licensed (with strong penalties for anyone who isn't) and to require all transactions to be recorded along with the ID of the seller. Its already done in many jurisdictions for pawn shops (where you need a license to operate one and where sellers have to provide ID when they sell it, why should scrap merchants be any different.
They should also have a delay before payout, with appointments required for the time to pick up the money. I imagine that even a one week delay would be enough to allow utilities to track down sales of stolen goods, and then the police can be waiting when the guy shows up to collect.
Stuff like this isn't going to be a problem for anybody legit. Electricians aren't going to trips to the dealer to sell one coil of wire, and be desperate for their $50.
In the event of a global collapse, these people will simply carry on as before.
If civilization collapses, there will be a reason that it collapsed. Such as a pandemic disease, crop destroying volcanic eruption, asteroid impact, nuclear winter, or runaway greenhouse effect. In any of these events, Africans will not "carry on as before". They will be the hardest hit, because they have nothing to fall back on.
Agree with this. The middle of Africa won't care if there are worldwide blackouts, as long as everybody has their plot of land to grow crops on. That kind of disaster will be very hard on the industrial world when you can't get food into your cities. On the other hand if the problem is that there are more bodies to feed than local land to feed them on, then the people of Africa will have a real problem on their hand, and will probably solve it by killing each other off until it is no longer a problem, since historically that is what tends to happen in these situations anywhere.
At the general election level; yes, you'd need a constitutional amendment. At the party primary level, however, such a system as you describe would be incredibly helpful, and probably for all parties.
How? A proportional representation system only works if you have multiple people elected into an office for any particular constituent.
The Republicans can't nominate 100 different people for the 3rd congressional district in Florida. Well, I guess they could, but then the first past the post system would practically guarantee that none would be elected, and that is why they only nominate one.
In a proportional system the Republicans would offer a prioritized list of 435 representatives for the entirety of the US House, and then their members could use proportional voting to decide who ends up in what slot. Then in the general election the Republicans would be awarded some number of seats, and the top n candidates would take office.
The president would still lead to the usual deadlock if separately elected since it is one man in one office. The only real solution to that is to replace the office with a prime minister, which is of course how most democracies handle the situation. That would require an amendment as well, and probably a rewrite of half the constitution.
Well, the real problem is a first-past-the-post election system, combined with only having one candidate elected to any particular office for any particular voter. This happens both at the general and party levels.
If you allowed each voter to elect a few hundred candidates, out of a much larger pool, using a proportional system, then you'd get a legislative body that better reflected the diversity of political views.
But, that would also require a constitutional amendment that serves neither of the parties in power, so you'll never see it happen.
Yes, let's not have any rules or oversight on "people" who were born in a lawyer's office, can potentially live forever, are motivated purely by greed, and will gladly break the law when it suits them. What could possibly go wrong?
That is a ridiculous assertion. There is no way they'll break the law, simply by virtue of the fact that without rules and oversight there won't be any laws for them to break.
Now get out there and sign some contracts of adhesion to help out the economy so that the big companies can spend more on lobbyists!
Ultimately, it is the immune system that keeps your body free of cancer. Cancers happen frequently in your body, and the immune system beats them down. When it fails at that for some reason, only then does clinical disease happen.
You could make the same argument about MANY functions within your body. There are many different mechanisms in cells that work to prevent uncontrolled cell growth. There are many controls on the cell cycle which have to fail. The tumor needs to create demand for more blood vessels to sustain growth. The tumor has to evade the immune system, etc.
It seems like cancer is the flip-side of multicellular life, and as a result humans have many different mechanisms for preventing it. Only when they all fail do cancers form. The problem is that you have billions of cells independently dividing and every time they do there is an opportunity for more mutations to creep in. Sooner or later any level of redundancy is bound to fail.
Slashdot Mirror
The right thing is to gracefully introduce the new capability. TSYNC isn't at all a bad thing, that's why Linus let it go in. The correct thing in userspace is to use the feature if it is there and do it the old way if it is not. Then, later on when the new kernel feature becomes universal through natural upgrade cycles you can remove the fallback.
The correct solution is the one that best solves your use case. If Google's use case for chromium isn't supporting Debian, then there is really nothing wrong with their approach. Presumably their main linux target for chromium is chromiumos, and I'm sure it works just fine there.
It only hit the kernel a few months ago, don't be surprised if every single process on Linux is using it.
Any other applications that adopt it will likely do so in a much more sane manner, though: use it if it is present and function without it if it isn't. Having a hard dependency on an optional feature that's only a few months old is kind of an insane approach. Who programs like that?
Somebody whose main linux distro target is one they publish themselves?
Sure, but that's not really the argument, is it?
I thought it was.
Because with a tracking/aiming system every weapon can hit anything which is not fast enough to escape.
If the target is maneuvering, the speed of the projectile makes a HUGE difference. If your target is 20 meters away, it is much easier to shoot it with a gun than to hit it with a water balloon, even if the balloon is thrown perfectly. Lasers travel at the speed of light, making target maneuvers almost negligible. Granted, at a distance like my geosync orbit example, a maneuvering target could evade attack unless you can direct enough energy in its general vicinity to make maneuvers unimportant (ie fire many beams over an area of space).
Lasers basically push the potential engagement ranges MUCH further out.
I said:
The feature is in several stable kernel branches.
You said:
No, it is not in any stable kernel branches other than those later than 3.17.
Your use of the word "no" suggests that there is a contradiction here, and there isn't.
It is present in 3.18 and 3.19, which are both stable kernel branches.
Maybe save it for the next systemd article then? To boost income, slashdot only has about 5 of them each week... :)
The application dropped support for production kernels ... because it wants a patch that isn't yet in production kernels.
The feature is in several stable kernel branches. Your distro might just not support them, so either don't use Chrome, or don't use that distro, or figure out how to use a newer kernel on your distro. :)
So "The Google Way" is the correct way, and every other application in Linux is insecure?
Security isn't a binary thing. Google's way is more secure, and I'm sure other applications will start using it. It only hit the kernel a few months ago, don't be surprised if every single process on Linux is using it.
2. Vendor doesn't seem to care about kernel/userspace boundaries and very happily writes code on both sides to an interface they've designed themselves, for themselves.
So, nothing ends up in Linux unless Linus lets it in, and it isn't like this was snuck in.
Really, Google is doing the right thing here. Instead of trying to hack inter-process security in on the userspace side, they're extending the kernel to improve things. Inter-process communications/relationships/etc is one of the things the kernel is supposed to do.
This is no more a violation of boundaries than putting modesetting in the kernel instead of running X11 as root and having it set device registers and such.
That's what airburst grenades are for. :) I'm not saying the laser is the solution to EVERY problem. I don't really see them as being replacements for hand-held rifles. This is something that would be combined with an automated control system.
Also, when somebody else throws a grenade at your wall, then your laser turret can blast it out of the air.
Of course. I'm not saying that it is a solved problem. I'm just saying that there is a reason people are interested in solving it.
They should do some tests to see how many seconds a 120mm shell fired from an M1A1 at ~1700m/s needs to be in contact with the truck to disable it.
Perhaps a more important question is how long the laser has to be pointed at the 120mm shell in-flight from the M1A1 at 1700m/s before the shell does nothing on contact with the target.
Blowing up targets sitting still on a range has been a solved problem for centuries. That doesn't mean that an Aegis air defense system has no more value than a medieval cannon.
0.1% of the kill rate of an A10 Warthog, for only 1000 times the cost. But hey, somebody's making a lot of money here, so we can't complain.
Speed-of-light weapons that have virtually linear trajectories certainly have the potential to change warfare though.
Put it this way - if you pointed that A10 cannon upwards at an aircraft at 70k feet, you'd be hard-pressed to hit it at all. On the other hand, a laser would have relatively little difficulty hitting the aircraft even if it were in geosync orbit, or even on the surface of the moon.
Stick something like this on a plane and you could use it to shoot down incoming missiles, shoot artillery shells in mid-flight, shoot aircraft, and so on.
Sure, the technology is immature, but it certainly is a capability that is valuable for a military to posses.
Politicians don't listen to grassroots activists, they listen to $100,000 contributors.
Yup. This is largely due to the fact that voters don't listen to grassroots activists either - they do whatever the ads on TV tell them to do, maybe augmented by whatever the union boss or preacher down the street tells them.
Of course, a fair and impartial trial will also require him to accept a very high likeliness of losing the case, based on the current evidence against him.
Yup. When the law is immoral, only the immoral are innocent of breaking the law.
Is this supported if you don't set a screen lock password/etc?
I'd love to have FDE, but I have no desire to enter a password when driving/etc. The two should be completely independent.
Whether Android does this I have no idea, but the device could be configured to power off if the wrong screen PIN was entered too many times. A FDE password has to withstand offline attack, which means unlimited attempts at a high rate.
It is completely appropriate to use a different level of security for each.
It's depressing that morons are willing to do so much damage for so little gain.
Here is the thing - they're not morons - they're sociopaths. The morons are the rest of us for not making simple legal changes to prevent this nonsense, largely because the beneficiaries of this theft want it to continue.
The answer to fixing this problem is to require scrap metal dealers to be licensed (with strong penalties for anyone who isn't) and to require all transactions to be recorded along with the ID of the seller. Its already done in many jurisdictions for pawn shops (where you need a license to operate one and where sellers have to provide ID when they sell it, why should scrap merchants be any different.
They should also have a delay before payout, with appointments required for the time to pick up the money. I imagine that even a one week delay would be enough to allow utilities to track down sales of stolen goods, and then the police can be waiting when the guy shows up to collect.
Stuff like this isn't going to be a problem for anybody legit. Electricians aren't going to trips to the dealer to sell one coil of wire, and be desperate for their $50.
A better example of a coding error might be this one.
In the event of a global collapse, these people will simply carry on as before.
If civilization collapses, there will be a reason that it collapsed. Such as a pandemic disease, crop destroying volcanic eruption, asteroid impact, nuclear winter, or runaway greenhouse effect. In any of these events, Africans will not "carry on as before". They will be the hardest hit, because they have nothing to fall back on.
Agree with this. The middle of Africa won't care if there are worldwide blackouts, as long as everybody has their plot of land to grow crops on. That kind of disaster will be very hard on the industrial world when you can't get food into your cities. On the other hand if the problem is that there are more bodies to feed than local land to feed them on, then the people of Africa will have a real problem on their hand, and will probably solve it by killing each other off until it is no longer a problem, since historically that is what tends to happen in these situations anywhere.
At the general election level; yes, you'd need a constitutional amendment. At the party primary level, however, such a system as you describe would be incredibly helpful, and probably for all parties.
How? A proportional representation system only works if you have multiple people elected into an office for any particular constituent.
The Republicans can't nominate 100 different people for the 3rd congressional district in Florida. Well, I guess they could, but then the first past the post system would practically guarantee that none would be elected, and that is why they only nominate one.
In a proportional system the Republicans would offer a prioritized list of 435 representatives for the entirety of the US House, and then their members could use proportional voting to decide who ends up in what slot. Then in the general election the Republicans would be awarded some number of seats, and the top n candidates would take office.
The president would still lead to the usual deadlock if separately elected since it is one man in one office. The only real solution to that is to replace the office with a prime minister, which is of course how most democracies handle the situation. That would require an amendment as well, and probably a rewrite of half the constitution.
Well, the real problem is a first-past-the-post election system, combined with only having one candidate elected to any particular office for any particular voter. This happens both at the general and party levels.
If you allowed each voter to elect a few hundred candidates, out of a much larger pool, using a proportional system, then you'd get a legislative body that better reflected the diversity of political views.
But, that would also require a constitutional amendment that serves neither of the parties in power, so you'll never see it happen.
Yes, let's not have any rules or oversight on "people" who were born in a lawyer's office, can potentially live forever, are motivated purely by greed, and will gladly break the law when it suits them. What could possibly go wrong?
That is a ridiculous assertion. There is no way they'll break the law, simply by virtue of the fact that without rules and oversight there won't be any laws for them to break.
Now get out there and sign some contracts of adhesion to help out the economy so that the big companies can spend more on lobbyists!
Ultimately, it is the immune system that keeps your body free of cancer. Cancers happen frequently in your body, and the immune system beats them down. When it fails at that for some reason, only then does clinical disease happen.
You could make the same argument about MANY functions within your body. There are many different mechanisms in cells that work to prevent uncontrolled cell growth. There are many controls on the cell cycle which have to fail. The tumor needs to create demand for more blood vessels to sustain growth. The tumor has to evade the immune system, etc.
It seems like cancer is the flip-side of multicellular life, and as a result humans have many different mechanisms for preventing it. Only when they all fail do cancers form. The problem is that you have billions of cells independently dividing and every time they do there is an opportunity for more mutations to creep in. Sooner or later any level of redundancy is bound to fail.
Nobody cared before, because CPUs and browser layout engines were the bottleneck not the network.
Nonsense. With some notable exceptions, network has always been the primary bottleneck.
Agree. Think about it - in 1999 a modem was a pretty common way of accessing web data.