The other big thing would be to require that the package is the most recent version in the repository. That alone would eliminate most non-0-day security issues -- assuming, obviously, that any such issue will get a prompt security fix.
That could be combined with a non-root-install blacklist in case there isn't a good fix.
I wouldn't be tremendously worried about J-random-user's desktop being compromised by a 0-day exploit...
The big security issues are with services and suid/sgid binaries. As long as installing the package doesn't cause any other users to run that code it shouldn't be any more of a problem than letting the user run their own code. Those probably should require root to install.
This isn't 100% watertight but would drastically reduce the scope for local root exploits.
It'd be stupid to have this policy on a server, obviously, but it doesn't seem to be that absurd at all for desktops/workstations.
Everytime nuclear fission comes up as a possible viable alternative. Peak Uranium is as real as peak oil, and it's here now.
No it's not. There is plenty of uranium out there if we're willing to pay enough for it -- and the cost of fuel is a tiny portion of the cost of running a fission plant. There isn't much uranium production right now because there isn't much demand for it -- and a large portion of that demand is filled by sources other than mining.
Am I the only person who finds it odd that nuke plants waste is more radio active the more it is processed or used as fuel? I would think that once it is used as fuel it would have less energy so less radioactive.
Uranium (even U-235) is not very radioactive, in that its half life is hundreds of millions of years -- so long that some of it has lasted for the billions of years since the Earth was formed. Pu-239 has a half-life of tens of thousands of years.
When these atoms undergo fission, many of the resulting isotopes have short half lives. A lot of the nasty ones last for a few years or decades.
Fission tends to take atoms with long half lives and turn them into (pairs of) atoms with short half lives. The latter have a lot less energy but they're a lot more radioactive, at least for the first few centuries before most of it decays.
Deuterium is limited by the amount of deuterium in ocean water, which is the largest source on Earth but remains quite limited.
If you put all the deuterium in the oceans through D-D fusion, you'd get several years worth of solar output out of it. When I say "year of solar output" I mean the heat radiated by the entire star in a year. The energy we could get from deuterium would be limited by the earth's ability to radiate heat long before it would be limited by the supply of deuterium.
The existence of a Schwartzchild radius assumes that gravity can ever be stronger than the repulsive forces within the nucleus. It cannot. Both increase simultaneously as you increase mass. Gravity's attractive force will never be stronger than the electrostatic forces that hold the particles apart.
To create a small black hole you don't need to overcome the repulsive forces (which are primarily Pauli Exclusion and not electrostatic) indefinitely. You only need to overcome them for an instant.
It need not be gravity which overcomes said forces, either.
If you fire a bunch of high energy particles at each other such that they are within their mutual Schwartzchild radius you'd get a black hole.
You might think that Pauli Exclusion would push everything apart. It does, in a sense -- but space curves in towards the singularity faster than *any* repulsion can act! The entire point of GR is that you need to think of gravitation as a geometric thing and not a force.
Polynomial time approximate, probabilistic or special case solutions to NP problems are wide spread.
That's all well and good, but the point is that any theory that requires the actors to have infinite computational power must be incorrect, irrespective of the rationality of the actors. Note that it is possible to prove it is NP-hard to even approximate solutions to some problems.
The problem is that real human being in economics can not be easily described by an equation
Which does not mean that their behavior in certain cases can not be usefully approximated by one.
and when they can be, they quickly change their behavior based on that knowledge.
Not necessarily.
A model may well tell them that they are already behaving correctly. It could also point out a way they could change their behavior to their own benefit, without invalidating the model.
Just because no model can effectively predict what the stock market will do tomorrow does not mean that economic modeling is useless.
Their numbering scheme is incremented in.01's. It would not be consistent with their scheme to increment by.10.
Although I agree -- if they aren't close to a 1.0 they should consider moving to a two-dot system ("0.2.3") which would make it clearer when a release has large destabilizing changes.
What I wonder about is how much material would be needed to make a cable, or ribbon, that long. Just think of the volume of material needed for even a cable a foot in diameter. How many Mount Everests would it take?
A cable a foot in diameter and 22K mi long would require on the order of a thousandth of a cubic mile, much less than the volume of Mount Everest.
With that said, the trick is that the cable doesn't need to be anywhere near a foot in diameter, at least to start. A seed cable would be small enough that we could plausibly manufacture it on the ground and ship it into orbit.
If you have a material with enough tensile strength it doesn't need to be very thick.
Block based deduplication does not have that problem. Writing to a deduplicated block only requires a copy of that block.
This isn't actually a matter of ZFS being a "copy on write" filesystem. Any filesystem implementing block level deduplication needs to support copy on write for duplicate blocks, but it doesn't need to support copy on write for everything.
This is block based. Changing one block of each file will only result in one new block written, not a full copy of the file -- unless the file is only one block.
Certain data will be more susceptible to collision
The entire point of a cryptographic hash is that this is not the case. If you can reliably generate such data you've broken the hash.
Okay, if the data is "blocks whose first 16 bits of SHA-256 is 0" then it'll be a little more susceptible to collision, but you can't generate such blocks without either breaking SHA-256 or generating ~65535 blocks you *don't* write.
You can type, right? Unless you use non-keyboard input I'm not sure how you're going to do better than LaTeX for math input.
If it is slow you are doing something wrong.
f_X(x) = integral(-infinity, infinity, f(x,y) dy)
This is $f_X(x) = \int_{-\infty}^{\infty} f(x,y)dy$, to a first order approximation that is good enough for notes or even homework. That's fewer characters than your example!
Obviously, especially if you're a novice, you might not know the commands to do what you want -- but you can always fix up the syntax later.
Note that superconductors are, in a way, inherently easier to cool, because they don't generate resistive heat. All you need to do is cool them enough to take away the heat they gain from the environment, which you can mitigate with insulation.
How are you going to figure out *which* cell you're going to read from the RAM without looking at each of the O(log N) bits of the address? The decision to read from cell N depends on all O(log N) bits of N. It's not immediately obvious to me that *must* require O(log N) depth in your control circuitry, but it certainly isn't constant unless you allow unlimited fanin/fanout.
My guess is that miniaturizing a optical processor into silicon is probably going to be less powerful than normal optical processors.
The power of a quantum computer, at this early stage, is limited by the number of qbits. The point of putting it onto silicon is that silicon fabrication is the easiest way, right now, to make large numbers of interesting small structures. (in this case qbits and controlling infrastructure)
I think the real question is whether or not quantum computing can solve the Travelling Salesman problem.
It can not.
There is no reason to believe QBP contains NP. (We might be wrong, but then we might be wrong about P != NP!)
Any approach along the lines of "do everything quantumly in parallel and somehow select the interesting results" will do no better than a Grover search, which is a quadratic speedup.
Slashdot Mirror
The other big thing would be to require that the package is the most recent version in the repository. That alone would eliminate most non-0-day security issues -- assuming, obviously, that any such issue will get a prompt security fix.
That could be combined with a non-root-install blacklist in case there isn't a good fix.
I wouldn't be tremendously worried about J-random-user's desktop being compromised by a 0-day exploit...
The big security issues are with services and suid/sgid binaries. As long as installing the package doesn't cause any other users to run that code it shouldn't be any more of a problem than letting the user run their own code. Those probably should require root to install.
This isn't 100% watertight but would drastically reduce the scope for local root exploits.
It'd be stupid to have this policy on a server, obviously, but it doesn't seem to be that absurd at all for desktops/workstations.
A fair portion of geothermal is nuclear, albeit not fission -- much of the heat in the Earth's core is due to radioactive decay.
Everytime nuclear fission comes up as a possible viable alternative. Peak Uranium is as real as peak oil, and it's here now.
No it's not. There is plenty of uranium out there if we're willing to pay enough for it -- and the cost of fuel is a tiny portion of the cost of running a fission plant. There isn't much uranium production right now because there isn't much demand for it -- and a large portion of that demand is filled by sources other than mining.
Am I the only person who finds it odd that nuke plants waste is more radio active the more it is processed or used as fuel? I would think that once it is used as fuel it would have less energy so less radioactive.
Uranium (even U-235) is not very radioactive, in that its half life is hundreds of millions of years -- so long that some of it has lasted for the billions of years since the Earth was formed. Pu-239 has a half-life of tens of thousands of years.
When these atoms undergo fission, many of the resulting isotopes have short half lives. A lot of the nasty ones last for a few years or decades.
Fission tends to take atoms with long half lives and turn them into (pairs of) atoms with short half lives. The latter have a lot less energy but they're a lot more radioactive, at least for the first few centuries before most of it decays.
This must be the start of the 12 year old season, because all the posts seem to be full of ignorant assholes at the moment.
I agree.
Deuterium is limited by the amount of deuterium in ocean water, which is the largest source on Earth but remains quite limited.
If you put all the deuterium in the oceans through D-D fusion, you'd get several years worth of solar output out of it. When I say "year of solar output" I mean the heat radiated by the entire star in a year. The energy we could get from deuterium would be limited by the earth's ability to radiate heat long before it would be limited by the supply of deuterium.
The existence of a Schwartzchild radius assumes that gravity can ever be stronger than the repulsive forces within the nucleus. It cannot. Both increase simultaneously as you increase mass. Gravity's attractive force will never be stronger than the electrostatic forces that hold the particles apart.
To create a small black hole you don't need to overcome the repulsive forces (which are primarily Pauli Exclusion and not electrostatic) indefinitely. You only need to overcome them for an instant.
It need not be gravity which overcomes said forces, either.
If you fire a bunch of high energy particles at each other such that they are within their mutual Schwartzchild radius you'd get a black hole.
You might think that Pauli Exclusion would push everything apart. It does, in a sense -- but space curves in towards the singularity faster than *any* repulsion can act! The entire point of GR is that you need to think of gravitation as a geometric thing and not a force.
While it does have its uses, when will people like you the realize that free market isn't a magical thing immune to the limitations of time and space?
And when will people like you stop attacking strawmen?
Polynomial time approximate, probabilistic or special case solutions to NP problems are wide spread.
That's all well and good, but the point is that any theory that requires the actors to have infinite computational power must be incorrect, irrespective of the rationality of the actors. Note that it is possible to prove it is NP-hard to even approximate solutions to some problems.
The problem is that real human being in economics can not be easily described by an equation
Which does not mean that their behavior in certain cases can not be usefully approximated by one.
and when they can be, they quickly change their behavior based on that knowledge.
Not necessarily.
A model may well tell them that they are already behaving correctly. It could also point out a way they could change their behavior to their own benefit, without invalidating the model.
Just because no model can effectively predict what the stock market will do tomorrow does not mean that economic modeling is useless.
enough billions that you outbid everybody else in the entire country who wants some concrete.
Or, you know, concrete producers would respond to the increase in demand/prices by producing more concrete.
Their numbering scheme is incremented in .01's. It would not be consistent with their scheme to increment by .10.
Although I agree -- if they aren't close to a 1.0 they should consider moving to a two-dot system ("0.2.3") which would make it clearer when a release has large destabilizing changes.
What I wonder about is how much material would be needed to make a cable, or ribbon, that long. Just think of the volume of material needed for even a cable a foot in diameter. How many Mount Everests would it take?
A cable a foot in diameter and 22K mi long would require on the order of a thousandth of a cubic mile, much less than the volume of Mount Everest.
With that said, the trick is that the cable doesn't need to be anywhere near a foot in diameter, at least to start. A seed cable would be small enough that we could plausibly manufacture it on the ground and ship it into orbit.
If you have a material with enough tensile strength it doesn't need to be very thick.
Block based deduplication does not have that problem. Writing to a deduplicated block only requires a copy of that block.
This isn't actually a matter of ZFS being a "copy on write" filesystem. Any filesystem implementing block level deduplication needs to support copy on write for duplicate blocks, but it doesn't need to support copy on write for everything.
This is block based. Changing one block of each file will only result in one new block written, not a full copy of the file -- unless the file is only one block.
Are there any other filesystems with that feature?
WAFL.
Certain data will be more susceptible to collision
The entire point of a cryptographic hash is that this is not the case. If you can reliably generate such data you've broken the hash.
Okay, if the data is "blocks whose first 16 bits of SHA-256 is 0" then it'll be a little more susceptible to collision, but you can't generate such blocks without either breaking SHA-256 or generating ~65535 blocks you *don't* write.
You can type, right? Unless you use non-keyboard input I'm not sure how you're going to do better than LaTeX for math input.
If it is slow you are doing something wrong.
f_X(x) = integral(-infinity, infinity, f(x,y) dy)
This is $f_X(x) = \int_{-\infty}^{\infty} f(x,y)dy$, to a first order approximation that is good enough for notes or even homework. That's fewer characters than your example!
Obviously, especially if you're a novice, you might not know the commands to do what you want -- but you can always fix up the syntax later.
And you can't build it from source because...?
The bug is in the source?
Note that superconductors are, in a way, inherently easier to cool, because they don't generate resistive heat. All you need to do is cool them enough to take away the heat they gain from the environment, which you can mitigate with insulation.
How are you going to figure out *which* cell you're going to read from the RAM without looking at each of the O(log N) bits of the address? The decision to read from cell N depends on all O(log N) bits of N. It's not immediately obvious to me that *must* require O(log N) depth in your control circuitry, but it certainly isn't constant unless you allow unlimited fanin/fanout.
ECC's hardness is based upon a generalization of discrete log to elliptic curves.
The same algorithm that breaks discrete log on integers mod N breaks the elliptic curve generalization.
My guess is that miniaturizing a optical processor into silicon is probably going to be less powerful than normal optical processors.
The power of a quantum computer, at this early stage, is limited by the number of qbits. The point of putting it onto silicon is that silicon fabrication is the easiest way, right now, to make large numbers of interesting small structures. (in this case qbits and controlling infrastructure)
I think the real question is whether or not quantum computing can solve the Travelling Salesman problem.
It can not.
There is no reason to believe QBP contains NP. (We might be wrong, but then we might be wrong about P != NP!)
Any approach along the lines of "do everything quantumly in parallel and somehow select the interesting results" will do no better than a Grover search, which is a quadratic speedup.
Someone did an analysis.
Cite or it didn't happen.