They collapse when one passes through one polarizer. This will result in the other, passing through the other polarizer in a "modulo 90" orientation to behave predictably, either pass at mathing polarity or vanish at wrong polarity, with no uncertainty; "the same" measurement gets repeated, producing "the same" results.
If one passes at a "superfluous" polarizer at 45 degrees orientation, their wavefunction collapses to that. Afterwards, being no longer entangled, they may pass two "matching orientation" polarizers and produce different results. That's all that's needed to confirm the wavefunction has collapsed.
Yes, that's not all that much, yes it's a way to share keys in a taper-evident way, and yes there's a hype with misunderstanding the consequence of simultaneous collapse of wavefunction, coming from belief that the fact of the collapse is somehow recognizable without standard communication between the two ends. That's the purpose of my post: to show that it can be useful but it doesn't allow faster-than-light communication because the "information" that did travel faster than light is undistinguishable from random noise without a-posteriori standard communication.
Would you mind explaining the LC concept? It sounds quite interesting.
World of Tanks uses BT for its updates. You may choose a http alternative but download times will balloon to days; everyone who can't use BT uses them, and overwhelms the update servers every time there is something to update, so using BT to get the update through peer players is no-brainer.
The few informed voters are simply drowned out in the sea of people who buy into the sociological tricks of the parties. The people who control the campaigns control the country. The nation is offered an illusion of choice and a colorful show, but the real decision is made by the people who write the candidates lists.
Guerilla warfare is the epitome of stealth. With drones armed with infrared gone are the days of partizan groups hiding in forests - currently they hide in plain sight, among civilians, undistinguishable until it's too late. It's still stealth, but of a different kind - blending into the background in such a way that the soldier staring you in the face doesn't know you're a combatant.
Mobility is another means of stealth: if firing reveals your position and draws enemy fire, you need to relocate quickly and hide elsewhere.
Stealth air warfare allows to strike at strategical units deep inside the enemy structures - destroy the command, supply and communications. Without that, the front lines are easy pickings for conventional weaponry.
It's only when the front line falls apart and the enemy falls back to stealth tactic = guerrilla warfare - that's when things take a turn for the worse.
Yep, the first free elections in Iraq. The Nation spoke. They wanted islamist fundamentalist party at the helm. The USA choose to simply disregard the results of the elections and set up a government of their choice. "Democracy."
"Puppet government" and "government friendly to us" is just two different names for the same thing, depending on which side you're sitting.
"self-interest in survival" is a very interesting way to put their motivations behind remaining friendly to the USA.
"Pack your particle in your car and send it there, no information is known still. Now measure the particle at your end and at the other end simultaneously. You'll get the same result instantly".
The problem is this doesn't really transmit information. You could get two sealed envelopes, put two identical cards, either with "1"s on them or with "0"s, one in each, then deliver one to the distant place. If two people now simultaneously open the envelopes, "the information about the number on the card will be transmitted instantly".
And one person may choose not to open the envelope.
No, no useful information was transmitted. The one who opened the envelope can't tell whether the other person opened the envelope by color of the marble alone. In this scenario the quantum entanglement boils down to the "hidden variables" theory and doesn't provide anything better than completely standard communication.
Take a different scenario.
The cards are written in special inks that require specific chemical to become visible. Specifically, you have one kind of ink in which you write "0" or "1" and another, in which you write "A" or "B". You have swabs of cotton with two chemicals - one will make the "0" or "1" to show up - but will simultaneously wash out the "A" or "B" unrecoverably. The other will do the opposite, washing the "0/1" off but showing the "A" or "B".
You prepare identical pairs of the cards - both cards of the pair have the same number and the same letter. You distribute the cards to the two parties, and each can choose a swab at random, to display either a number or a letter.
Later they compare notes, which swabs they used - if they both used the same type of swab, they know they got the same number, or the same letter. If they used different swabs - they just agree to disregard the number-letter pair. They also share some of the results.
If a third party wants to intercept it, they can read the number or the letter, but if they choose a number, they'll never know the letter and vice versa. They can create a fake card which contains the same number or the same letter, but not both for sure - they got a card and found it to be "1", so they can create a card with "1" but put "A" or "B" randomly.
Now if the two parties compare their partial results, find that both checked for a letter on a particular card, and one got "A" while the other got "B" that means someone faked one of the card and the channel was compromised.
That's quantum encryption. A useful exchange was made, and it was absolutely taper-proof (or taper-evident if you prefer), but no superluminal communication occurred - actually no communication occurred over the quantum channel; simply both parties obtained *some* common information; later standard communication allowing them to verify whether there were any superfluous measurements.
You set up the experiment so that you can polarize a photon at 4 angles: 0, 45, 90 and 135 degrees ( | / - \ ).
There are two distant terminals, let's call them A and B, where the photons can be polarized and then checked whether they passed the polarizer or not. There's also a (dumb) source of entangled pairs in the middle, that sends one photon from the pair to each of the terminals.
Take a single (non-entangled) photon: If you polarize it at 0 degrees, it will pass the 0 degrees polarizer 100% of the time, 90 degrees 0% of the time, and the two diagonal ones 50% of the time. Extend to three other cases by rotating the initial polarizer by multiplies of 45 degrees; it's analogous.
Pass the same photon through three polarizers now: 0, 45 and 90 degrees. Unlike with just two (0 and 90) It will pass in 25% cases.
Take an entangled pair of photons. If you polarize one of them, the other behaves as if it was polarized the same as the first.
As entangled photons are sent, both A and B choose random orientations of their polarizers (each with own, locally generated random sequence); they write down the sequence: angle, result (photon passed or not).
Now the result is a string of zeros and ones each with an angle. If both randomly choose | or - then the result is valid, 1 means the other side had the same orientation, 0 means the other side had a perpendicular orientation. If one choose | or -, but the other choose \ or/, the result is random junk. The problem though is that neither of them knows which ones are right and which ones were faulty. There's a lot of data on both ends but none useful. No *actual* information was exchanged, because any that really did, was hidden behind the randomness of the polarizer setting.
Now, using normal, non-encrypted channels, A and B exchange the recorded random sequence of polarizer settings.
Each compares this with own recording and converts: Mine was |, their was |, got 1, record 1. Mine was |, their was -, got 0, record 1. Mine was -, their was -, got 1, record 1. Mine was |, their was / - discard record; it's junk.
And again, no information was passed from A to B because all the information was *generated*, in two copies at two ends. A couldn't send B a single bit. It's the dummy emitter that sent a random bit in two directions, and it was recorded on two ends. Still, to be actually read, it required normal subluminal exchange for decoding.
Nevertheless, both A and B now have the same sequence of bits, which they can use as a key for a common encryption - and they know the key had not been intercepted; no third party has it.
How do they know? Because for a third party to get any data from the photons, they'd need to put a polarizer along the way and since they don't know the sequence, they'd have to turn it at random.
Now remember the "three polarizers" case from the beginning?
"Mine was |, their was -, I got 1. Alarm! Somebody put a 45 degrees polarizer along the way! The communication has been intercepted!"
Afghanistan Iraq Kuwait Serbia South Korea West Germany Nikaragua
Of course they don't include them in the territory of the USA. That would be economically disadvantageous. They just set up the puppet governments that do the US bidding.
Care to check the debt Iraq has to the USA? All these refineries rebuilt in place of the bombed ones? Who authorized the loan?
Yep. Better flak jacket for soldiers to survive more powerful rounds? Harder tank armor to withstand multiple hits from armor-piercing rockets? It's all a dead end.
We've developed so efficient assault capabilities - rockets, ammo, smart bombs - that they've outpaced the defensive technology so far they just won't catch up. The only way not to die currently is to either not be stopped or to kill them before they can kill you. Dodging or soaking damage is no longer really an option.
It's still better than nothing. A passive broadband radar will take milliseconds to adapt the coating to any single radar frequency. With more than one it will run into trouble, but again, multiple blips at predictable intervals are quite easy to program, so the airplane would be able to hide from multiple radars.
This can be thwarted relatively easily with a radar that constantly shift frequencies randomly, but... you need to build one. The old infrastructure becomes obsolete, necessitating costly upgrades. It may provide exactly zero battlefield advantage in the long run but if it forces the opponent to suffer costs of performing upgrades of their infrastructure which would be otherwise not needed, it's already a win. And by publishing the paper (instead of costly building air force exploiting the new tech, which *might* leak through the opponent spies (and become obsolete fast), or might remain secret and become obsolete much slower) they are simply running the costs up for the USA.
"Upgrade all your radars to frequency-hopping right now, before anyone builds a plane that will be invisible to them". And the plane *will* be built somewhere, because there are many countries that just can't afford upgrading their radar infrastructure, and their opponents would benefit greatly from planes invisible against them - even if they are useless against the (upgraded) USA radars. And the USA will definitely dislike the option of having 3rd party airplanes in the air which they can't see, even if they don't mean a direct threat currently.
Economical warfare: at relatively low cost for yourself force the enemy to spend a bunch of money on defenses they will probably never need. (but will never need them only if they build them... if they don't, they'll regret they didn't.)
China alone can't maintain air superiority against the USA. They can (and will) use the technology but the numbers alone mean the US will keep ahead of them.
But after releasing it to the public, every half-civilized country will be able to make their own stealth fighters. And who has more airplanes: the USA, or the rest of the world? With the cat out of the box, USA will be facing competent opposition in any major conflict involving aerial forces, which will slow down their advance of power, giving China a chance in the race.
China, not being nearly as expansionist as the USA has far less to lose by having random countries all over the world armed with technology to oppose state-of-the-art air forces. This paper directly strikes a blow to USA's assault capabilities while not really hurting China, which doesn't have nearly as much interest in assault capabilities in the first place, concentrating on a mighty defensive force.
There are quite few utilities suid root currently. The usual approach to "common day" privilege escalation is some server/service/daemon working with privileges set to exactly what it needs for its work (if root, so be it) and a "frontend app" that runs on user level and contacts the demon for said service. This allows for an additional choke point as only data that needs the extra privileges gets through, while the "client" handles all the rest; no hundreds of options that could exploit the escalated privileges, just a dozen or so that are designed to be easy to sanitize.
Debian never abandoned glibc, AFAIK. It just changed the maintainer of the package, replacing the person who'd get angry at people submitting patches that only benefitted embedded.
And yes, device tree has many shortcomings that should be addressed one way or another - if replacing the whole subsystem with something better is the way, so be it. Nevertheless it's the embedded that benefits from it primarily (early Linux had maybe a dozen architectures, which could be all managed by #ifdef and.config entries; it was a bit of a mess but not a total road-block. Now it goes into hundreds if not thousands.) and so it's a proof things are done for the embedded.
15 years ago Linux for Embedded was a pipe dream. It was something tinkerers would do "because they could" without any practical, commercial application. Linux was a system for servers and nerds, with a slow promise of general desktop. Embedded Linux? Only as a joke, "he installed Linux on his toaster."
Only a limited of manufacturers understand that the only things really needed today is full Linux mainline support and a full support to a good boot loader.
No, there's one more thing that is needed: Kernel support for all the peripherals. Device Tree reduced the totally unmaintainable zoo of millions of implementations of peripherals into a more maintainable format where you just need how given hardware interacts with the core - which pins, which interrupts, what timings etc. HAL again removed one more ugly thing that was each device appearing as something completely different in the OS - is CAN a socket, or a/dev/, or a/sys entry? But still, the hardware half of each new peripheral device needs to be provided, preferably by the manufacturer.
Linux will not change in a fundamental way to replace the smallest of them, at least in the case of your description of "your PC at least a dozen different OS installations start up in various controllers" (I have big doubt that you can prove that on my PC).
OSD processors on 2 monitors. On-board hardware drivers for 2 hard disks. Controller on the gfx card. BIOS. SATA controller on motherboard. HDMI controller in your gfx card. Trusted Computing chip on the motherboard. A key dongle for some proprietary software you have. True, most of them run specialized OS'es designed for embedded, like COS, but they all have a potential to run Linux. I know of modems that run it. I'm also fairly sure the SAS card in my work PC runs it.
"Linux will not change..." - because Busybox never happened? I'm not talking about a fundamental change that would impact all installations globally. I'm talking about some adaptation, e.g. a fork of systemd - aimed specifically at smaller embedded. And once again, the slice of the market may seem tiny, but it isn't - trust an insider.
The manufacturer may change the chip to be able to run Linux, but they won't compromise the primary niche for given chip: being radiation-proof, being ultra-low-power, handling extreme temperatures, and such, just so that it could run "a fully-featured distribution". If the chip needs to have limited resources because the conditions require it, either don't run Linux on it or trim your Linux to fit.
The chips that great most space probes fly on, are built in a technology that is 30 years old, and while the software is advancing, the technology is not - because nothing newer was successfully developed as radiation-proof ever since. It's a good while until a running Linux install leaves the Earth orbit - and the first time it happens, I'm fairly sure it won't be running systemd.
But I'm lazy. In this mentioned case I'd receive a fair credit rating. In case of tax returns I'm getting whatever goods my taxes pay for, plus not getting bullied by the country. I see no benefits in passing this information to you. Provide an incentive (say, 0.1BTC to my wallet, 18vsEWJizN2dEdEhykXF4kFQ1udS4n7wSr ) and you can have it.
Slashdot Mirror
Do they need to interact further?
They collapse when one passes through one polarizer. This will result in the other, passing through the other polarizer in a "modulo 90" orientation to behave predictably, either pass at mathing polarity or vanish at wrong polarity, with no uncertainty; "the same" measurement gets repeated, producing "the same" results.
If one passes at a "superfluous" polarizer at 45 degrees orientation, their wavefunction collapses to that. Afterwards, being no longer entangled, they may pass two "matching orientation" polarizers and produce different results. That's all that's needed to confirm the wavefunction has collapsed.
Yes, that's not all that much, yes it's a way to share keys in a taper-evident way, and yes there's a hype with misunderstanding the consequence of simultaneous collapse of wavefunction, coming from belief that the fact of the collapse is somehow recognizable without standard communication between the two ends. That's the purpose of my post: to show that it can be useful but it doesn't allow faster-than-light communication because the "information" that did travel faster than light is undistinguishable from random noise without a-posteriori standard communication.
Would you mind explaining the LC concept? It sounds quite interesting.
Because it's already on a service that can't be overwhelmed: Bittorrent.
And is it?
World of Warcraft. World of Tanks. Linux. BTSync. LibreOffice updater. Lots of Creative Commons licensed music. Project Gutenberg e-books.
Anyone care to provide some statistics? Because my hunch tells me they account for more than 1% of the traffic.
World of Tanks uses BT for its updates. You may choose a http alternative but download times will balloon to days; everyone who can't use BT uses them, and overwhelms the update servers every time there is something to update, so using BT to get the update through peer players is no-brainer.
RTFA is so last century.
RTFS was new until recently.
Now we have RTFT. Read the f***ng title.
Advances in Sociology have destroyed Democracy.
The few informed voters are simply drowned out in the sea of people who buy into the sociological tricks of the parties. The people who control the campaigns control the country. The nation is offered an illusion of choice and a colorful show, but the real decision is made by the people who write the candidates lists.
Guerilla warfare is the epitome of stealth. With drones armed with infrared gone are the days of partizan groups hiding in forests - currently they hide in plain sight, among civilians, undistinguishable until it's too late. It's still stealth, but of a different kind - blending into the background in such a way that the soldier staring you in the face doesn't know you're a combatant.
Mobility is another means of stealth: if firing reveals your position and draws enemy fire, you need to relocate quickly and hide elsewhere.
Stealth air warfare allows to strike at strategical units deep inside the enemy structures - destroy the command, supply and communications. Without that, the front lines are easy pickings for conventional weaponry.
It's only when the front line falls apart and the enemy falls back to stealth tactic = guerrilla warfare - that's when things take a turn for the worse.
It worked for space shuttles (actually for the enormous fuel tank of the space shuttle) and with 2.5" of foam layer.
These are all engineering problems to be solved, not show-stoppers that make it useless.
Plus wanna bet how much this can be improved if one tosses a couple billion at improvement research?
Yep, the first free elections in Iraq. The Nation spoke. They wanted islamist fundamentalist party at the helm. The USA choose to simply disregard the results of the elections and set up a government of their choice. "Democracy."
"Puppet government" and "government friendly to us" is just two different names for the same thing, depending on which side you're sitting.
"self-interest in survival" is a very interesting way to put their motivations behind remaining friendly to the USA.
The misconception comes from:
"Pack your particle in your car and send it there, no information is known still. Now measure the particle at your end and at the other end simultaneously. You'll get the same result instantly".
The problem is this doesn't really transmit information. You could get two sealed envelopes, put two identical cards, either with "1"s on them or with "0"s, one in each, then deliver one to the distant place. If two people now simultaneously open the envelopes, "the information about the number on the card will be transmitted instantly".
And one person may choose not to open the envelope.
No, no useful information was transmitted. The one who opened the envelope can't tell whether the other person opened the envelope by color of the marble alone. In this scenario the quantum entanglement boils down to the "hidden variables" theory and doesn't provide anything better than completely standard communication.
Take a different scenario.
The cards are written in special inks that require specific chemical to become visible. Specifically, you have one kind of ink in which you write "0" or "1" and another, in which you write "A" or "B". You have swabs of cotton with two chemicals - one will make the "0" or "1" to show up - but will simultaneously wash out the "A" or "B" unrecoverably. The other will do the opposite, washing the "0/1" off but showing the "A" or "B".
You prepare identical pairs of the cards - both cards of the pair have the same number and the same letter. You distribute the cards to the two parties, and each can choose a swab at random, to display either a number or a letter.
Later they compare notes, which swabs they used - if they both used the same type of swab, they know they got the same number, or the same letter. If they used different swabs - they just agree to disregard the number-letter pair. They also share some of the results.
If a third party wants to intercept it, they can read the number or the letter, but if they choose a number, they'll never know the letter and vice versa. They can create a fake card which contains the same number or the same letter, but not both for sure - they got a card and found it to be "1", so they can create a card with "1" but put "A" or "B" randomly.
Now if the two parties compare their partial results, find that both checked for a letter on a particular card, and one got "A" while the other got "B" that means someone faked one of the card and the channel was compromised.
That's quantum encryption. A useful exchange was made, and it was absolutely taper-proof (or taper-evident if you prefer), but no superluminal communication occurred - actually no communication occurred over the quantum channel; simply both parties obtained *some* common information; later standard communication allowing them to verify whether there were any superfluous measurements.
Quantum encryption protocol BB-84
You set up the experiment so that you can polarize a photon at 4 angles: 0, 45, 90 and 135 degrees ( | / - \ ).
There are two distant terminals, let's call them A and B, where the photons can be polarized and then checked whether they passed the polarizer or not. There's also a (dumb) source of entangled pairs in the middle, that sends one photon from the pair to each of the terminals.
Take a single (non-entangled) photon: If you polarize it at 0 degrees, it will pass the 0 degrees polarizer 100% of the time, 90 degrees 0% of the time, and the two diagonal ones 50% of the time. Extend to three other cases by rotating the initial polarizer by multiplies of 45 degrees; it's analogous.
Pass the same photon through three polarizers now: 0, 45 and 90 degrees. Unlike with just two (0 and 90) It will pass in 25% cases.
Take an entangled pair of photons. If you polarize one of them, the other behaves as if it was polarized the same as the first.
As entangled photons are sent, both A and B choose random orientations of their polarizers (each with own, locally generated random sequence); they write down the sequence: angle, result (photon passed or not).
Now the result is a string of zeros and ones each with an angle. If both randomly choose | or - then the result is valid, 1 means the other side had the same orientation, 0 means the other side had a perpendicular orientation. If one choose | or -, but the other choose \ or /, the result is random junk. The problem though is that neither of them knows which ones are right and which ones were faulty. There's a lot of data on both ends but none useful. No *actual* information was exchanged, because any that really did, was hidden behind the randomness of the polarizer setting.
Now, using normal, non-encrypted channels, A and B exchange the recorded random sequence of polarizer settings.
Each compares this with own recording and converts: Mine was |, their was |, got 1, record 1. Mine was |, their was -, got 0, record 1. Mine was -, their was -, got 1, record 1. Mine was |, their was / - discard record; it's junk.
And again, no information was passed from A to B because all the information was *generated*, in two copies at two ends. A couldn't send B a single bit. It's the dummy emitter that sent a random bit in two directions, and it was recorded on two ends. Still, to be actually read, it required normal subluminal exchange for decoding.
Nevertheless, both A and B now have the same sequence of bits, which they can use as a key for a common encryption - and they know the key had not been intercepted; no third party has it.
How do they know? Because for a third party to get any data from the photons, they'd need to put a polarizer along the way and since they don't know the sequence, they'd have to turn it at random.
Now remember the "three polarizers" case from the beginning?
"Mine was |, their was -, I got 1. Alarm! Somebody put a 45 degrees polarizer along the way! The communication has been intercepted!"
Afghanistan
Iraq
Kuwait
Serbia
South Korea
West Germany
Nikaragua
Of course they don't include them in the territory of the USA. That would be economically disadvantageous. They just set up the puppet governments that do the US bidding.
Care to check the debt Iraq has to the USA? All these refineries rebuilt in place of the bombed ones? Who authorized the loan?
That still depends on density. If it's a solid of >1g/cm^3 that's a lot. If it's a foam-like coating, that's not such a problem.
Yep. Better flak jacket for soldiers to survive more powerful rounds? Harder tank armor to withstand multiple hits from armor-piercing rockets? It's all a dead end.
We've developed so efficient assault capabilities - rockets, ammo, smart bombs - that they've outpaced the defensive technology so far they just won't catch up. The only way not to die currently is to either not be stopped or to kill them before they can kill you. Dodging or soaking damage is no longer really an option.
It's still better than nothing. A passive broadband radar will take milliseconds to adapt the coating to any single radar frequency. With more than one it will run into trouble, but again, multiple blips at predictable intervals are quite easy to program, so the airplane would be able to hide from multiple radars.
This can be thwarted relatively easily with a radar that constantly shift frequencies randomly, but... you need to build one. The old infrastructure becomes obsolete, necessitating costly upgrades. It may provide exactly zero battlefield advantage in the long run but if it forces the opponent to suffer costs of performing upgrades of their infrastructure which would be otherwise not needed, it's already a win. And by publishing the paper (instead of costly building air force exploiting the new tech, which *might* leak through the opponent spies (and become obsolete fast), or might remain secret and become obsolete much slower) they are simply running the costs up for the USA.
"Upgrade all your radars to frequency-hopping right now, before anyone builds a plane that will be invisible to them". And the plane *will* be built somewhere, because there are many countries that just can't afford upgrading their radar infrastructure, and their opponents would benefit greatly from planes invisible against them - even if they are useless against the (upgraded) USA radars. And the USA will definitely dislike the option of having 3rd party airplanes in the air which they can't see, even if they don't mean a direct threat currently.
Economical warfare: at relatively low cost for yourself force the enemy to spend a bunch of money on defenses they will probably never need. (but will never need them only if they build them... if they don't, they'll regret they didn't.)
Shifting the balance of power.
China alone can't maintain air superiority against the USA. They can (and will) use the technology but the numbers alone mean the US will keep ahead of them.
But after releasing it to the public, every half-civilized country will be able to make their own stealth fighters. And who has more airplanes: the USA, or the rest of the world? With the cat out of the box, USA will be facing competent opposition in any major conflict involving aerial forces, which will slow down their advance of power, giving China a chance in the race.
China, not being nearly as expansionist as the USA has far less to lose by having random countries all over the world armed with technology to oppose state-of-the-art air forces. This paper directly strikes a blow to USA's assault capabilities while not really hurting China, which doesn't have nearly as much interest in assault capabilities in the first place, concentrating on a mighty defensive force.
Go away, troll.
Casuals?
So, how many pro e-sports gamers choose controller over mouse in a game that supports both?
There are quite few utilities suid root currently. The usual approach to "common day" privilege escalation is some server/service/daemon working with privileges set to exactly what it needs for its work (if root, so be it) and a "frontend app" that runs on user level and contacts the demon for said service. This allows for an additional choke point as only data that needs the extra privileges gets through, while the "client" handles all the rest; no hundreds of options that could exploit the escalated privileges, just a dozen or so that are designed to be easy to sanitize.
Analog movement is slightly nicer than WASD. Mouse aiming/turning is vastly nicer than stick aiming/turning.
Debian never abandoned glibc, AFAIK. It just changed the maintainer of the package, replacing the person who'd get angry at people submitting patches that only benefitted embedded.
And yes, device tree has many shortcomings that should be addressed one way or another - if replacing the whole subsystem with something better is the way, so be it. Nevertheless it's the embedded that benefits from it primarily (early Linux had maybe a dozen architectures, which could be all managed by #ifdef and .config entries; it was a bit of a mess but not a total road-block. Now it goes into hundreds if not thousands.) and so it's a proof things are done for the embedded.
15 years ago Linux for Embedded was a pipe dream. It was something tinkerers would do "because they could" without any practical, commercial application. Linux was a system for servers and nerds, with a slow promise of general desktop. Embedded Linux? Only as a joke, "he installed Linux on his toaster."
No, there's one more thing that is needed: Kernel support for all the peripherals. Device Tree reduced the totally unmaintainable zoo of millions of implementations of peripherals into a more maintainable format where you just need how given hardware interacts with the core - which pins, which interrupts, what timings etc. HAL again removed one more ugly thing that was each device appearing as something completely different in the OS - is CAN a socket, or a /dev/, or a /sys entry? But still, the hardware half of each new peripheral device needs to be provided, preferably by the manufacturer.
OSD processors on 2 monitors. On-board hardware drivers for 2 hard disks. Controller on the gfx card. BIOS. SATA controller on motherboard. HDMI controller in your gfx card. Trusted Computing chip on the motherboard. A key dongle for some proprietary software you have. True, most of them run specialized OS'es designed for embedded, like COS, but they all have a potential to run Linux. I know of modems that run it. I'm also fairly sure the SAS card in my work PC runs it.
"Linux will not change..." - because Busybox never happened? I'm not talking about a fundamental change that would impact all installations globally. I'm talking about some adaptation, e.g. a fork of systemd - aimed specifically at smaller embedded. And once again, the slice of the market may seem tiny, but it isn't - trust an insider.
The manufacturer may change the chip to be able to run Linux, but they won't compromise the primary niche for given chip: being radiation-proof, being ultra-low-power, handling extreme temperatures, and such, just so that it could run "a fully-featured distribution". If the chip needs to have limited resources because the conditions require it, either don't run Linux on it or trim your Linux to fit.
The chips that great most space probes fly on, are built in a technology that is 30 years old, and while the software is advancing, the technology is not - because nothing newer was successfully developed as radiation-proof ever since. It's a good while until a running Linux install leaves the Earth orbit - and the first time it happens, I'm fairly sure it won't be running systemd.
If they manage to get to me they can have the data too.
Nah, it's more of a sporting thing. Come and get it if you can.
But I'm lazy.
In this mentioned case I'd receive a fair credit rating. In case of tax returns I'm getting whatever goods my taxes pay for, plus not getting bullied by the country. I see no benefits in passing this information to you. Provide an incentive (say, 0.1BTC to my wallet, 18vsEWJizN2dEdEhykXF4kFQ1udS4n7wSr ) and you can have it.
What will I get in exchange?