Client of mine says he's putting data on an external drive, but then the drive shows up empty on the other computer.
Turns out he was formatting and writing to sdb1 instead of sdc1. One of the internal drive's partitions. Ooops!
After figuring that one out I get to clean up the mess. sdb1 is supposed to be a swap partition. mkswap/dev/sdb2.... Ooops typo! that was supposed to be sdb1! sdb2 is their active data partition.... He'd been formatting their (never used: someone had made a typo) swap partition. I had overwritten (part of) their data partition....
The only way to truly verify a system is good is to do it from a known good system. For a standalone PC that means booting off known-good read-only media, usually a CDROM, Here you have a slight problem with implementing your suggestion: The CPU boots off the read-write flash chips on the motherboard, not off the CDROM.
We have radios. They work for a few km. Some drivers like or don't mind driving right behind the other driver. Others prefer just to "lose track" of the other car, and then one party gets to pause for 15 minutes while the other gets a 20 minute break... Those 5 minutes means you're about 10km apart, too far for the radios.
In a few days I'll be among a group of 6 people driving over 1000km for a trip. We'll be in two cars. So when we want to plan our next stop, we'll call one of the people in the other car. We have a 66% chance of calling the not-driver-at-the-time. If we do happen to call the driver, we will pass the phone on to someone who is not driving.
More than 20 years ago, my "basic programming" course wanted to teach us "top down" programming. Break the top level thing from the assignment into smaller pieces until you more or less have chunks that fit into one line in your programming language.
This fails horribly if you're learning that programming language.
A friend wanted me to do his assignments. I refused, but offered to help him do his assignments.... So repeatedly saying: "and how do you break that into smaller pieces" resulted in him trying to parse floating point numbers....
So.... I would suggest: First give some assignments that end up building on the previous assignment. i.e. bottom up. Build something more complex after having built some basic building blocks. THEN introduce top down design, and if at all possible, have the building blocks from previous assingments fit in!
Indeed. The western world is in a "good times" part of a cycle.
What do you think happens when an ice age subsides? Suddenly there is lots and lots of food. Population numbers of all sorts of animals skyrocket. This allows the creation of a genetic pool, ready to be tested on the next "bad times" period.
Some genetic improvements require two steps, either a slight disadvantage, but together an advantage. It is extremely unlikely that those two mutations happen in the same individual at the same time. In a good times period, the individuals with the slight disadvantage get to survive, and mate with the individuals with the other genetic change.
Evolution is driven most strongly when there are cycles: Times are good, times are bad.
In the good times there is plenty of everything to multiply like mad, in the bad times, all the inferior products die.
Humanity is in a "multiply like mad" phase. Even the individuals with "something wrong" are surviving and reproducing. We currently live in an environment where certain medical problems are not fatal. It's simply a matter of the environment where you happen to live.
Certain fish are in a "bad times" period: Due to too much fishing by humans, only the smaller ones survive. So they have evolved to be smaller over the course of the last 20 years.
The theoretical case where hunters and prey-animals are in a constant arms-race and improve all the time is theoretical. Nothing much happens. Most of evolution happens when there is suddenly a period of fierce competition.
That's why a planet like mercury that doesn't turn on its axis, and doesn't have a moon will not have life. Too little cycles to drive evolution. (Astronomers decided that mercury would be lifeless because it's too hot before they found bacteria on earth that survive those temperatures....)
The device pictured in the video is not going to be practical.
It has a so called aspect ratio of about one (Width devided by length, if it were rectangular). This is inefficient. This is inefficient both in air and in water.
The low aspect ratio means it flies inefficient: I estimate it will fly at a glide ratio of about 1:3. So it requires a thrust of about a third of its weight to fly.
If it's going to go underwater, its going to have a specific weight that is comparable to water. That's not going to fly.
You need a higher aspect ratio to fly. At least to do so efficiently.
To be able to submerge, you need the specific weight to be about that of water. Normal submarines have a mixture of air and steel so that taking on some water makes them heavier than water. But their specific density needs to be quite close to that of water to have a little water make the difference between floating and sinking.
Having a specific density of water makes flying difficult.
The depicted vehicle with the door next to the jet engines has another disadvantage: you can't get off the plane for half an hour after landing... But thats just a practical issue that can be fixed easily....
I still hold to the rule of thumb: twice RAM. (I convinced SuSE and apparently red hat too, to stick to this rule a long time ago...) Anyway, it is also an "emergency brake". If a program in your system starts taking memory uncontrollably, your system will lock up for minutes at a time without swap, before it starts doing the "out of memory" thing. That usually results in killing off the wrong program. With swap your system will show signs of life during that time, and you might be able to kill the correct program in time....
When the first digital trespassing laws were introduced, the techies started saying that they were too broad. You could also punish most "good" guys with those laws. Those proposing these laws indicated that of course, the good guys wouldn't be punished, only the guys really doing bad things....
A few years later, the guys who were considered "good guys" are now being convicted under these laws. They are bad guys because the broke the law didn't they?
As security professionals, we can spot a hole in a computer system, and know it's exploitable without actually going in and exploiting it. So we don't actually have to go in to be convinced there actually IS a security hole.
When I was a student, I reported a hole in a computer system, to be told by the sysadmin that this was NOT a hole, that what I thought was a hole was in fact not a hole, because there were further security measures in place. I then decided I had other things to do, but I still think he was wrong. As a sysadmin, it is very difficult to acknowledge that a hole exists if you can't actually see the results. As an experiment I once checked: I sent a pre-release exploitable bug with exploit to interested people and asked if it was exploitable. About 50% could not reproduce the problem, and reported: safe, bug does not exist, no security advisory neccesary.
So, in a University environment, you train students for different jobs. Some of them need to become security experts. They learn the trade by breaking into computer systems at the university. They should then report the problems they found to the "authorities", who should request enough information to close said problems.
When I was a power user, and a responsible net citizen, I called my provider, and asked them: "When during the day are your top usage moments, so that I can reduce my usage during those moments?" They refused to answer.
But they should. If you ask the power-users to reduce file-sharing bandwidth during the top hours, I bet you can cause a significant improvement in throughput. And for the power users, a script to e.g. throttle the file-sharing program at 8 PM, and set it going again at 10PM is easy to install. And it makes a minor impact on their file sharing habbits. But for the bandwidth of the provider this will make a big difference.... The provider however needs to "grow up" and politely ask their power-users to move their usage away from the top hours-of-the-day.
Twelve years ago, I didn't think I needed the ".com" domain for my company. I just registered the ".nl". Stupid mistake.
I'd really love to be able to take over the ".com" domain from the guys that have it now.
On the other side, when I had a domain someone wanted, I wouldn't be too harsh on them. However, you're calling the shots: So I would demand that they leave your Email address in place. Or you can offer them that they specify what they need of the domain (web, certain EMail addreses) and then offer them to provide those.
Roger.
P.S. Guessing which domain is even easier than with the poster....
It's wonderful to have a tiny computer, but if you need to slap on a monitor, keyboard, and mouse to use it it's really not all that tiny, is it?
What do you use a computer for then? Ah. I can guess. You use it to do Email and type documents....
This thing sound to me as a great computer for some other uses. It would be great as the controlling unit in my washing machine. An UTP connection on the machine, and I can tell the washing machine to start at a certain time, instead of now, I have to start it before work and come back to already-smelly clothes after work.
Probably desgined for use in a cube-sat. Those are 4 inches on a side. So you'd have 7/8ths of the space available for batteries, radio and experiments.
Still, I can imagine a top Nvidia engineer spending a couple of weekends on adapting the GPU to run X86 code. If that would show promising, they could put a team on it, finish the project, and make a surpise move.
You correctly state that a GPU is usually an SIMD machine. So, they have an instruction fecth for the "I" in SIMD. They also have huge IO bandwidth for the "MD" part. If you go MIMD (muliticore in modern CPU terminology), you also need the huge IO bandwidth for the MI part. That's already done! They have that.
They just might come up with a 16 or 32 core CPU. It might perform on parallel tasks say 2x better than a top-of-the-line Intel or AMD CPU. (or per core 4x worse).
Now, building something that is 4x behind Intel performance is not that hard. A friend working on CPUs at the university had his design made into a chip, and it worked! Intel CPUs were running 200MHz at the time, his ran at 40. Intel CPUs had much smaller features on the chips. Scaling down his design would easily speed it up a lot. Nvidia already has access to the top-of-the-line fabs. We didn't.
CPUs are easily bandwidth limited. Doing the IO well, means you can easily outperform a current CPU. And doing the IO well is something NVIDIA already has the plans for.
At the computer architecture lab here at the university of Delft, we built a CPU, and then tried to emulate x86 on it. Didn't go fast.
Then a guy from HP visits. A year later HP comes with a design awkwardly similar to what we came up with. But they did emulate x86 quickly. The trick to a quick emulator is that you don't have to handle corner cases. So if your architecture has an "add" instruction and leaves the flags register exactly as the emulated architecture would, then you'll be able to emulate quickly.
HP's x86 effort merged into a cooperation with Intel. HP dropped out. Our design now powers Pentium IV chips behind the scenes.
The good thing is that besides generating better keys after the patch, patched debian systems will refuse to authenticate on a broken key!
Thus, your broken keys become useless, and you're forced to generate new ones. Moreover, if there are "old" users on a system that no longer actively login on a system, they wouldn't notice the key no longer working. However, the patched system will refuse to authenticate the broken key.
Hmm. come to think of it, there remains a risk of people copying bad debian keys onto other not-patched Unix systems, and the bad guys figuring that one out...
If you try to jail the CEO, he will say it's the CTO's job to secure the systems. He in turn blames the head-of-IT-ops, who in turn blames the lonely sysop. So who's going to jail? All of them? The top? The bottom?
If YOU do something bad, YOU have to pay the price. We've got several gradations here: pay a fine, go to jail, both in different amounts.
If a company does something bad, what can we do to make it pay? Well, exactly that: Make it pay.
Now, if YOU know that a fine for XYZ is $1, and it's easier for you to do XYZ than something else, then you'll easily do XYZ. Besides that the chances of getting caught are usually small, the fine is such that you can easily pay up. If you have to pay $10000 as the fine most of us will think twice, and be really careful.
In the case of a big company, $10000 is nothing. So fines you put on companies should be proportional to their size. Faking profits or losses is easy. So it should be proportional to their turnover.
Here in Europe, MicroSoft got fined EUR 1 billion for ignoring antitrust laws. This is an amount that even a company like MicroSoft feels.
With several situations, legally someone is responsible. But after they have "paid" in whatever way that is, they might then be able to hold someone else responsible. For example, if I buy a stereo here in The Netherlands, I've got warranty service from the shop. They can claim: "factory warranty: 1 year" all they want, but the law gives me the right to ask the shop to fix problems in the product during a "reasonable time" no matter what they claim. (i.e. warranty: 1 week will not work either!).
So, if a company pays a fine, and finds that this evidently the fault of a certain employee, they can sue that employee afterwards.
The problem of scale then kicks in. If the company pays a $1M fine, but this is evidently the fault of precisely one employee. (Say he was told not to do X, but he did so anyway, finding clever ways to escape the regular checks of the company to see if he was complying with the order) Then how can that single employee pay the $1M "damages" to the company?
Certificates from trusted parties should be used to certify that the certificate signed to belong to www.yourbank.com actually does belong to yourbank.
When certificate authorities break down, and issue www.yourbank.com certificates to somecrook, things break down.
The master certificate of the certificate authority that issues such bad nonsense should be revoked ASAP, and things can go on as designed.
The photos hows strawberries, and the machine suggests all kinds of fruit that have "red" in them.
Sure it's an improvement over the old system, where you always had to chose from 70 or so choices, but it's not that it only doesn't know the difference between roma tomatoes, and regular ones.
Why is your airconditioner noisy? Because you hear the pump (which is normally outside) or because lots of airflow is required along the heat exchanging elements?
So now you have a material that can cool on command by an electrical signal. Nice.
So now you make it touch your fridge, and tell it to go to the "cool state". Next it absorbs heat (that leaked through the walls of the fridge), and you need to expell that heat. So now you turn it to the "warm" state, Now it's heating your fridge? No you need to make it insulated from the fridge, and thermally connected to the outside to pump the heat out. How are you going to do that?
The easiest way would be to have two of those electro-thermal-active-plastics built as a heat exchanger. One of them (the one in the "hot" state) circulates an appropriate fluid with the heat exchanger on the back. The other circulates the fluid with the heat exchanger inside the fridge.
So, how about we get rid of those nasty ozone-layer-affecting CFKs? Nice try, but no go! These ARE CFKs we're talking about. Maybe easier to contain than CFK gasses, but CFKs notheless.
Next, when your element is exchanging heat with the fridge, and it has come to an equilibirium.... Then you change it to the "warm" state. Now it becomes 12.5 degrees centigrade warmer! So my fridge is 4 degrees, and the element becomes 16.5 But in the summer my home is warmer than that (actually in the winter as well!). It has to become warmer than the environment to expell heat. So we're going to need a two-step heatpump.
So instead of a fridge with one pump, two heat exchangers, and a replacement for the CFKs for the old days, we might go to a frige with three pumps, two valves, and four CFK -containing active essential elements!
I predict that everyone will have one of these in their house in 5 years! Not!
To me it looks as if they had a serious engine failure: The color of the exhaust is not normal. It starts with some clouds, then it becomes black smoke. Normally they tune the engines to the point that there is certainly no black smoke coming from them. (that would indicate incomplete oxidation, which would mean inefficient, and you would not want to hoist unusable fuel in your rocket)
Slashdot Mirror
Client of mine says he's putting data on an external drive, but then the drive shows up empty on the other computer.
Turns out he was formatting and writing to sdb1 instead of sdc1. One of the internal drive's partitions. Ooops!
After figuring that one out I get to clean up the mess. sdb1 is supposed to be a swap partition. mkswap /dev/sdb2 .... Ooops typo! that was supposed to be sdb1! sdb2 is their active data partition.... He'd been formatting their (never used: someone had made a typo) swap partition. I had overwritten (part of) their data partition....
Oh shit!
The only way to truly verify a system is good is to do it from a known good system. For a standalone PC that means booting off known-good read-only media, usually a CDROM,
Here you have a slight problem with implementing your suggestion: The CPU boots off the read-write flash chips on the motherboard, not off the CDROM.
By popular convention here on slashdot, I didn't RTFA. That's why.
We have radios. They work for a few km. Some drivers like or don't mind driving right behind the other driver. Others prefer just to "lose track" of the other car, and then one party gets to pause for 15 minutes while the other gets a 20 minute break... Those 5 minutes means you're about 10km apart, too far for the radios.
In a few days I'll be among a group of 6 people driving over 1000km for a trip. We'll be in two cars. So when we want to plan our next stop, we'll call one of the people in the other car. We have a 66% chance of calling the not-driver-at-the-time. If we do happen to call the driver, we will pass the phone on to someone who is not driving.
So this technology would cut us off. Brilliant!
More than 20 years ago, my "basic programming" course wanted to teach us "top down" programming. Break the top level thing from the assignment into smaller pieces until you more or less have chunks that fit into one line in your programming language.
This fails horribly if you're learning that programming language.
A friend wanted me to do his assignments. I refused, but offered to help him do his assignments.... So repeatedly saying: "and how do you break that into smaller pieces" resulted in him trying to parse floating point numbers....
So.... I would suggest: First give some assignments that end up building on the previous assignment. i.e. bottom up. Build something more complex after having built some basic building blocks. THEN introduce top down design, and if at all possible, have the building blocks from previous assingments fit in!
Indeed. The western world is in a "good times" part of a cycle.
What do you think happens when an ice age subsides? Suddenly there is lots and lots of food. Population numbers of all sorts of animals skyrocket. This allows the creation of a genetic pool, ready to be tested on the next "bad times" period.
Some genetic improvements require two steps, either a slight disadvantage, but together an advantage. It is extremely unlikely that those two mutations happen in the same individual at the same time. In a good times period, the individuals with the slight disadvantage get to survive, and mate with the individuals with the other genetic change.
How long until that happens do you guess?
Evolution is generally not continuous.
Evolution is driven most strongly when there are cycles: Times are good, times are bad.
In the good times there is plenty of everything to multiply like mad, in the bad times, all the inferior products die.
Humanity is in a "multiply like mad" phase. Even the individuals with "something wrong" are surviving and reproducing. We currently live in an environment where certain medical problems are not fatal. It's simply a matter of the environment where you happen to live.
Certain fish are in a "bad times" period: Due to too much fishing by humans, only the smaller ones survive. So they have evolved to be smaller over the course of the last 20 years.
The theoretical case where hunters and prey-animals are in a constant arms-race and improve all the time is theoretical. Nothing much happens. Most of evolution happens when there is suddenly a period of fierce competition.
That's why a planet like mercury that doesn't turn on its axis, and doesn't have a moon will not have life. Too little cycles to drive evolution. (Astronomers decided that mercury would be lifeless because it's too hot before they found bacteria on earth that survive those temperatures....)
The device pictured in the video is not going to be practical.
It has a so called aspect ratio of about one (Width devided by length, if it were rectangular). This is inefficient. This is inefficient both in air and in water.
The low aspect ratio means it flies inefficient: I estimate it will fly at a glide ratio of about 1:3. So it requires a thrust of about a third of its weight to fly.
If it's going to go underwater, its going to have a specific weight that is comparable to water. That's not going to fly.
You need a higher aspect ratio to fly. At least to do so efficiently.
To be able to submerge, you need the specific weight to be about that of water. Normal submarines have a mixture of air and steel so that taking on some water makes them heavier than water. But their specific density needs to be quite close to that of water to have a little water make the difference between floating and sinking.
Having a specific density of water makes flying difficult.
The depicted vehicle with the door next to the jet engines has another disadvantage: you can't get off the plane for half an hour after landing... But thats just a practical issue that can be fixed easily....
Two sensible answers above.
I still hold to the rule of thumb: twice RAM. (I convinced SuSE and apparently red hat too, to stick to this rule a long time ago...) Anyway, it is also an "emergency brake". If a program in your system starts taking memory uncontrollably, your system will lock up for minutes at a time without swap, before it starts doing the "out of memory" thing. That usually results in killing off the wrong program. With swap your system will show signs of life during that time, and you might be able to kill the correct program in time....
Why does the space shuttle orbit earth in about one and a half hour? Because low earth orbit takes you around the earth in about 1.5 hours.
Orbital speed is over 7000 m/s and 5200 is simply not enough.
When the first digital trespassing laws were introduced, the techies started saying that they were too broad. You could also punish most "good" guys with those laws. Those proposing these laws indicated that of course, the good guys wouldn't be punished, only the guys really doing bad things....
A few years later, the guys who were considered "good guys" are now being convicted under these laws. They are bad guys because the broke the law didn't they?
As security professionals, we can spot a hole in a computer system, and know it's exploitable without actually going in and exploiting it. So we don't actually have to go in to be convinced there actually IS a security hole.
When I was a student, I reported a hole in a computer system, to be told by the sysadmin that this was NOT a hole, that what I thought was a hole was in fact not a hole, because there were further security measures in place. I then decided I had other things to do, but I still think he was wrong. As a sysadmin, it is very difficult to acknowledge that a hole exists if you can't actually see the results. As an experiment I once checked: I sent a pre-release exploitable bug with exploit to interested people and asked if it was exploitable. About 50% could not reproduce the problem, and reported: safe, bug does not exist, no security advisory neccesary.
So, in a University environment, you train students for different jobs. Some of them need to become security experts. They learn the trade by breaking into computer systems at the university. They should then report the problems they found to the "authorities", who should request enough information to close said problems.
When I was a power user, and a responsible net citizen, I called my provider, and asked them: "When during the day are your top usage moments, so that I can reduce my usage during those moments?" They refused to answer.
But they should. If you ask the power-users to reduce file-sharing bandwidth during the top hours, I bet you can cause a significant improvement in throughput. And for the power users, a script to e.g. throttle the file-sharing program at 8 PM, and set it going again at 10PM is easy to install. And it makes a minor impact on their file sharing habbits. But for the bandwidth of the provider this will make a big difference.... The provider however needs to "grow up" and politely ask their power-users to move their usage away from the top hours-of-the-day.
Twelve years ago, I didn't think I needed the ".com" domain for my company. I just registered the ".nl". Stupid mistake.
I'd really love to be able to take over the ".com" domain from the guys that have it now.
On the other side, when I had a domain someone wanted, I wouldn't be too harsh on them. However, you're calling the shots: So I would demand that they leave your Email address in place. Or you can offer them that they specify what they need of the domain (web, certain EMail addreses) and then offer them to provide those.
Roger.
P.S. Guessing which domain is even easier than with the poster....
It's wonderful to have a tiny computer, but if you need to slap on a monitor, keyboard, and mouse to use it it's really not all that tiny, is it?
What do you use a computer for then? Ah. I can guess. You use it to do Email and type documents....
This thing sound to me as a great computer for some other uses. It would be great as the controlling unit in my washing machine. An UTP connection on the machine, and I can tell the washing machine to start at a certain time, instead of now, I have to start it before work and come back to already-smelly clothes after work.
Probably desgined for use in a cube-sat. Those are 4 inches on a side. So you'd have 7/8ths of the space available for batteries, radio and experiments.
Still, I can imagine a top Nvidia engineer spending a couple of weekends on adapting the GPU to run X86 code. If that would show promising, they could put a team on it, finish the project, and make a surpise move.
You correctly state that a GPU is usually an SIMD machine. So, they have an instruction fecth for the "I" in SIMD. They also have huge IO bandwidth for the "MD" part. If you go MIMD (muliticore in modern CPU terminology), you also need the huge IO bandwidth for the MI part. That's already done! They have that.
They just might come up with a 16 or 32 core CPU. It might perform on parallel tasks say 2x better than a top-of-the-line Intel or AMD CPU. (or per core 4x worse).
Now, building something that is 4x behind Intel performance is not that hard. A friend working on CPUs at the university had his design made into a chip, and it worked! Intel CPUs were running 200MHz at the time, his ran at 40. Intel CPUs had much smaller features on the chips. Scaling down his design would easily speed it up a lot. Nvidia already has access to the top-of-the-line fabs. We didn't.
CPUs are easily bandwidth limited. Doing the IO well, means you can easily outperform a current CPU. And doing the IO well is something NVIDIA already has the plans for.
At the computer architecture lab here at the university of Delft, we built a CPU, and then tried to emulate x86 on it. Didn't go fast.
Then a guy from HP visits. A year later HP comes with a design awkwardly similar to what we came up with. But they did emulate x86 quickly. The trick to a quick emulator is that you don't have to handle corner cases. So if your architecture has an "add" instruction and leaves the flags register exactly as the emulated architecture would, then you'll be able to emulate quickly.
HP's x86 effort merged into a cooperation with Intel. HP dropped out. Our design now powers Pentium IV chips behind the scenes.
The good thing is that besides generating better keys after the patch, patched debian systems will refuse to authenticate on a broken key!
Thus, your broken keys become useless, and you're forced to generate new ones. Moreover, if there are "old" users on a system that no longer actively login on a system, they wouldn't notice the key no longer working. However, the patched system will refuse to authenticate the broken key.
Hmm. come to think of it, there remains a risk of people copying bad debian keys onto other not-patched Unix systems, and the bad guys figuring that one out...
If you try to jail the CEO, he will say it's the CTO's job to secure the systems. He in turn blames the head-of-IT-ops, who in turn blames the lonely sysop. So who's going to jail? All of them? The top? The bottom?
If YOU do something bad, YOU have to pay the price. We've got several gradations here: pay a fine, go to jail, both in different amounts.
If a company does something bad, what can we do to make it pay? Well, exactly that: Make it pay.
Now, if YOU know that a fine for XYZ is $1, and it's easier for you to do XYZ than something else, then you'll easily do XYZ. Besides that the chances of getting caught are usually small, the fine is such that you can easily pay up. If you have to pay $10000 as the fine most of us will think twice, and be really careful.
In the case of a big company, $10000 is nothing. So fines you put on companies should be proportional to their size. Faking profits or losses is easy. So it should be proportional to their turnover.
Here in Europe, MicroSoft got fined EUR 1 billion for ignoring antitrust laws. This is an amount that even a company like MicroSoft feels.
With several situations, legally someone is responsible. But after they have "paid" in whatever way that is, they might then be able to hold someone else responsible. For example, if I buy a stereo here in The Netherlands, I've got warranty service from the shop. They can claim: "factory warranty: 1 year" all they want, but the law gives me the right to ask the shop to fix problems in the product during a "reasonable time" no matter what they claim. (i.e. warranty: 1 week will not work either!).
So, if a company pays a fine, and finds that this evidently the fault of a certain employee, they can sue that employee afterwards.
The problem of scale then kicks in. If the company pays a $1M fine, but this is evidently the fault of precisely one employee. (Say he was told not to do X, but he did so anyway, finding clever ways to escape the regular checks of the company to see if he was complying with the order) Then how can that single employee pay the $1M "damages" to the company?
Certificates from trusted parties should be used to certify that the certificate signed to belong to www.yourbank.com actually does belong to yourbank.
When certificate authorities break down, and issue www.yourbank.com certificates to somecrook, things break down.
The master certificate of the certificate authority that issues such bad nonsense should be revoked ASAP, and things can go on as designed.
The photos hows strawberries, and the machine suggests all kinds of fruit that have "red" in them.
Sure it's an improvement over the old system, where you always had to chose from 70 or so choices, but it's not that it only doesn't know the difference between roma tomatoes, and regular ones.
Why is your airconditioner noisy? Because you hear the pump (which is normally outside) or because lots of airflow is required along the heat exchanging elements?
So now you have a material that can cool on command by an electrical signal. Nice.
So now you make it touch your fridge, and tell it to go to the "cool state". Next it absorbs heat (that leaked through the walls of the fridge), and you need to expell that heat. So now you turn it to the "warm" state, Now it's heating your fridge? No you need to make it insulated from the fridge, and thermally connected to the outside to pump the heat out. How are you going to do that?
The easiest way would be to have two of those electro-thermal-active-plastics built as a heat exchanger. One of them (the one in the "hot" state) circulates an appropriate fluid with the heat exchanger on the back. The other circulates the fluid with the heat exchanger inside the fridge.
So, how about we get rid of those nasty ozone-layer-affecting CFKs? Nice try, but no go! These ARE CFKs we're talking about. Maybe easier to contain than CFK gasses, but CFKs notheless.
Next, when your element is exchanging heat with the fridge, and it has come to an equilibirium.... Then you change it to the "warm" state. Now it becomes 12.5 degrees centigrade warmer! So my fridge is 4 degrees, and the element becomes 16.5 But in the summer my home is warmer than that (actually in the winter as well!). It has to become warmer than the environment to expell heat. So we're going to need a two-step heatpump.
So instead of a fridge with one pump, two heat exchangers, and a replacement for the CFKs for the old days, we might go to a frige with three pumps, two valves, and four CFK -containing active essential elements!
I predict that everyone will have one of these in their house in 5 years! Not!
To me it looks as if they had a serious engine failure: The color of the exhaust is not normal. It starts with some clouds, then it becomes black smoke. Normally they tune the engines to the point that there is certainly no black smoke coming from them. (that would indicate incomplete oxidation, which would mean inefficient, and you would not want to hoist unusable fuel in your rocket)