When I was in high school in the 80's and was taught the Wien and black body radiation model, our teacher told us that there is way to produce a "black body surface" which absorbs virtually all incoming radiation (more than the TFA:s 98-99%), and that is to make a chamber with the inside walls painted black and to drill a hole in one of the walls of the chamber. That orifice will be a very, very good approximation of a black body.
Granted, it's kind of hard to construct objects with orifices only...
... that determines its success or not in a non-nieche segment.
It's the mass of developers that already know it, it's the accumulated code base (both locally and globally) and most importantly: the eco system of tools surrounding it: the compilers, the IDEs, the debuggers, the static/dynamic code analysis, build systems, code generators, mock tools, coverage tools etc.
The new kids on the block have a lot of catching up to do in areas which are not directly language related.
the Mars One people. They'll have it up and running in 2017. Heck, they may event skip the rockets, why not build this thing from Los Angeles to the first city on mars?
Agile doesn't make a crappy team good. An excellent team is good with or without Agile.
Too much effort is spent on methodology compared to construction of teams which work really well. Usually, people just end up in a team. "Here's Steve, he starts today. Hi Steve!".
Setting up an excellent team is usually very hard, and take a lot of skill. I.e. it requires excellent leadership. I've seen it once in 20 years, and fortunately, that time is now. I'm happy to go to work, despite crappy customers, budget constraints, slow hardware. The people are awesome and make the best of it.
You need a wafer with the same crystal structure and lattice constant. If there is a mismatch between the inter-atomic distance (aka lattice parameter or lattice constant), the atoms deposited on this wafer will try to adjust to this lattice.
If the layer is thin, the deposited crystal will in effect be compressed or expanded. While this is OK from a mechanical and crystal point of view, the electronic properties of the grown semiconductor will change. E.g. the bandgap (energy distance between filled and empty energy levels) will shift, which will change the electronic properties of the material.
If/when the layer becomes thick (~5 atomic layers), the grown crystal will (try to) revert to its native lattice constant. However, there's no ordered way to do this, so the grown crystal will contain lots of defects, or on the worst case, become amorphous (loose its crystal structure alltogether). Defects destroy the material from an electronic point; it provides ample opportunities for electrones and holes to recombine. It increases the leakage current and power dissipation and alse change the electronic band structure.
I haven't read the article, but what I don't understand how they intend to separate the circuitry from the wafer... it's not exactly a tape that you can peel off, or a thick slab which you can hammer away.
Yes, seriously. Traffic Sign Recognition is a very easy problem, but it's not directly related to safety concerns. Usually the speed can be picked up from the navigation system, so TSR is mostly improving the accuracy of the speed detection. Using standard classification techniques, this is mostly a problem on how much computing power you can throw at the problem (the limiting factors, in an automotive setting, are: dissipated heat, budget, physical size and environmental resilience). It can be mostly be solved using well-known computer vision mechanisms and there's no merit at all in showing "current speed". That's 101-level.
Lane detection is directly related to safety. It supports services like Lane Keep Assist, Lane Departure Warning and Lane Departure Prevention. These are all ADAS services. Lane Depature Prevention steers (or brakes) you back into your lane if you inadvertantly change lanes. The safety analysis of this function alone is many, many man-years of effort. Let me take one example: wrong sign. A hazard is that the activation is done with the wrong sign, sending your vehicle into the lane of oncoming traffic. How do you design your system (from visual detection throughout the car, to the braking system) to make sure that there's not a bit flip anywhere (in any message or code executing in RAM or flash) that changes the sign of the activation?
These additional safety (ADAS) systems are constantly on, while the situations where they save you from an accident are constitute a vanishingly small percentage of the time. You have to *prove* that the additional safety it brings in that tiny, tine time, is not invalidated by a small risk in the normal driving use case. So, a TP rate of 0.9 cannot be combined with a FP rate of 0.1. The FP rate has to be on the order of 1E-6 given the large accumuluated normal driving time.
If we're talking about AD, then the TP rate must be upped to 5 or 6 nines, while the FP rate is still 1E-5 or 1E-6.
So, if you think you can solve this, I have a well-paid job opening for you.
People just don't get that AD is severly plauged by the curse of dimensionality (http://en.wikipedia.org/wiki/Curse_of_dimensionality).
How do you define test coverage for "traffic situations"? How would it at all be possible to make a "safe system" without even a concept of how to define or measure test coverage? How do you *prove* that the remainder of the situations pose an acceptable risk to the public?
... that work on the new holy grail, autonomous vehicles. Somehow, the level of confidence in this new technology seems to be inversely proportional to the distance to the nitty-gritty details of actually doing this. Can someone please tell me, exactly, how this is supposed to be done? Without using the phrase "how hard can it be".
Let's take the simplest of all the detection problems. How many lines of code does it take to reliably and safely detect the lane markings of a road? Nobody knows, because nobody has done it yet. Yes, there are prototypes that can handle some sub sets of all cases. The best I've seen handles 90% of the cases. That takes 1 MSLOC and still counting. How expensive will the last 10% be? How many hours of recorded video data does it take to verify the last 10%? The last 1%? The 90% takes a room full of TB harddisks and thousands of units parallel verification.
But yeah, how hard can it be to make a fully autonomous vehicle? I'll bet we'll have the fusion, flying car and AI analog: constantly 30 years in the future with winters interspearsed.
T=30fs, P=1PW => E=30J Each pulse carries about 30 joules of energy. The pulse only lasts for 30 fs. In that time, it travels s=30fs*300 000 000 m/s = 9 um. The pulse is thus only 9 microns thick.
They don't state the wavelength in the article, but since they say laser and not maser or IR laser, it's visible. With a nice green colour, i.e. 500 nm, the pulse is only 18 wavelengths long.
The energy of each (green) photon is E=hc/lamda = 2.2E-20 J. Thus, each pulse packs about 30J/2.2E-20J = 1.36E21 photons. [That's 1.5E26 photons per meter (mass of earth ~5E24 kg).]
Good job, coordinating that rather sizable pack of riotous photons in a timely manner.
This "four booter rocket" configuration is not new to the russians. It was introduced by Sergei Korlev http://en.wikipedia.org/wiki/S... already in the 50:s, with the R7 line of rockets http://en.wikipedia.org/wiki/R... . In fact, it was one such R7 carrier rocket (8K71PS) that launched Sputnik-1 (and -2) into orbit. The detachment of the booster rockts were such a common sight, that it got its own name: Korolev cross https://www.google.se/search?q....
Tor falls into the same category as many other items which can be used for both good and bad. A knife can be used to cut bread, but also throats. Morphine is be best pain killer there is, but is also a killer when abused. Bitcoin (& co) can revolutionize the monetary system, but also be used for non-tracable financing for all sorts of illegal activities.
"Non-tracable financing for all sorts of illegal activities" is also a a well-known property of... cash! Cash is not really under political questioning (a bit more so from banking): it's common, under relative control, and it's not new.
This leads me to believe that the banning things which can have adverse side effects is not primarily motivated by care for the public best, but rather fear of the new/unknown and fear of loosing control.
With Si, it's easy to create a great gate insulator, just oxidize the silicon, and you're done. With GaAs/InGaAs, there's still research going on how to create a good gate inulator. A paper from 2012 (http://iopscience.iop.org/0268-1242/27/11/115002) proposes to use Al oxide, which complicates things by adding another material and and an extra process step to the manufacturing.
On a side note: in the paper, the GaAs gate length was 1.5 um. A bit more than 5 nm.
The material part of the cost of the chip is likely to go up. I think however, that part today is minuscle, so that part of the price impact with be small. However, I do think the volume benefits to Si technology (50 years of development and industrial support, and with 13 gazillion Si units produced every year) will be very, very hard to beat with any III-V technology. There's so much new stuff to be done: defect density, passivation, via technology, lithography chemistry etc. The investment in III-V to reach current Si position will be huge and ultimately paid by the customers with higher unit prices.
If you're familiar with Daniel Kahneman's works, you'll immediately recognize that this
prediction suffers from the "validity illusion". Kahneman and associates have identified
under which circumstances you can trust the predictions of an expert . They've found out
that in order to create an expert with some kind of credibility when it comes to prediction,
the feedback loop of the system of study must be short and quick. This makes it possible
for the entity in the mind called "system 1", to actually be trained to the to a sufficient
level to make experience-based predictions with some kind of actual validity. Examples
of these kind of experts are doctors, firefighters and athletes; they've dealt with the same
types of situations hundreds and thousands times forming a vast, intuitive base of experience.
Financial and political experts do not have that kind of feedback and therefore their predictions
are naturally much less valid.
How many times have the author experienced the "break through" of an alternative desktop?
His prediction shall be viewed in that light....
Slashdot Mirror
When I was in high school in the 80's and was taught the Wien and black body radiation model, our teacher told us that there is way to produce a "black body surface" which absorbs virtually all incoming radiation (more than the TFA:s 98-99%), and that is to make a chamber with the inside walls painted black and to drill a hole in one of the walls of the chamber. That orifice will be a very, very good approximation of a black body.
Granted, it's kind of hard to construct objects with orifices only...
http://www.theonion.com/articl...
It definitely Donald Trumps it!
... that determines its success or not in a non-nieche segment.
It's the mass of developers that already know it, it's the accumulated code base (both locally and globally) and most importantly: the eco system of tools surrounding it: the compilers, the IDEs, the debuggers, the static/dynamic code analysis, build systems, code generators, mock tools, coverage tools etc.
The new kids on the block have a lot of catching up to do in areas which are not directly language related.
To prevent that the vacuum collapse spreads beyond LHC, just put it into a vacuum-vacuum flask!
the Mars One people. They'll have it up and running in 2017. Heck, they may event skip the rockets, why not build this thing from Los Angeles to the first city on mars?
Thake the tube to mars. How difficult can it be?
Betterige's law (aka Hinchliffe's Rule) is neither true, nor false.
Since 1995, it cannot be evaluated, see https://newtonexcelbach.wordpr...
Gödel and Heisenberg would have been proud!
Agile doesn't make a crappy team good.
An excellent team is good with or without Agile.
Too much effort is spent on methodology compared to construction of teams which work really well.
Usually, people just end up in a team. "Here's Steve, he starts today. Hi Steve!".
Setting up an excellent team is usually very hard, and take a lot of skill. I.e. it requires excellent leadership.
I've seen it once in 20 years, and fortunately, that time is now. I'm happy to go to work, despite crappy
customers, budget constraints, slow hardware. The people are awesome and make the best of it.
Wow, then they've managed to scale up production to a whopping 0.00003% of the US consumption http://en.wikipedia.org/wiki/D....
You need a wafer with the same crystal structure and lattice constant. If there is a mismatch
between the inter-atomic distance (aka lattice parameter or lattice constant), the atoms
deposited on this wafer will try to adjust to this lattice.
If the layer is thin, the deposited crystal will in effect be compressed or expanded. While this is OK from a mechanical and
crystal point of view, the electronic properties of the grown semiconductor will change. E.g. the bandgap (energy distance
between filled and empty energy levels) will shift, which will change the electronic properties of the material.
If/when the layer becomes thick (~5 atomic layers), the grown crystal will (try to) revert to its native lattice constant. However, there's
no ordered way to do this, so the grown crystal will contain lots of defects, or on the worst case, become amorphous
(loose its crystal structure alltogether). Defects destroy the material from an electronic point; it provides ample opportunities
for electrones and holes to recombine. It increases the leakage current and power dissipation and alse change the electronic band structure.
I haven't read the article, but what I don't understand how they intend to separate the circuitry from the wafer... it's not exactly
a tape that you can peel off, or a thick slab which you can hammer away.
Or, if you choose the year on Earth, with 365 days, that's about 34 000 days.
Yes, seriously. Traffic Sign Recognition is a very easy problem, but it's not directly related to safety concerns. Usually the speed can be picked up from the navigation system, so TSR is mostly improving the accuracy of the speed detection. Using standard classification techniques, this is mostly a problem on how much computing power you can throw at the problem (the limiting factors, in an automotive setting, are: dissipated heat, budget, physical size and environmental resilience). It can be mostly be solved using well-known computer vision mechanisms and there's no merit at all in showing "current speed". That's 101-level.
Lane detection is directly related to safety. It supports services like Lane Keep Assist, Lane Departure Warning and Lane Departure Prevention. These are all ADAS services. Lane Depature Prevention steers (or brakes) you back into your lane if you inadvertantly change lanes. The safety analysis of this function alone is many, many man-years of effort. Let me take one example: wrong sign. A hazard is that the activation is done with the wrong sign, sending your vehicle into the lane of oncoming traffic. How do you design your system (from visual detection throughout the car, to the braking system) to make sure that there's not a bit flip anywhere (in any message or code executing in RAM or flash) that changes the sign of the activation?
These additional safety (ADAS) systems are constantly on, while the situations where they save you from an accident are constitute a vanishingly small percentage of the time. You have to *prove* that the additional safety it brings in that tiny, tine time, is not invalidated by a small risk in the normal driving use case. So, a TP rate of 0.9 cannot be combined with a FP rate of 0.1. The FP rate has to be on the order of 1E-6 given the large accumuluated normal driving time.
If we're talking about AD, then the TP rate must be upped to 5 or 6 nines, while the FP rate is still 1E-5 or 1E-6.
So, if you think you can solve this, I have a well-paid job opening for you.
(TP=true positive, FP=false positive)
Amen to that.
People just don't get that AD is severly plauged by the curse of dimensionality (http://en.wikipedia.org/wiki/Curse_of_dimensionality).
How do you define test coverage for "traffic situations"? How would it at all be possible to make a "safe system" without even a concept of how to define or measure test coverage? How do you *prove* that the remainder of the situations pose an acceptable risk to the public?
... that work on the new holy grail, autonomous vehicles. Somehow, the level of confidence in this new technology seems to be inversely proportional to the distance to the nitty-gritty details of actually doing this. Can someone please tell me, exactly, how this is supposed to be done? Without using the phrase "how hard can it be".
Let's take the simplest of all the detection problems. How many lines of code does it take to reliably and safely detect the lane markings of a road? Nobody knows, because nobody has done it yet. Yes, there are prototypes that can handle some sub sets of all cases. The best I've seen handles 90% of the cases. That takes 1 MSLOC and still counting. How expensive will the last 10% be? How many hours of recorded video data does it take to verify the last 10%? The last 1%? The 90% takes a room full of TB harddisks and thousands of units parallel verification.
But yeah, how hard can it be to make a fully autonomous vehicle? I'll bet we'll have the fusion, flying car and AI analog: constantly 30 years in the future with winters interspearsed.
A standard solar panel (like http://www.wholesalesolar.com/...) is about $150/m^2.
The size of California: 423 000 km^2. One-fifth of 8% of that, to meet the current need, is about 6768 km^2.
At $150/m^2, that would be approximately $1E12. That's [only] $26 000 per citizen. Start the haggle!
T=30fs, P=1PW => E=30J
Each pulse carries about 30 joules of energy.
The pulse only lasts for 30 fs. In that time, it travels s=30fs*300 000 000 m/s = 9 um.
The pulse is thus only 9 microns thick.
They don't state the wavelength in the article, but since they say laser and not maser or IR laser, it's visible.
With a nice green colour, i.e. 500 nm, the pulse is only 18 wavelengths long.
The energy of each (green) photon is E=hc/lamda = 2.2E-20 J. Thus, each pulse packs about
30J/2.2E-20J = 1.36E21 photons. [That's 1.5E26 photons per meter (mass of earth ~5E24 kg).]
Good job, coordinating that rather sizable pack of riotous photons in a timely manner.
This "four booter rocket" configuration is not new to the russians. It was introduced by Sergei Korlev http://en.wikipedia.org/wiki/S... already in the 50:s, with the R7 line of rockets http://en.wikipedia.org/wiki/R... . In fact, it was one such R7 carrier rocket (8K71PS) that launched Sputnik-1 (and -2) into orbit. The detachment of the booster rockts were such a common sight, that it got its own name: Korolev cross https://www.google.se/search?q....
Weren't they afraid that the Cape Town archipelago https://www.google.se/search?q... would hereafter be named a bit differently?
Tor falls into the same category as many other items which can be used for both good and bad.
A knife can be used to cut bread, but also throats. Morphine is be best pain killer there is,
but is also a killer when abused. Bitcoin (& co) can revolutionize the monetary system,
but also be used for non-tracable financing for all sorts of illegal activities.
"Non-tracable financing for all sorts of illegal activities" is also a a well-known property of... cash!
Cash is not really under political questioning (a bit more so from banking): it's common,
under relative control, and it's not new.
This leads me to believe that the banning things which can have adverse side effects
is not primarily motivated by care for the public best, but rather fear of the new/unknown and
fear of loosing control.
I doubt that, there are no sharks in space.
No, not yet. But at this rate, it's just a matter of time.
http://en.wikipedia.org/wiki/S...
No, it wasn't a laser test.It was my weather satellite joke above that back-fired.
Even the military grade tech deteriorates. Surely it should withstand a two degree increase - especially over a century!
With Si, it's easy to create a great gate insulator, just oxidize the silicon,
and you're done. With GaAs/InGaAs, there's still research going on how
to create a good gate inulator. A paper from 2012
(http://iopscience.iop.org/0268-1242/27/11/115002) proposes to use Al oxide,
which complicates things by adding another material and and an extra process
step to the manufacturing.
On a side note: in the paper, the GaAs gate length was 1.5 um. A bit more than 5 nm.
From Metal-Pages:
In: $600/kg
Ga: $220/kg
vs
Si: $3/kg
The material part of the cost of the chip is likely to go up. I think however, that part today is minuscle,
so that part of the price impact with be small. However, I do think the volume benefits to Si technology
(50 years of development and industrial support, and with 13 gazillion Si units produced every year)
will be very, very hard to beat with any III-V technology. There's so much new stuff to be done: defect
density, passivation, via technology, lithography chemistry etc. The investment in III-V to reach current Si
position will be huge and ultimately paid by the customers with higher unit prices.
If you're familiar with Daniel Kahneman's works, you'll immediately recognize that this prediction suffers from the "validity illusion". Kahneman and associates have identified under which circumstances you can trust the predictions of an expert . They've found out that in order to create an expert with some kind of credibility when it comes to prediction, the feedback loop of the system of study must be short and quick. This makes it possible for the entity in the mind called "system 1", to actually be trained to the to a sufficient level to make experience-based predictions with some kind of actual validity. Examples of these kind of experts are doctors, firefighters and athletes; they've dealt with the same types of situations hundreds and thousands times forming a vast, intuitive base of experience. Financial and political experts do not have that kind of feedback and therefore their predictions are naturally much less valid. How many times have the author experienced the "break through" of an alternative desktop? His prediction shall be viewed in that light....