As a grad student researcher, you *are* an employee of the university, and a portion of your paycheck happens to pay your tuition. It's standard practice at just about every university to have such student researchers sign an intellectual property agreement which divides up who owns how much of students' work. I also had to sign one of these when I worked in a campus research lab as an undergrad.
The university could probably make a case for partial ownership of your work, on the basis that it was performed in part with university property and university consulting (professor). But there is no way they could make a case for owning all of your work unless all of you signed agreements stating that.
I think your uni just had some assholes who try to make sure that the university owns as much IP as possible, and bullies it out of students, so that every now and then they'll own something valuable and make a bit of cash from licensing it.
VLC also needs an extremely fast computer to play high definition H.264 video without skipping and dropping tons of frames. The CoreAVC codec seems to be the best available for watching hi-def media on typical computers, with CPU time to spare to do other things.
VLC is great and I use it frequently - but it has never had particularly well-optimized codecs.
X-prizes type systems are nice, but they are mostly a PR stunt. This isn't to belittle them, or deny effectiveness, but you need real money to do research, not the prospect of maybe winning a prize. You can't hire researchers and buy equipment with prize money that you haven't won yet. So there still needs to be *real* grant money, *real* research contracts, or *real* investors to accomplish anything.
Furthermore, the prize money only defrays the cost of eventual success, it costs more money to win an X prize than the prize is worth. This works out because by that point they usually have a product that will start making back the investment, but prize money cannot be confused with research funding.
And you are right, I doubt $40 million won't be enough to build a bionic eye. But they will probably make progress, and it's better than jumping straight in with $400 million. People in charge of public funding tend not to be risk-averse, so they start with small steps.
And if you in turn wish to be enlightened, then I raise your unscientific and speculative debunking with... another unscientific and speculative debunking: http://dvorak.mwbrooks.com/dissent.html
There are many discussions arguing both sides of this issue for those interested. There is certainly controversy about the validity of the Navy's Dvorak study which concluded Dvorak was superior. And yet the head of the GSA study which concluded Dvorak offered no significant improvement (or worsening) of typing speed appeared to be strongly biased in favor of retaining Qwerty as a standard, and there are scientific validity issues with this study too.
I am not aware of any definitive and scientifically valid studies which compare Dvorak and Qwerty typing speeds and training times. And there has not been any study, even the GSA study, to make a reasonable claim that Dvorak was at all worse than Qwerty.
My personal unscientific opinion, as a converted Dvorak typist, is that if I had put the effort into learning Qwerty touch-typing that I did with Dvorak, then my speed would probably be comparable in each keyboard. Currently I can do about 80WPM in Dvorak, 50WPM in Qwerty, and I use Dvorak for over 90% of my typing. Perhaps more importantly is that in my personal experience the Dvorak keyboard is much more comfortable to use because my fingers move much less from the home row positions and have less same-hand repeat letter combinations than they would with Qwerty. This alone is enough for me to keep using Dvorak in a Qwerty world. This increased comfort *may* make it easier to type at yet faster speeds, 100WPM+, but I'm sure I could hit 80WPM in Qwerty if I had a reason to.
In TFA he states: The dice are "Michigan Red Eyes", which have different colored pips for each value. The different colors make it pretty easy to count rolls. For example, if 6 yellow dots are found in the image, there were three 2s rolled, no need to worry about determining the proper grouping or orientation of pips.
It seems pretty reliable since his software is measuring both the color and number of pips at once. And I would be amazed if he didn't check it pretty thoroughly for accuracy. My main concern would be wear and tear on the dice themselves, but I'm sure he's aware of that eventuality. And the corners of dice will wear down first, not the faces.
That is SO true. I see this pretty regularly on my campus, around the semiconductor clean rooms and chem labs which are in my corner of campus. In addition to walking all around the buildings and outside with their gloves still on, people routinely hold their pens in their gloved hands in the cleanroom, and then again later with no gloves.
I think a lot of people have a simple thought process: If I'm wearing gloves then I'm being safe. They don't have that second level of thinking, which should be common sense, that gloves can have dangerous chemicals on them - which is why you're wearing them in the first place!
To be fair, in a clean room you wear the gloves also to avoid contaminating the environment and your samples. But that doesn't make me feel any better about sloppy glove protocol.
You are completely correct. Gravity is proportional to mass, and only mass. I think the main subtlety you can encounter is with a full general relativity solution when the object is spinning, but the effect is rather small.
The interesting thing about black holes is that their mass is compressed into a much smaller volume. The sun's Schwarzchild radius is about 3km, which is how big the sun would be if it were a black hole. If you were 3km from the center of the regular sun the gravitational force would be rather small, because you'd have most of the sun's mass distributed almost evenly around you. If you were 3km from the center of the black hole though, then that same mass would all be to one side of you, pulling you with incredible force. If you were at the surface of the sun, the gravitational force is less because you're a further distance from most of the sun's mass.
This is why black holes can have far more intense gravitational fields, but once you reach a distance greater than the original radius of the sun, the gravitational force you feel is the same.
Textbooks are a huge ripoff, but you should realize that your situation sounds atypical. I have *never* had a textbook cost more than about $160, and most cost about $100 when purchased new, putting my textbook costs at a maximum of $500 per quarter, typically much less. Luckily, as I got to my junior and senior years and was mainly taking engineering elective classes, many professors were smart and didn't require textbooks for us to buy, they simply recommended a few different books so that we could choose for ourselves. If the professor had really good online lecture notes, I sometimes found I didn't need the book very much at all.
But it is important to give professors' their share of the blame, publishers aren't the only problem. They shouldn't jump out and require new editions when nothing important has been added. They shouldn't assign homework problems from the textbook which forces you to buy it or jump through hoops, i.e. borrow the book from a friend who *did* waste money on it, or hope it's available at the library. They should NEVER use a nonstandard edition - I once dropped a class mainly because the professor put the homework assignments in a special edition just for our campus, and refused to create a course website with the assignments or send it out via email.
While you might ultimately be correct for mass market software, there is a lot of very expensive specialized software which probably would lose money if it were easily pirated.
Some of the engineering software that I use, mainly through my university, can cost in excess of $100k per seat depending on the optional components that you get. If it were easy to get away with I'm sure some companies would pirate it who would otherwise pay for the license, because the expense is enormous. There isn't much direct competition with specialized engineering software, and the features are far more important, so getting rid of the DRM (usually a hardware key) isn't much of a selling point. And it's honestly pretty reliable, usually the hardware key is on a license server, so a few times a year the server goes down but it's not a big deal.
I live in California, and I haven't even had an air conditioner for the last 5 years. I rarely use my heater. Granted, I live on the coast where the weather is about as perfect as it could possibly be, but even back when I lived in a hot inland desert region of California, the A/C was used a maximum of 5 months a year, probably less. Not 90% of the year.
Besides, Holland is a little far from the US. I don't think many people are willing to move overseas simply so they can save $50/month or less on their electric bill.
"Google" isn't meaning baby chatter... they had originally meant to call the website "googol", the number 10^100, but they misspelled it and decided that they liked it: http://graphics.stanford.edu/~dk/google_name_origin.html
Emergency stops aren't necessarily a problem, it has "limit" wheels front and back. If I were designing the controller, when I detected a hard stop which would destabilize the vehicle, I would apply a measured braking force until the front limit wheels hit ground and then apply full braking force. Would this be as effective as just slamming the brakes in a 3/4 wheeled vehicle? I'm not sure without running simulations or tests, but I can't imagine that the designers just forgot about this problem.
Adding a third wheel also reduces the maneuverability compared to a two-wheeled vehicle, and introduces increased losses which reduce the overall efficiency.
This is a new piece of technology which can make a smaller, lighter, more maneuverable, and hopefully cheaper vehicle viable. They have their design goals, and those goals do not include mimicking old three-wheeled vehicles, they want to use new technology to make a new vehicle and see how well it ultimately works.
Thousands of dollars? This isn't a specialized military guidance unit, it will only need gyros on par with what the Segway already has. A hobbyist can buy a good 3-axis gyro on a ready to use board for ~$100 from Sparkfun, and I'm willing to bet that Segway already knows of some good sources for buying large numbers of gyros. The prices are also always decreasing, with the ever increasing MEMs and micro-manufacturing capabilities.
Now in exchange for the gyros and software to use them you can eliminate a moving part which introduces loss, requires maintenance, and added frame size and weight to accommodate. So some simple sensors and processing power is all it takes to eliminate a physical part from the vehicle. Without seeing their specific numbers it's hard to say if this will make the vehicle any cheaper or more expensive, but it does allow it to be lighter and more efficient. And I'm sure this prototype will let them find out if the concept can be done for cheaper overall.
That's quite possible, contamination of the gate oxide with various ions is a big issue with MOSFETs. Sodium ions in particular tend to be rather common in the air, so they caused a lot of problems. If it's particularly contaminated then you can't make enhancement-mode ("normally-off") devices, because the charges stored in the oxide lowers the threshold voltage below 0V. So to turn it off, you need a negative voltage, which is much less convenient than just having ground and a positive voltage.
In a small cleanroom at my university, just a simple one for the device processing classes, the lab manager said that in the past some students had opened a window because it was very hot in the room with the oxidation and doping ovens. Now it just so happens that the Pacific Ocean is about 200 yards from this window, so we have *lots* of sodium in the air. It allegedly took months before the sodium fully dissipated, so the mosfets that students made were poor-quality until the sodium levels dropped.
I like the suggestion that people are somehow lax in security because their mail client remembers their password. Who are these guys who type the password in every 3 minutes when they check their mail?
Every three minutes? In my day we were checking our mail every 20 seconds, both ways uphill, and tapping out the password in binary!
And I'm guessing the two most popular passwords were "1" and "0".
Making the circuit board bendy is no problem, they've got plenty of substrates for that which do it with varying degrees of flexibility, thickness, and cost.
The hard part is that you need to follow new design rules for the things you put on the board: copper traces and components which have not been designed with flexibility in mind. The non-flexible components limit the amount you can bend the board to an extent much less than the substrate itself can flex, otherwise you crack the solder joint or the component. It is worth noting that it is very possible to get around a lot of these issue, using embedded components and extremely small components, but as usual it's all basically a matter of how much you are willing to spend and what tradeoffs you can make.
Universities don't hide this fact from students... when we start grad school we sign a very specific intellectual property agreement. Universities aren't stupid, they have legal departments or at least consultants. Keep in mind that we aren't *just* students, most grad student researchers are really paid employees who happen to do their work at a school instead of a private company, instead of a manager we have an adviser (professor), and we also sometimes take or TA for classes.
Just like with a regular job you are free to ask that this agreement is altered or even nullified, but that's up to you and your employer (the school). If you invent something unrelated to your job (your academic research), then of course you are free to do whatever you want with it. If you invent something directly related to your job and you want to go start a company with that invention, you might be in for a fight, just like you would be with a regular employer.
Technically you're correct, but that's still not helpful. The problem with power LEDs is the large voltage step conversion that must be made. House electricity is delivered at 120VAC or 240VAC, and white LEDs usually run at 4-5V, so you are going to have loss. Granted you can put strings of LEDs in series to reduce the magnitude of the voltage step, but then you get reliability issues (one LED fails and the string fails), and the higher output voltage increases component cost and design complexity.
There are problems associated with running LEDs with AC, even though as a diode it would show up as only half of a sine wave. With a changing voltage you have a changing current, and the relationship between the two is exponential for a diode, meaning a small increase in voltage past its so-called "forward voltage" where it is happy can create a large increase in current; if that current is big enough the diode is destroyed. So you just decrease the peak voltage, right? No, because then you aren't running the LED at full capacity, so you need more LEDs. Because the current keeps changing you also can't run it at its most efficient current.
Finally, you would never just "put it to an AC circuit", diodes are finicky loads because of their complicated I-V relationship, even if you had a string large enough to work at house supply voltages, a small overvoltage condition could destroy them, and the inevitable operating temperature fluctuations of the junction (between room temp and maybe 60-100C) would mean that you can't operate the LEDs at an optimum point, you have to design it to operate over a wide range of operating conditions, so you have to design conservatively, thus it costs more and takes up more space. In practice most LED lighting is run by switching power regulators with closed loop current sensing feedback.
This is incorrect. Modern white LEDs are quite similar to fluorescent bulbs. They use a high frequency (large bandgap) junction to generate UV light, which is "down converted" by a cocktail of phosphor chemicals, to produce a smooth output covering wide swatches of the visual light range. It is customizable, but it is based on what our eye and brain actually consider to be white (our eye and brain have a very nonlinear response to different wavelengths of light). We're not used to it since other lights are not as white as LEDs are.
There are a host of ways that you can solve start-up surge current problems, the main trade off is better solutions require more components, thus more cost. This way is simple and can work.
Of course, I can't imagine why they would go with such a simplistic design as being discussed for this. Considering the large number of LEDs, the power levels, and the desire for long-life operation, the obvious candidate is an adjustable current switch-mode driver.
Read about the Flat Earth Myth. By the early middle ages, and certainly by Columbus's time, it was almost universally accepted that the earth is round.
The reason why nobody wanted to fund Columbus's voyage is that he wildly miscalculated the distance to Asia, thinking that it was far closer than it really was. Everybody else had more or less accurately calculated the westward traveling distance to Asia and knew that current ships couldn't make that far of a journey. Coincidentally, the American continents happened to be roughly where Columbus thought Asia was supposed to be, and now he has a national holiday in his honor.
As GP stated: "P.S. If you've got some way to analytically solve any constrained optimization problem with 50+ variables, there's probably a long line of people with medals and/or piles of cash to give you."
Simplex is a numerical, not an analytical, way to solve problems. You get paid well to consult on problems like this because it's extremely useful to solve such problems, and it's not always easy to (A) get a solution and (B) do it with a minimum of computing resources. The whole point of finding analytical solutions to problems is so that you don't need to throw a consultant and a really powerful computer at the problem.
I don't have any of the "missing features" you list on my current desktop computer. I used to be a technophile, but since I became an engineer I'm too busy and lazy for that. Now I use my computer for: internet, watching TV/movies, bittorrent, Mathematica and Matlab (when I work at home), a touch of occasional photo editing (The Gimp), and word processing (Open Office).
My computer has a 3GHZ (or so) P4, a gig of ram, an old 64MB video card, and 1.25TB of hard drive space. The processor is probably way overpowered for what I need. I built it over 4 years ago and haven't had any need to upgrade it it since, and I don't expect that I will need to for another several years. If XP gets too out of date I'll move to Linux before installing the Vista resource hog. I don't know why I'd need any of the features you list unless I was running a server, doing lots of photo/video editing, or playing the latest games, but like most computer users I'm not.
To address you point by point:
ECC: Who cares, it honestly does *not* matter if you have ECC ram. 4cm fan: Have you ever seen a northbridge with a big fan? They don't need big fans. If it matters to you, take it off, replace it with a quieter one, or put on a bigger heatsink. PCI-E: Most people (including me) can get by on still-available PCI video cards. And people who buy mini-ITX video cards are not usually concerned about RAID. Side note: my work computer just got upgraded with a PCI video card because it has an apparently obscure PCI-E x8 port on the motherboard instead of x16 or AGP. Gigabit: Again, who cares. It's not a server. I rarely find myself transferring gigabytes of data between two computers on my network. Only one ethernet: When was the last desktop motherboard you bought with two ethernet ports? Old SATA: For the third time, it doesn't fucking matter even for most power users. DVI: All of the several monitors I own still use VGA. Hardware virtualization: Hell, I don't even know what that means, and I'm too lazy to google it. PS/2, IDE, parallel: I am typing on my IBM model M keyboard, it's PS/2. Ever try developing stuff to run off USB? It takes a lot more work than a parallel port to implement a USB connection. I know, because I've done both. The IDE may be unnecessary but the biggest cost is in implementing it is either board space or the connector cost itself, and both of those are pretty small. Made in a communist dictatorship: Yeah, a lot of stuff is made over there if you haven't noticed yet. "Pre-teen" Overclocking options: Why would that be bad? I won't use them, but I don't mind that they are there. It takes very little engineering, and even less production cost to include that. 12 month warranty too long: Why exactly is it "too long"? Would you prefer a 90 day warranty? Linux support: Well, that's the only potentially valid point, but since it is using an Intel chipset and Intel graphics, I bet: (A) It is supported, and (B) you can confirm or deny that point easily via a web search.
You should be modded Troll, I don't know why you're +5 Insightful.
The reason that your company's products are so energy inefficient has nothing inherent to do with the product.
It is expensive, but quite possible to make an electronically controlled dimmable CFL. They even have dimmable CFL's designed to work with standard incandescent dimmer switches - the CFL has to handle the dimming in a different way internally, but its internal electronics make it a drop-in replacement.
You don't say exactly what your modules/switches do, but I'm going to assume that among other possibilities they sit there and wait for a signal to turn on a light. A device like that does not need to consume 12-15 watts of electricity. Although most of my work is in other areas, I have designed embedded systems so I know how these work. When I was in school we were actually graded on our electronics labs on the basis of power consumption, because it's often very easy to solve problems by throwing power at them, but it is an inelegant solution, so for the best grade you must also have an efficient design. For a system like a switch waiting for a signal to turn on, you barely need any power. The power converter which takes the 120V AC and converts it to the 5V or so DC that the circuit runs on will probably use up most of the power, because it can be more expensive to make a low power circuit which has a big change in voltage very efficient. However beyond that you basically only need a microcontroller (which can drop into power saving mode, and be woken up instantly by an interrupt whenever a signal is coming in), a few external components for whatever interface is implemented (signals over power lines, separate control line, button on the wall, or wireless), and some semiconductor switches to control and dim the lights. A system like that can be built under a watt.
It honestly is not much harder to design systems to be power efficient, but sometimes people don't do it because they are either lazy or they think that it will be less reliable if they do "tricky" things like utilize power-savings mode on chips and lower bias currents.
Slashdot Mirror
They were probably bluffing.
As a grad student researcher, you *are* an employee of the university, and a portion of your paycheck happens to pay your tuition. It's standard practice at just about every university to have such student researchers sign an intellectual property agreement which divides up who owns how much of students' work. I also had to sign one of these when I worked in a campus research lab as an undergrad.
The university could probably make a case for partial ownership of your work, on the basis that it was performed in part with university property and university consulting (professor). But there is no way they could make a case for owning all of your work unless all of you signed agreements stating that.
I think your uni just had some assholes who try to make sure that the university owns as much IP as possible, and bullies it out of students, so that every now and then they'll own something valuable and make a bit of cash from licensing it.
I think your school just had some assholes
VLC also needs an extremely fast computer to play high definition H.264 video without skipping and dropping tons of frames. The CoreAVC codec seems to be the best available for watching hi-def media on typical computers, with CPU time to spare to do other things.
VLC is great and I use it frequently - but it has never had particularly well-optimized codecs.
X-prizes type systems are nice, but they are mostly a PR stunt. This isn't to belittle them, or deny effectiveness, but you need real money to do research, not the prospect of maybe winning a prize. You can't hire researchers and buy equipment with prize money that you haven't won yet. So there still needs to be *real* grant money, *real* research contracts, or *real* investors to accomplish anything.
Furthermore, the prize money only defrays the cost of eventual success, it costs more money to win an X prize than the prize is worth. This works out because by that point they usually have a product that will start making back the investment, but prize money cannot be confused with research funding.
And you are right, I doubt $40 million won't be enough to build a bionic eye. But they will probably make progress, and it's better than jumping straight in with $400 million. People in charge of public funding tend not to be risk-averse, so they start with small steps.
And if you in turn wish to be enlightened, then I raise your unscientific and speculative debunking with... another unscientific and speculative debunking: http://dvorak.mwbrooks.com/dissent.html
There are many discussions arguing both sides of this issue for those interested. There is certainly controversy about the validity of the Navy's Dvorak study which concluded Dvorak was superior. And yet the head of the GSA study which concluded Dvorak offered no significant improvement (or worsening) of typing speed appeared to be strongly biased in favor of retaining Qwerty as a standard, and there are scientific validity issues with this study too.
I am not aware of any definitive and scientifically valid studies which compare Dvorak and Qwerty typing speeds and training times. And there has not been any study, even the GSA study, to make a reasonable claim that Dvorak was at all worse than Qwerty.
My personal unscientific opinion, as a converted Dvorak typist, is that if I had put the effort into learning Qwerty touch-typing that I did with Dvorak, then my speed would probably be comparable in each keyboard. Currently I can do about 80WPM in Dvorak, 50WPM in Qwerty, and I use Dvorak for over 90% of my typing. Perhaps more importantly is that in my personal experience the Dvorak keyboard is much more comfortable to use because my fingers move much less from the home row positions and have less same-hand repeat letter combinations than they would with Qwerty. This alone is enough for me to keep using Dvorak in a Qwerty world. This increased comfort *may* make it easier to type at yet faster speeds, 100WPM+, but I'm sure I could hit 80WPM in Qwerty if I had a reason to.
In TFA he states:
The dice are "Michigan Red Eyes", which have different colored pips for each value. The different colors make it pretty easy to count rolls. For example, if 6 yellow dots are found in the image, there were three 2s rolled, no need to worry about determining the proper grouping or orientation of pips.
It seems pretty reliable since his software is measuring both the color and number of pips at once. And I would be amazed if he didn't check it pretty thoroughly for accuracy. My main concern would be wear and tear on the dice themselves, but I'm sure he's aware of that eventuality. And the corners of dice will wear down first, not the faces.
That is SO true. I see this pretty regularly on my campus, around the semiconductor clean rooms and chem labs which are in my corner of campus. In addition to walking all around the buildings and outside with their gloves still on, people routinely hold their pens in their gloved hands in the cleanroom, and then again later with no gloves.
I think a lot of people have a simple thought process: If I'm wearing gloves then I'm being safe. They don't have that second level of thinking, which should be common sense, that gloves can have dangerous chemicals on them - which is why you're wearing them in the first place!
To be fair, in a clean room you wear the gloves also to avoid contaminating the environment and your samples. But that doesn't make me feel any better about sloppy glove protocol.
You are completely correct. Gravity is proportional to mass, and only mass. I think the main subtlety you can encounter is with a full general relativity solution when the object is spinning, but the effect is rather small.
The interesting thing about black holes is that their mass is compressed into a much smaller volume. The sun's Schwarzchild radius is about 3km, which is how big the sun would be if it were a black hole. If you were 3km from the center of the regular sun the gravitational force would be rather small, because you'd have most of the sun's mass distributed almost evenly around you. If you were 3km from the center of the black hole though, then that same mass would all be to one side of you, pulling you with incredible force. If you were at the surface of the sun, the gravitational force is less because you're a further distance from most of the sun's mass.
This is why black holes can have far more intense gravitational fields, but once you reach a distance greater than the original radius of the sun, the gravitational force you feel is the same.
Textbooks are a huge ripoff, but you should realize that your situation sounds atypical. I have *never* had a textbook cost more than about $160, and most cost about $100 when purchased new, putting my textbook costs at a maximum of $500 per quarter, typically much less. Luckily, as I got to my junior and senior years and was mainly taking engineering elective classes, many professors were smart and didn't require textbooks for us to buy, they simply recommended a few different books so that we could choose for ourselves. If the professor had really good online lecture notes, I sometimes found I didn't need the book very much at all.
But it is important to give professors' their share of the blame, publishers aren't the only problem. They shouldn't jump out and require new editions when nothing important has been added. They shouldn't assign homework problems from the textbook which forces you to buy it or jump through hoops, i.e. borrow the book from a friend who *did* waste money on it, or hope it's available at the library. They should NEVER use a nonstandard edition - I once dropped a class mainly because the professor put the homework assignments in a special edition just for our campus, and refused to create a course website with the assignments or send it out via email.
While you might ultimately be correct for mass market software, there is a lot of very expensive specialized software which probably would lose money if it were easily pirated.
Some of the engineering software that I use, mainly through my university, can cost in excess of $100k per seat depending on the optional components that you get. If it were easy to get away with I'm sure some companies would pirate it who would otherwise pay for the license, because the expense is enormous. There isn't much direct competition with specialized engineering software, and the features are far more important, so getting rid of the DRM (usually a hardware key) isn't much of a selling point. And it's honestly pretty reliable, usually the hardware key is on a license server, so a few times a year the server goes down but it's not a big deal.
I live in California, and I haven't even had an air conditioner for the last 5 years. I rarely use my heater. Granted, I live on the coast where the weather is about as perfect as it could possibly be, but even back when I lived in a hot inland desert region of California, the A/C was used a maximum of 5 months a year, probably less. Not 90% of the year.
Besides, Holland is a little far from the US. I don't think many people are willing to move overseas simply so they can save $50/month or less on their electric bill.
"Google" isn't meaning baby chatter... they had originally meant to call the website "googol", the number 10^100, but they misspelled it and decided that they liked it:
http://graphics.stanford.edu/~dk/google_name_origin.html
Emergency stops aren't necessarily a problem, it has "limit" wheels front and back. If I were designing the controller, when I detected a hard stop which would destabilize the vehicle, I would apply a measured braking force until the front limit wheels hit ground and then apply full braking force. Would this be as effective as just slamming the brakes in a 3/4 wheeled vehicle? I'm not sure without running simulations or tests, but I can't imagine that the designers just forgot about this problem.
Adding a third wheel also reduces the maneuverability compared to a two-wheeled vehicle, and introduces increased losses which reduce the overall efficiency.
This is a new piece of technology which can make a smaller, lighter, more maneuverable, and hopefully cheaper vehicle viable. They have their design goals, and those goals do not include mimicking old three-wheeled vehicles, they want to use new technology to make a new vehicle and see how well it ultimately works.
Thousands of dollars? This isn't a specialized military guidance unit, it will only need gyros on par with what the Segway already has. A hobbyist can buy a good 3-axis gyro on a ready to use board for ~$100 from Sparkfun, and I'm willing to bet that Segway already knows of some good sources for buying large numbers of gyros. The prices are also always decreasing, with the ever increasing MEMs and micro-manufacturing capabilities.
Now in exchange for the gyros and software to use them you can eliminate a moving part which introduces loss, requires maintenance, and added frame size and weight to accommodate. So some simple sensors and processing power is all it takes to eliminate a physical part from the vehicle. Without seeing their specific numbers it's hard to say if this will make the vehicle any cheaper or more expensive, but it does allow it to be lighter and more efficient. And I'm sure this prototype will let them find out if the concept can be done for cheaper overall.
That's quite possible, contamination of the gate oxide with various ions is a big issue with MOSFETs. Sodium ions in particular tend to be rather common in the air, so they caused a lot of problems. If it's particularly contaminated then you can't make enhancement-mode ("normally-off") devices, because the charges stored in the oxide lowers the threshold voltage below 0V. So to turn it off, you need a negative voltage, which is much less convenient than just having ground and a positive voltage.
In a small cleanroom at my university, just a simple one for the device processing classes, the lab manager said that in the past some students had opened a window because it was very hot in the room with the oxidation and doping ovens. Now it just so happens that the Pacific Ocean is about 200 yards from this window, so we have *lots* of sodium in the air. It allegedly took months before the sodium fully dissipated, so the mosfets that students made were poor-quality until the sodium levels dropped.
Lisa: "But then we're stuck with anti-anti-fanbois!"
Skinner: "When wintertime rolls around, the anti-anti-fanbois will simply freeze to death"
See, it's simple!
I like the suggestion that people are somehow lax in security because their mail client remembers their password. Who are these guys who type the password in every 3 minutes when they check their mail?
Every three minutes? In my day we were checking our mail every 20 seconds, both ways uphill, and tapping out the password in binary!
And I'm guessing the two most popular passwords were "1" and "0".
Making the circuit board bendy is no problem, they've got plenty of substrates for that which do it with varying degrees of flexibility, thickness, and cost.
The hard part is that you need to follow new design rules for the things you put on the board: copper traces and components which have not been designed with flexibility in mind. The non-flexible components limit the amount you can bend the board to an extent much less than the substrate itself can flex, otherwise you crack the solder joint or the component. It is worth noting that it is very possible to get around a lot of these issue, using embedded components and extremely small components, but as usual it's all basically a matter of how much you are willing to spend and what tradeoffs you can make.
Universities don't hide this fact from students... when we start grad school we sign a very specific intellectual property agreement. Universities aren't stupid, they have legal departments or at least consultants. Keep in mind that we aren't *just* students, most grad student researchers are really paid employees who happen to do their work at a school instead of a private company, instead of a manager we have an adviser (professor), and we also sometimes take or TA for classes.
Just like with a regular job you are free to ask that this agreement is altered or even nullified, but that's up to you and your employer (the school). If you invent something unrelated to your job (your academic research), then of course you are free to do whatever you want with it. If you invent something directly related to your job and you want to go start a company with that invention, you might be in for a fight, just like you would be with a regular employer.
Technically you're correct, but that's still not helpful. The problem with power LEDs is the large voltage step conversion that must be made. House electricity is delivered at 120VAC or 240VAC, and white LEDs usually run at 4-5V, so you are going to have loss. Granted you can put strings of LEDs in series to reduce the magnitude of the voltage step, but then you get reliability issues (one LED fails and the string fails), and the higher output voltage increases component cost and design complexity.
There are problems associated with running LEDs with AC, even though as a diode it would show up as only half of a sine wave. With a changing voltage you have a changing current, and the relationship between the two is exponential for a diode, meaning a small increase in voltage past its so-called "forward voltage" where it is happy can create a large increase in current; if that current is big enough the diode is destroyed. So you just decrease the peak voltage, right? No, because then you aren't running the LED at full capacity, so you need more LEDs. Because the current keeps changing you also can't run it at its most efficient current.
Finally, you would never just "put it to an AC circuit", diodes are finicky loads because of their complicated I-V relationship, even if you had a string large enough to work at house supply voltages, a small overvoltage condition could destroy them, and the inevitable operating temperature fluctuations of the junction (between room temp and maybe 60-100C) would mean that you can't operate the LEDs at an optimum point, you have to design it to operate over a wide range of operating conditions, so you have to design conservatively, thus it costs more and takes up more space. In practice most LED lighting is run by switching power regulators with closed loop current sensing feedback.
This is incorrect. Modern white LEDs are quite similar to fluorescent bulbs. They use a high frequency (large bandgap) junction to generate UV light, which is "down converted" by a cocktail of phosphor chemicals, to produce a smooth output covering wide swatches of the visual light range. It is customizable, but it is based on what our eye and brain actually consider to be white (our eye and brain have a very nonlinear response to different wavelengths of light). We're not used to it since other lights are not as white as LEDs are.
Look and page 19 and 20 of this PDF to see what I mean: http://www.philipslumileds.com/pdfs/DS51.pdf
In a word, yes.
There are a host of ways that you can solve start-up surge current problems, the main trade off is better solutions require more components, thus more cost. This way is simple and can work.
Of course, I can't imagine why they would go with such a simplistic design as being discussed for this. Considering the large number of LEDs, the power levels, and the desire for long-life operation, the obvious candidate is an adjustable current switch-mode driver.
Read about the Flat Earth Myth. By the early middle ages, and certainly by Columbus's time, it was almost universally accepted that the earth is round.
The reason why nobody wanted to fund Columbus's voyage is that he wildly miscalculated the distance to Asia, thinking that it was far closer than it really was. Everybody else had more or less accurately calculated the westward traveling distance to Asia and knew that current ships couldn't make that far of a journey. Coincidentally, the American continents happened to be roughly where Columbus thought Asia was supposed to be, and now he has a national holiday in his honor.
As GP stated: "P.S. If you've got some way to analytically solve any constrained optimization problem with 50+ variables, there's probably a long line of people with medals and/or piles of cash to give you."
Simplex is a numerical, not an analytical, way to solve problems. You get paid well to consult on problems like this because it's extremely useful to solve such problems, and it's not always easy to (A) get a solution and (B) do it with a minimum of computing resources. The whole point of finding analytical solutions to problems is so that you don't need to throw a consultant and a really powerful computer at the problem.
I don't have any of the "missing features" you list on my current desktop computer. I used to be a technophile, but since I became an engineer I'm too busy and lazy for that. Now I use my computer for: internet, watching TV/movies, bittorrent, Mathematica and Matlab (when I work at home), a touch of occasional photo editing (The Gimp), and word processing (Open Office).
My computer has a 3GHZ (or so) P4, a gig of ram, an old 64MB video card, and 1.25TB of hard drive space. The processor is probably way overpowered for what I need. I built it over 4 years ago and haven't had any need to upgrade it it since, and I don't expect that I will need to for another several years. If XP gets too out of date I'll move to Linux before installing the Vista resource hog. I don't know why I'd need any of the features you list unless I was running a server, doing lots of photo/video editing, or playing the latest games, but like most computer users I'm not.
To address you point by point:
ECC: Who cares, it honestly does *not* matter if you have ECC ram.
4cm fan: Have you ever seen a northbridge with a big fan? They don't need big fans. If it matters to you, take it off, replace it with a quieter one, or put on a bigger heatsink.
PCI-E: Most people (including me) can get by on still-available PCI video cards. And people who buy mini-ITX video cards are not usually concerned about RAID. Side note: my work computer just got upgraded with a PCI video card because it has an apparently obscure PCI-E x8 port on the motherboard instead of x16 or AGP.
Gigabit: Again, who cares. It's not a server. I rarely find myself transferring gigabytes of data between two computers on my network.
Only one ethernet: When was the last desktop motherboard you bought with two ethernet ports?
Old SATA: For the third time, it doesn't fucking matter even for most power users.
DVI: All of the several monitors I own still use VGA.
Hardware virtualization: Hell, I don't even know what that means, and I'm too lazy to google it.
PS/2, IDE, parallel: I am typing on my IBM model M keyboard, it's PS/2. Ever try developing stuff to run off USB? It takes a lot more work than a parallel port to implement a USB connection. I know, because I've done both. The IDE may be unnecessary but the biggest cost is in implementing it is either board space or the connector cost itself, and both of those are pretty small.
Made in a communist dictatorship: Yeah, a lot of stuff is made over there if you haven't noticed yet.
"Pre-teen" Overclocking options: Why would that be bad? I won't use them, but I don't mind that they are there. It takes very little engineering, and even less production cost to include that.
12 month warranty too long: Why exactly is it "too long"? Would you prefer a 90 day warranty?
Linux support: Well, that's the only potentially valid point, but since it is using an Intel chipset and Intel graphics, I bet: (A) It is supported, and (B) you can confirm or deny that point easily via a web search.
You should be modded Troll, I don't know why you're +5 Insightful.
The reason that your company's products are so energy inefficient has nothing inherent to do with the product.
It is expensive, but quite possible to make an electronically controlled dimmable CFL. They even have dimmable CFL's designed to work with standard incandescent dimmer switches - the CFL has to handle the dimming in a different way internally, but its internal electronics make it a drop-in replacement.
You don't say exactly what your modules/switches do, but I'm going to assume that among other possibilities they sit there and wait for a signal to turn on a light. A device like that does not need to consume 12-15 watts of electricity. Although most of my work is in other areas, I have designed embedded systems so I know how these work. When I was in school we were actually graded on our electronics labs on the basis of power consumption, because it's often very easy to solve problems by throwing power at them, but it is an inelegant solution, so for the best grade you must also have an efficient design. For a system like a switch waiting for a signal to turn on, you barely need any power. The power converter which takes the 120V AC and converts it to the 5V or so DC that the circuit runs on will probably use up most of the power, because it can be more expensive to make a low power circuit which has a big change in voltage very efficient. However beyond that you basically only need a microcontroller (which can drop into power saving mode, and be woken up instantly by an interrupt whenever a signal is coming in), a few external components for whatever interface is implemented (signals over power lines, separate control line, button on the wall, or wireless), and some semiconductor switches to control and dim the lights. A system like that can be built under a watt.
It honestly is not much harder to design systems to be power efficient, but sometimes people don't do it because they are either lazy or they think that it will be less reliable if they do "tricky" things like utilize power-savings mode on chips and lower bias currents.