Having been an addict level Everquest player twice, who eventually lost interest twice when I realized it was taking me away from other things I enjoy in life, I have one point to make..
Especially now that most MMORPG'ing is accompanied by Teamspeak etc, chances are your friend has quite good friends in the game, that he chats with, spends a lot of time with (more time probably than with any real life friend). Don't short the importance of these friendships, if he were to stop playing the game, he might feel like he was betraying friends, or walking away from people who counted on him. I know I definitely felt like that, possibly because I played an enchanter, and the guilds I was in often couldn't run certain raids unless I was on.
If you're going to get him to quit the game, you have to make sure he feels like he's not walking out on the people he's spent literally weeks or months of his time interacting with. I know I still keep in touch with a fair number of my EQ friends, and have since met several of them IRL and turned game friendships into more tangible ones. Giving him support in that process, not belittling the time he spends with his online friends, and making sure he doesn't feel like he's leaving them in a lurch is probably the most important support you can give him.
Indeed, scientists' work is certainly important, imaginative and repetitive... I'd leave out the boring part, because I think nearly everybody who has ever published a scientific paper would agree with me that doing so is actually extremely exciting and interesting. Knowing that you (and your colleagues) have discovered or observered something that nobody else ever has is something that is incredibly exciting. I just wish that there was a better method in place to convey that sense of excitement over to people before or while they are in college, since unfortunately unless you wind up working in a lab you probably don't experience it.
What is missing in that survey is "How do scientists communicate?" and the answer is in general "very poorly." The thing that is missing in science is a wealth of communicators that seek to make science interesting to those who don't have the time or inclination to study it. In the recent past that's been somewhat handled by major science fiction writers like Asimov... but now it seems to be a major void in the entire scientific enterprise, which is why so many interest groups can convincingly distort science to their own agendas.
One thing that scientists have to remember, no matter how well otherwise elucidated (like in Richard Hamming's Talk http://www.cs.virginia.edu/~robins/YouAndYourResea rch.html) is that unless you can communicate your research with others, no matter how brilliant, it won't matter unless you spend as much time learning how to write up your findings as you do on the actual experiments themselves.
The amusing thing about this is that for much of medieval history, many people did try to leave perpetuities for them and their families, not for anything in life, but to ensure that there were enough monks/masses for them to ensure their eternal salvation.. Same thing happened in the Ottomon empire... check out Barbara Tuchman's "A Distant Mirror" or Halil Inalcik's "The Ottoman Empire 1300-1600" as far as one can tell, the half life of any of these was ~100 years, with a max of about 300 before they were taken over... and if you look, please do so in order, since Barbara Tuchman is a much much better writer:)
I would have no problem with this if they charged for guaranteed speed above and beyond what's normally available, which would invovle building more infrastructure and then using it as a 'premium channel' so to speak, but the way this article is phrased... ie. "Apple would asked to pay 5 or 10 cents per song" and "Online game companies would be targeted" makes the whole thing sound like extortion rather than providing a value-added service for a fee.
Speaking as somebody who just this passed Wednesday had a doctor's appointment related to repetitive stress from typing..
The report is most likely technichally correct. What many people call carpel tunnel is actually various ligament overuse disorders (which are typing related), rather than nerve compression. One main way to tell, is that nearly all the wrist/forearm/elbow pain, 'itchiness' etc, is related to ligament issues, the nerve compression (which is carpel tunnel disorder) part causes numbness, 'falling alseep' type symptoms etc.
However, the ligament overuse problems, if left untreated for too long, can eventually cause carpel tunnel, because the ligaments and the nerves go through the same tunnels in the wrist, so if the ligaments are inflamed for too long, it can cause long term nerve compression and carpel tunnel disorder.
Basically my advice, is if you're having any wrist/forearm issues: see a doctor early rather than later, because it can get dramatically worse if left untreated.
You then still have to deal with the fact that neurons 'generally' connect to about ~10k others, (actual range something like 10-100k). And that's before you get to details like what neurons are where, with what densities, that long range connections in mammalian brains are generally not very well understood, etc. etc. etc.
The brain is a lot more complicated than you think. We're still many many years away from modeling a mouse brain, at a purely neuronal level. I mean, there still isn't a definitive model of the Aplysia, neuron count ~10k...
Saying we are in the 3rd millenium (2001-3000) is the same thing as saying we're in the 21st century (2001-2100), or that you're in your 25th year.... it simply implies that we're no longer in the 2nd millenium (1001-2000) or 20th century (1901-2000).
I think that I've been unclear in what I meant by the terms 'model' and 'physical representation'... which doesn't totally surprise me since most of my work involves mathematical modeling.
I did not mean to imply that some bricks-and-mortar, or any other tangible prototype had to exist. I was addressing the idea that it used to be required, before software patents, that any patent be accompanied by a clear description of a physical device, whether the patent was for the device itself, or whether that device instantiated an algorithm.
To my knowledge, except for perpetual motion devices, patents have never required a prototype to be submitted (which hasn't stopped perpetual motion devices from receiving patents without that requirement, if the language has been sufficiently obscured), but beyond things that violate generally accepted 'natural laws', I don't see any need for any prototype device to accompany the application.
Many of the problems with the current patent system go to the fact that you can now patent things that can't be represented as physical devices. Reinstating the requirement that all patents are accompanied by a detailed description of a physical device would remove all the absurd business concept patents, as well as many unreasonable types of software patents.
Things that require code, like hyper-links, one click web ordering, and other patents that most people consider ridiculous would still be protected by copyright on the code... and last time I checked MS, Oracle, Sun, IBM, Apple etc.. didn't have all that much trouble protecting their intellectual property as start-ups without software patents.
Rolling back the clock to require a description of a physical device would both make patents a lot less vague as well as making the obvious harder to obsfucate, without requiring a massive paradigm change for what patents are supposed to protect (this isn't to say that I'm not in favor a more rational system for challenging a patent, especially for prior art issues, but reform tends to move in baby-steps)
The article only mentions 5 data points over ~50 years, 1957, 1981, 1992, 1998 and now 2005.. which is not a lot to go on, likewise it mentions that the last time the current stopped was 12,000 years ago, at the end of the last Ice Age, and that it may have slowed down between 1300-1850 which was a "mini" ice-age.
I assume that the last 2 things were speculation, since the only way I could think of these things being measured is if it's somehow preserved in glacial layers etc.. could anyone who knows more explain what types of evidence back up these long term speculations? And if not, why we should draw any major conclusions from 5 data points over 50 years, when we don't know the variance of the system over hundreds or thousands of years, which 'seems' to be a 'normal' timescale for change?
I'm not saying this isn't a big deal, but the information in the article is woefully incomplete.
I guess, I did come across sounding pretty anti-fMRI, but that wasn't my intent. fMRI is an extremely valuable technique, that we can learn a lot from when it is used well. My main point is that in neuroscience in general--fMRI and optical imaging being the biggest culprits--the precision of measurement techniques, particularly spatially, is almost always considerably overstated. We need to be aware of the limitations in precision of the measurement techniques available to us, just like any other science.
On timing information, the ~50 ms, etc. is from macaque electrophysiology.. see
Lamme VAF, Super H, Spekreijse H Feedforward, horizontal, and feedback processing in the visual cortex. CURR OPIN NEUROBIOL 8: (4) 529-535 AUG 1998 or Bullier J. (2001). Integrated model of visual processing. Brain Res Brain Res Rev.36(2-3):96-107
There is also good psychophysics evidence that humans can perceive major features of a scene in ~100-200ms, RSVP (rapid serial visual presentation) experiments being an example.
Hochstein S & Ahissar M. View from the top: hierarchies and reverse hierarchies in the visual system. Neuron 2002 Dec 5;36(5):791-804.
First, most fMRI stastics are voxel-based. That is, independent statistics are done on each spatial location at the finest resolution you acquire. Then, you make some corrections for the fact that you are doing so many independent analyses. It's quite rare to "average across areas" and I'm not sure why anyone would do such a thing.
A voxel, by definition, is the smallest spatial resolution unit. So you if you're doing any non-temporal statistics there, you're getting 'noise in'-'noise out'... So I'm assuming you're referring to the fact that most fMRI experiments do many iterations for each subject and then average between them. But you run into 2 major problems: People move their heads, and even a millimeter head movement creates major noise issues compare to voxel size, particularly in 3D.. and a person's cortical representation for some function will vary depending on their history (ie. How much they've empolyed the sense in some range, or their motor control for certain muscles). So nearly all fMRI statistics routines (that I'm aware of) do inter-voxel statistics, and most of them simply do them via nearest neighbor, or nearest match after projection to an idealized brain (which DOES seperate the gyri in principle, but human brains have a surprisingly high variance in terms of lesser gyri). These methods, by their nature, don't preserve the independence of independent gyri lamina.
The other option is mapping the sulci (which I'm assuming you mention as the 2D surface... but there are not many techniques.. I'm only aware of 2 groups that have seriously pursued this.. but this isn't my specialty, so I could well be wrong), and then using some minimum distance length type process to aggregate. However, that technique is much more difficult to use, and so is not often done... nearly all of the papers I read use either just plain spatial smearing, or projection to an ideal brain which involves spatial uncertainty. And even if you DO use one of the very good techniques, you still have aggregation problems because, for example, you play the piano and I don't, your cortical areas corresponding to fingers will generally be larger than mine, even if our brain were structurally identical.
Although my professor meant that remark as a joke, it also has a fair grain of truth, in terms of the actual precision of fMRI (and that doens't even get into whether metabolic rates are mediated by average membrane potential or spike rates in any area, and other major questions).
The whole idea of 'top-down' or cognitive drive for the sensory systems is very addictive, since among other things, it allows you to explain perception as some type of baysian method. However it is simply untrue. The visual system is replete with examples, from the Craik-O'Brien-Cornsweet illusion, to stuff like Julesz' Random Dot stereograms (that CANNOT have top down effects), that defy a top-down framework.
Even with effects that might be top-down modulated (like illusory contours) the physiological evidence is totally towards these things happening in the early nervous system. Although there is definitely some feedback present even in this area, one has to consider that RC constants for most neurons are about ~10ms, and much of our perception takes place in ~100ms. These timeframes are VERY well studied, and generally accepted.. and of that 100ms, about 50ms of the time is the signal travelling from the retina to the cortex (see Bullier & DeAngelis, among others). That doesn't leave much room for dramatic top down feedback for general sensory perception.... Your visual system, bottom up, manages to figure out edges, what colors to fill them in with, various levels of depth, what's moving (in relation to your eye movements.. no easy challenge.. how can you tell when your eye moves whether you're looking at a pen, or a moving streak?) and in relation to what else, all within 100-150ms of the stimulus. That just doesn't leave time for very dramatic 'high level' feedback like this article assumes.
Although I've only mentioned vision, there are similar issues in all sensory modalities except audition, which is a special case, since audition is optimized for temporal accuity, but it has its own issues that make it look like much of your perception happens without much top-down activity.
From our current understanding it appears that top-down activity does two things: 1) Equalize 'gain' in the sensory system.. if the amplification levels across you're visual field were different, you wouldn't be able to tell whether a line was something that had to do with the outside world or noise. And 2) Modulate acuity for attention.. which is very complicated in and of itself, but there is good evidence that most early perception occurs even in areas we aren't attending to.
The main 'evidence' in this article is from a 'brain scanner' which is probably fMRI. As one of my professor's liked to say, "In fMRI we show people a picture of their ass, then a picture of a hole in the ground, and subract them." Most fMRI statistics include averaging across areas... which is nice, until you remember that our brain isn't on a sphere, but something with fissures in it, and so you just averaged two things that were (cortically speaking) in other worlds (since because of the fissure they might be centimeters apart! Remember the Cortex is a laminar archiecture around the surface)... so I'm highly skeptical, to say the least.
This is the most intelligent post that I've read on this issue..
If Mills is right, then not only QM but also Maxwell-Boltzman and our entire understanding of thermodynamics is thrown out the window.. and if he's saying that Maxwell-Boltzman is wrong then he has not only to re-explain QM, but just about everything in chemistry and (micro-)biology. All of both those fields are predicated on the essential correctness of Maxwell-Boltzman, which implies that matter, nearly all the time, must exist in its lowest energy state.
If that is incorrect, then among other issues, Mills also has to re-explain diffusion, kinetic theory, Nernst-Planck, etc, etc, etc.. that haven't been under any type of serious debate for over 50 years.
Well, then I guess you consider astronomy and astrophysics soft sciences too, since they rely overwhelmingly on non-experimental observation.
There also are experiments that show some of the major ideas in evolution, using cloned bacteria on pin-heads in completely identical conditions, that show that within 20-40 generations the bacteria start to differentiate, and the populations of these various differentiations cover different percentages of the pin. There have also been recent studies in Michigan that have drilled lake bottom cores, and found that several microbial forms of life adapt over time periods of decades to changing environmental conditions (they actually woke up spores that had been trapped in the lake silt for decades).
Not all the evidence that points towards evolution is based on fossils, or timescales that are not observable. Also, if you discard the theory of evolution, then you need some other theory with similar predictive and explanatory power, and there aren't any currently available alternatives that even remotely fit the bill (ID has no predictive power at all, and thus is not testable)--which is why I totally support the AAAS telling Kansas they can't use their planning material unless they acknowledge this fact.
I get Mathematica under an annual subscription system (called Premium Service), which allows you to always download the newest version and it works great. Of course, Wolfram also lets you buy a specific version of Mathematica and keep it, so one is not forced into the subscription model. For pieces of software that are in the vein of Mathematica (or Matlab, or any other specialized technical application), I think that this model works well, because you're going to want the updates. I personally prefer paying an annual fee and getting a new license code every year (and yes, the subscription Mathematica stops working after a certain date unless you put in the new license code, but it doesn't have to phone home) than having to shell out a new version every year or two to stay up to date.
However, for something like my operating system, or any other program that I rarely need to upgrade versions, I think this is a horrible idea, because I'm more concerned that the damn thing work, and continue to work with minimal expense and/or effort on my part. The possibility of the software not working because I don't have internet access on some day (or everyday, with some phone-home verification system) would be intolerable.
Even beyond that though, the engineering-math curricula suck as well as the teaching. I double majored in math and history in college, and many of my friends were engineers. I helped a lot of them do many assignments that were basically specialized subsets of some theory (often graph theory or combinatorics) that they didn't understand at all since they were only being taught through examples and problem sets instead starting with the concepts, so you can actually understand the detailed applications as something more than looking up the right formula.
At some point, I agree that for engineering you need a set of bootcamp classes that weed out people who, while intelligent, make silly errors like dropping negative signs, or moving decimal places (I fully admit to doing both), but that shouldn't be combined with actually trying to teach new material, or you're going to get a situation like the article author described: You don't understand it, and you can't understand why you don't understand.
Whether the author would have been a good engineer is besides the point, if he was a good engineer, he wouldn't have time to write the article. But the main point, that the teaching (whether you view it as the profs or the curriculum's fault) is horribly designed, and that a lot of people who would make good engineers get dissuaded from continuing is a good one, and is more easily examined, I think, in engineering than in most other fields.
I wound up going into grad school for cognitive & neurual science, which is something I never did in college--the reason I could do this is because I was taught about the concepts, or "how to think," and I'm doing fine despite my background. You take an IOE who probably spent more time than I did doing problems sets in any given area of math (especially since they probably started as an EE or ME), and I doubt they could make the switch, because they never really learned that optimizing the layout of a factory floor, solving a 10k+ node electrical circuit, or analyzing traffic problems all are actually based around the same area of graph theory... they just learned about the factory floor.
The other amusing thing is why everybody is saying that engineering teaching isn't highly valued. That's probably true, but the explanation for that 'research' is hilarious--engineering research?-- thats a laugh. Engineering is all about DOING stuff in the real world, my guess is that a very small proportion of really top notch engineers are at universities (as opposed to (research people in) math, english or history teachers, for example). So why the hell is the engin teaching so bad? In fact, is there any other job for a college engin professor other than teaching? If they were so good at research, the private sector is a ton more lucrative.
My advice to undergrad engins or HS'ers: If you're interested in engineering, do your undergrad degree in math, physics, statistics, etc, take a few engin courses on the side (maybe do a minor), and then if you still like engineering, grab a masters in whatever field you want to go into (math majors, for example, have no problem getting into any engineering masters program). You'll leave your options quite a bit more open, not to mention having a less stressful college experience.
You know, although this is whats called an extrinsic event, ie has nothing to do with the election per se, that doesn't mean that these events don't have highly significant correllations in a statistical sense.
There was a conference, back in 2000, at NYU on extrensic uncertainty, which presented, among other things, that sun spot activity and the stock market are correllated.
If you believe in the entire idea of 'statistical signifigance', then its difficult, whether you believe those correllations to be intrinsic or extrinsic, to ignore them.
Back in the 60s-70s, they used to try driving marked police cars 3-abreast (or however many lanes there were) down the highway at 55 to try to enforce the speed limits. The result? Accident rates in the areas that used this type of enforcement went waaay up, because the police were basically creating artifical bubbles of dense traffic that was moving much faster than the same density of traffic would normally move down the same road.
The first thing you really need to make him understand is that the number of people out there who are really interesting and motivated to learn is pretty small, and those who he meets should be valued--but since most people out there aren't terribly interested in being serious intellectuals, he needs to be able to function with them. If he's interested in economics, introduce it as an opportunity cost argument. If not, maybe introduce it to him as a prisoner's dilemma *grin*?
He doesn't need to embrace the culture of his peers, he just needs to find things that interest him that create lanes of communication. I know I wasn't interested in football when I first got to Michigan (I grew up in NYC and attended Stuy), but I quickly realized that it was something that allowed you to chat with a lot of people, and since I've always enjoyed sports, I took the time to learn a lot about it. This doesn't mean that he has to do some rote memorization job so he has something to talk about, it just means he should look into what his peers are interested in, and if any of it interests him, he should take the time to learn about it (his way).
His peers will be interested in chatting with him, and knowing him, if he can discuss the subjects they are interested in talking about it in a novel and interesting way for them. If he's smart, and interested in coming up with novel ways of thinking about things, I doubt he'd have a problem with doing this.
Your student has to understand that we have to function in the world of our peers--whatever that is. Maybe he'll eventually become an academic and be able to lock himself up in an ivory tower, or some cube farm with a whole bunch of other people that are interested in programming. But until he gets to that point, it makes sense to at least try to understand his surroundings rather than trying to make them understand him. Ask him if he thinks its easier to understand him, or its easier for him to understand his class mates?
Unfortunately not very many of us can surround ourselves completely with people who all share our interests, but I've certainly had a lot more fun socially trying to get engage myself in what my friends and aquaintances are interested in, instead of just trying to engage them in my interests.
If he needs a jump start, try getting him to do something, whether its cutting his hair, wearing jeans (or not wearing jeans), etc, that you know he thinks *I* can't do that, but is actually something minor... He's probably backed himself into a niche with his peers that will take him some effort to widen.
Outperform it in terms of Mathematical sophistication?
Whats that supposed to mean? Answer a question Mathematica can't? Nearly every college math student I know can do that.
Or perhaps answer every question that Mathematica CAN answer? I highly doubt that there's a mathematician in the world today that can do that.
Mathematica is a tool, the results you get out are only as useful as your understanding of them.
Oh, and intellectually stunted generation? Intellectually stunted because students of today no longer learn several dozen arbitrary tricks to manually solve differential equations? Pullease. Mathematica (or Matlab, or Maple etc) gives students a tool to investigate problems that were previously inaccessible--that hardly makes them intellectually stunted.
Slashdot Mirror
http://www.patentlyo.com/patent/2006/05/supreme_co urt_v.html
Having been an addict level Everquest player twice, who eventually lost interest twice when I realized it was taking me away from other things I enjoy in life, I have one point to make..
Especially now that most MMORPG'ing is accompanied by Teamspeak etc, chances are your friend has quite good friends in the game, that he chats with, spends a lot of time with (more time probably than with any real life friend). Don't short the importance of these friendships, if he were to stop playing the game, he might feel like he was betraying friends, or walking away from people who counted on him. I know I definitely felt like that, possibly because I played an enchanter, and the guilds I was in often couldn't run certain raids unless I was on.
If you're going to get him to quit the game, you have to make sure he feels like he's not walking out on the people he's spent literally weeks or months of his time interacting with. I know I still keep in touch with a fair number of my EQ friends, and have since met several of them IRL and turned game friendships into more tangible ones. Giving him support in that process, not belittling the time he spends with his online friends, and making sure he doesn't feel like he's leaving them in a lurch is probably the most important support you can give him.
Indeed, scientists' work is certainly important, imaginative and repetitive... I'd leave out the boring part, because I think nearly everybody who has ever published a scientific paper would agree with me that doing so is actually extremely exciting and interesting. Knowing that you (and your colleagues) have discovered or observered something that nobody else ever has is something that is incredibly exciting. I just wish that there was a better method in place to convey that sense of excitement over to people before or while they are in college, since unfortunately unless you wind up working in a lab you probably don't experience it.
a rch.html) is that unless you can communicate your research with others, no matter how brilliant, it won't matter unless you spend as much time learning how to write up your findings as you do on the actual experiments themselves.
What is missing in that survey is "How do scientists communicate?" and the answer is in general "very poorly." The thing that is missing in science is a wealth of communicators that seek to make science interesting to those who don't have the time or inclination to study it. In the recent past that's been somewhat handled by major science fiction writers like Asimov... but now it seems to be a major void in the entire scientific enterprise, which is why so many interest groups can convincingly distort science to their own agendas.
One thing that scientists have to remember, no matter how well otherwise elucidated (like in Richard Hamming's Talk http://www.cs.virginia.edu/~robins/YouAndYourRese
The amusing thing about this is that for much of medieval history, many people did try to leave perpetuities for them and their families, not for anything in life, but to ensure that there were enough monks/masses for them to ensure their eternal salvation.. Same thing happened in the Ottomon empire... check out Barbara Tuchman's "A Distant Mirror" or Halil Inalcik's "The Ottoman Empire 1300-1600" as far as one can tell, the half life of any of these was ~100 years, with a max of about 300 before they were taken over... and if you look, please do so in order, since Barbara Tuchman is a much much better writer :)
I would have no problem with this if they charged for guaranteed speed above and beyond what's normally available, which would invovle building more infrastructure and then using it as a 'premium channel' so to speak, but the way this article is phrased... ie. "Apple would asked to pay 5 or 10 cents per song" and "Online game companies would be targeted" makes the whole thing sound like extortion rather than providing a value-added service for a fee.
Speaking as somebody who just this passed Wednesday had a doctor's appointment related to repetitive stress from typing..
The report is most likely technichally correct. What many people call carpel tunnel is actually various ligament overuse disorders (which are typing related), rather than nerve compression. One main way to tell, is that nearly all the wrist/forearm/elbow pain, 'itchiness' etc, is related to ligament issues, the nerve compression (which is carpel tunnel disorder) part causes numbness, 'falling alseep' type symptoms etc.
However, the ligament overuse problems, if left untreated for too long, can eventually cause carpel tunnel, because the ligaments and the nerves go through the same tunnels in the wrist, so if the ligaments are inflamed for too long, it can cause long term nerve compression and carpel tunnel disorder.
Basically my advice, is if you're having any wrist/forearm issues: see a doctor early rather than later, because it can get dramatically worse if left untreated.
Modelling 'real' neurons in detail is generally done with ~10k compartmental models, which are generally described by something like:
s pikes.htm
http://neuron.duke.edu/cells/
and modelled in something like:
http://www.neuron.yale.edu/neuron/
Even using vastly simplified neurons, like integrate & fire types, for example: http://www.nsi.edu/users/izhikevich/publications/
you still have many vastly different types of spiking behaviors.
You then still have to deal with the fact that neurons 'generally' connect to about ~10k others, (actual range something like 10-100k). And that's before you get to details like what neurons are where, with what densities, that long range connections in mammalian brains are generally not very well understood, etc. etc. etc.
The brain is a lot more complicated than you think. We're still many many years away from modeling a mouse brain, at a purely neuronal level. I mean, there still isn't a definitive model of the Aplysia, neuron count ~10k...
Saying we are in the 3rd millenium (2001-3000) is the same thing as saying we're in the 21st century (2001-2100), or that you're in your 25th year.... it simply implies that we're no longer in the 2nd millenium (1001-2000) or 20th century (1901-2000).
I think that I've been unclear in what I meant by the terms 'model' and 'physical representation'... which doesn't totally surprise me since most of my work involves mathematical modeling.
I did not mean to imply that some bricks-and-mortar, or any other tangible prototype had to exist. I was addressing the idea that it used to be required, before software patents, that any patent be accompanied by a clear description of a physical device, whether the patent was for the device itself, or whether that device instantiated an algorithm.
To my knowledge, except for perpetual motion devices, patents have never required a prototype to be submitted (which hasn't stopped perpetual motion devices from receiving patents without that requirement, if the language has been sufficiently obscured), but beyond things that violate generally accepted 'natural laws', I don't see any need for any prototype device to accompany the application.
Many of the problems with the current patent system go to the fact that you can now patent things that can't be represented as physical devices. Reinstating the requirement that all patents are accompanied by a detailed description of a physical device would remove all the absurd business concept patents, as well as many unreasonable types of software patents.
h tml/srchnum.htm&Sect1=PTO1&Sect2=HITOFF&p=1&r=1&l= 50&f=G&d=PALL&s1=4803736.WKU.&OS=PN/4803736&RS=PN/ 4803736).
Before you scream that novel and non-trivial algorithms wouldn't be patenable (like, for example, a new algorithm for encoding images etc), all algorithms can be represented by specifically designed analog or digital electronics (example of a non-trivial algorithm that can also be represented by a physical device: http://patft.uspto.gov/netacgi/nph-Parser?u=/neta
Things that require code, like hyper-links, one click web ordering, and other patents that most people consider ridiculous would still be protected by copyright on the code... and last time I checked MS, Oracle, Sun, IBM, Apple etc.. didn't have all that much trouble protecting their intellectual property as start-ups without software patents.
Rolling back the clock to require a description of a physical device would both make patents a lot less vague as well as making the obvious harder to obsfucate, without requiring a massive paradigm change for what patents are supposed to protect (this isn't to say that I'm not in favor a more rational system for challenging a patent, especially for prior art issues, but reform tends to move in baby-steps)
The article only mentions 5 data points over ~50 years, 1957, 1981, 1992, 1998 and now 2005.. which is not a lot to go on, likewise it mentions that the last time the current stopped was 12,000 years ago, at the end of the last Ice Age, and that it may have slowed down between 1300-1850 which was a "mini" ice-age.
I assume that the last 2 things were speculation, since the only way I could think of these things being measured is if it's somehow preserved in glacial layers etc.. could anyone who knows more explain what types of evidence back up these long term speculations? And if not, why we should draw any major conclusions from 5 data points over 50 years, when we don't know the variance of the system over hundreds or thousands of years, which 'seems' to be a 'normal' timescale for change?
I'm not saying this isn't a big deal, but the information in the article is woefully incomplete.
I guess, I did come across sounding pretty anti-fMRI, but that wasn't my intent. fMRI is an extremely valuable technique, that we can learn a lot from when it is used well. My main point is that in neuroscience in general--fMRI and optical imaging being the biggest culprits--the precision of measurement techniques, particularly spatially, is almost always considerably overstated. We need to be aware of the limitations in precision of the measurement techniques available to us, just like any other science.
On timing information, the ~50 ms, etc. is from macaque electrophysiology.. see
Lamme VAF, Super H, Spekreijse H Feedforward, horizontal, and feedback processing in the visual cortex. CURR OPIN NEUROBIOL 8: (4) 529-535 AUG 1998
or
Bullier J. (2001). Integrated model of visual processing. Brain Res Brain Res Rev.36(2-3):96-107
There is also good psychophysics evidence that humans can perceive major features of a scene in ~100-200ms, RSVP (rapid serial visual presentation) experiments being an example.
Hochstein S & Ahissar M. View from the top: hierarchies and reverse hierarchies in the visual system. Neuron 2002 Dec 5;36(5):791-804.
Provides a decent review of that evidence.
First, most fMRI stastics are voxel-based. That is, independent statistics are done on each spatial location at the finest resolution you acquire. Then, you make some corrections for the fact that you are doing so many independent analyses. It's quite rare to "average across areas" and I'm not sure why anyone would do such a thing.
A voxel, by definition, is the smallest spatial resolution unit. So you if you're doing any non-temporal statistics there, you're getting 'noise in'-'noise out'... So I'm assuming you're referring to the fact that most fMRI experiments do many iterations for each subject and then average between them. But you run into 2 major problems: People move their heads, and even a millimeter head movement creates major noise issues compare to voxel size, particularly in 3D.. and a person's cortical representation for some function will vary depending on their history (ie. How much they've empolyed the sense in some range, or their motor control for certain muscles). So nearly all fMRI statistics routines (that I'm aware of) do inter-voxel statistics, and most of them simply do them via nearest neighbor, or nearest match after projection to an idealized brain (which DOES seperate the gyri in principle, but human brains have a surprisingly high variance in terms of lesser gyri). These methods, by their nature, don't preserve the independence of independent gyri lamina.
The other option is mapping the sulci (which I'm assuming you mention as the 2D surface... but there are not many techniques.. I'm only aware of 2 groups that have seriously pursued this.. but this isn't my specialty, so I could well be wrong), and then using some minimum distance length type process to aggregate. However, that technique is much more difficult to use, and so is not often done... nearly all of the papers I read use either just plain spatial smearing, or projection to an ideal brain which involves spatial uncertainty. And even if you DO use one of the very good techniques, you still have aggregation problems because, for example, you play the piano and I don't, your cortical areas corresponding to fingers will generally be larger than mine, even if our brain were structurally identical.
Although my professor meant that remark as a joke, it also has a fair grain of truth, in terms of the actual precision of fMRI (and that doens't even get into whether metabolic rates are mediated by average membrane potential or spike rates in any area, and other major questions).
The whole idea of 'top-down' or cognitive drive for the sensory systems is very addictive, since among other things, it allows you to explain perception as some type of baysian method. However it is simply untrue. The visual system is replete with examples, from the Craik-O'Brien-Cornsweet illusion, to stuff like Julesz' Random Dot stereograms (that CANNOT have top down effects), that defy a top-down framework.
Even with effects that might be top-down modulated (like illusory contours) the physiological evidence is totally towards these things happening in the early nervous system. Although there is definitely some feedback present even in this area, one has to consider that RC constants for most neurons are about ~10ms, and much of our perception takes place in ~100ms. These timeframes are VERY well studied, and generally accepted.. and of that 100ms, about 50ms of the time is the signal travelling from the retina to the cortex (see Bullier & DeAngelis, among others). That doesn't leave much room for dramatic top down feedback for general sensory perception.... Your visual system, bottom up, manages to figure out edges, what colors to fill them in with, various levels of depth, what's moving (in relation to your eye movements.. no easy challenge.. how can you tell when your eye moves whether you're looking at a pen, or a moving streak?) and in relation to what else, all within 100-150ms of the stimulus. That just doesn't leave time for very dramatic 'high level' feedback like this article assumes.
Although I've only mentioned vision, there are similar issues in all sensory modalities except audition, which is a special case, since audition is optimized for temporal accuity, but it has its own issues that make it look like much of your perception happens without much top-down activity.
From our current understanding it appears that top-down activity does two things: 1) Equalize 'gain' in the sensory system.. if the amplification levels across you're visual field were different, you wouldn't be able to tell whether a line was something that had to do with the outside world or noise. And 2) Modulate acuity for attention.. which is very complicated in and of itself, but there is good evidence that most early perception occurs even in areas we aren't attending to.
The main 'evidence' in this article is from a 'brain scanner' which is probably fMRI. As one of my professor's liked to say, "In fMRI we show people a picture of their ass, then a picture of a hole in the ground, and subract them." Most fMRI statistics include averaging across areas... which is nice, until you remember that our brain isn't on a sphere, but something with fissures in it, and so you just averaged two things that were (cortically speaking) in other worlds (since because of the fissure they might be centimeters apart! Remember the Cortex is a laminar archiecture around the surface)... so I'm highly skeptical, to say the least.
This is the most intelligent post that I've read on this issue..
If Mills is right, then not only QM but also Maxwell-Boltzman and our entire understanding of thermodynamics is thrown out the window.. and if he's saying that Maxwell-Boltzman is wrong then he has not only to re-explain QM, but just about everything in chemistry and (micro-)biology. All of both those fields are predicated on the essential correctness of Maxwell-Boltzman, which implies that matter, nearly all the time, must exist in its lowest energy state.
If that is incorrect, then among other issues, Mills also has to re-explain diffusion, kinetic theory, Nernst-Planck, etc, etc, etc.. that haven't been under any type of serious debate for over 50 years.
Well, then I guess you consider astronomy and astrophysics soft sciences too, since they rely overwhelmingly on non-experimental observation.
There also are experiments that show some of the major ideas in evolution, using cloned bacteria on pin-heads in completely identical conditions, that show that within 20-40 generations the bacteria start to differentiate, and the populations of these various differentiations cover different percentages of the pin. There have also been recent studies in Michigan that have drilled lake bottom cores, and found that several microbial forms of life adapt over time periods of decades to changing environmental conditions (they actually woke up spores that had been trapped in the lake silt for decades).
Not all the evidence that points towards evolution is based on fossils, or timescales that are not observable. Also, if you discard the theory of evolution, then you need some other theory with similar predictive and explanatory power, and there aren't any currently available alternatives that even remotely fit the bill (ID has no predictive power at all, and thus is not testable)--which is why I totally support the AAAS telling Kansas they can't use their planning material unless they acknowledge this fact.
I get Mathematica under an annual subscription system (called Premium Service), which allows you to always download the newest version and it works great. Of course, Wolfram also lets you buy a specific version of Mathematica and keep it, so one is not forced into the subscription model. For pieces of software that are in the vein of Mathematica (or Matlab, or any other specialized technical application), I think that this model works well, because you're going to want the updates. I personally prefer paying an annual fee and getting a new license code every year (and yes, the subscription Mathematica stops working after a certain date unless you put in the new license code, but it doesn't have to phone home) than having to shell out a new version every year or two to stay up to date.
However, for something like my operating system, or any other program that I rarely need to upgrade versions, I think this is a horrible idea, because I'm more concerned that the damn thing work, and continue to work with minimal expense and/or effort on my part. The possibility of the software not working because I don't have internet access on some day (or everyday, with some phone-home verification system) would be intolerable.
I fully agree...
Even beyond that though, the engineering-math curricula suck as well as the teaching. I double majored in math and history in college, and many of my friends were engineers. I helped a lot of them do many assignments that were basically specialized subsets of some theory (often graph theory or combinatorics) that they didn't understand at all since they were only being taught through examples and problem sets instead starting with the concepts, so you can actually understand the detailed applications as something more than looking up the right formula.
At some point, I agree that for engineering you need a set of bootcamp classes that weed out people who, while intelligent, make silly errors like dropping negative signs, or moving decimal places (I fully admit to doing both), but that shouldn't be combined with actually trying to teach new material, or you're going to get a situation like the article author described: You don't understand it, and you can't understand why you don't understand.
Whether the author would have been a good engineer is besides the point, if he was a good engineer, he wouldn't have time to write the article. But the main point, that the teaching (whether you view it as the profs or the curriculum's fault) is horribly designed, and that a lot of people who would make good engineers get dissuaded from continuing is a good one, and is more easily examined, I think, in engineering than in most other fields.
I wound up going into grad school for cognitive & neurual science, which is something I never did in college--the reason I could do this is because I was taught about the concepts, or "how to think," and I'm doing fine despite my background. You take an IOE who probably spent more time than I did doing problems sets in any given area of math (especially since they probably started as an EE or ME), and I doubt they could make the switch, because they never really learned that optimizing the layout of a factory floor, solving a 10k+ node electrical circuit, or analyzing traffic problems all are actually based around the same area of graph theory... they just learned about the factory floor.
The other amusing thing is why everybody is saying that engineering teaching isn't highly valued. That's probably true, but the explanation for that 'research' is hilarious--engineering research?-- thats a laugh. Engineering is all about DOING stuff in the real world, my guess is that a very small proportion of really top notch engineers are at universities (as opposed to (research people in) math, english or history teachers, for example). So why the hell is the engin teaching so bad? In fact, is there any other job for a college engin professor other than teaching? If they were so good at research, the private sector is a ton more lucrative.
My advice to undergrad engins or HS'ers: If you're interested in engineering, do your undergrad degree in math, physics, statistics, etc, take a few engin courses on the side (maybe do a minor), and then if you still like engineering, grab a masters in whatever field you want to go into (math majors, for example, have no problem getting into any engineering masters program). You'll leave your options quite a bit more open, not to mention having a less stressful college experience.
You know, although this is whats called an extrinsic event, ie has nothing to do with the election per se, that doesn't mean that these events don't have highly significant correllations in a statistical sense.
There was a conference, back in 2000, at NYU on extrensic uncertainty, which presented, among other things, that sun spot activity and the stock market are correllated.
If you believe in the entire idea of 'statistical signifigance', then its difficult, whether you believe those correllations to be intrinsic or extrinsic, to ignore them.
Back in the 60s-70s, they used to try driving marked police cars 3-abreast (or however many lanes there were) down the highway at 55 to try to enforce the speed limits. The result? Accident rates in the areas that used this type of enforcement went waaay up, because the police were basically creating artifical bubbles of dense traffic that was moving much faster than the same density of traffic would normally move down the same road.
The first thing you really need to make him understand is that the number of people out there who are really interesting and motivated to learn is pretty small, and those who he meets should be valued--but since most people out there aren't terribly interested in being serious intellectuals, he needs to be able to function with them. If he's interested in economics, introduce it as an opportunity cost argument. If not, maybe introduce it to him as a prisoner's dilemma *grin*?
He doesn't need to embrace the culture of his peers, he just needs to find things that interest him that create lanes of communication. I know I wasn't interested in football when I first got to Michigan (I grew up in NYC and attended Stuy), but I quickly realized that it was something that allowed you to chat with a lot of people, and since I've always enjoyed sports, I took the time to learn a lot about it. This doesn't mean that he has to do some rote memorization job so he has something to talk about, it just means he should look into what his peers are interested in, and if any of it interests him, he should take the time to learn about it (his way).
His peers will be interested in chatting with him, and knowing him, if he can discuss the subjects they are interested in talking about it in a novel and interesting way for them. If he's smart, and interested in coming up with novel ways of thinking about things, I doubt he'd have a problem with doing this.
Your student has to understand that we have to function in the world of our peers--whatever that is. Maybe he'll eventually become an academic and be able to lock himself up in an ivory tower, or some cube farm with a whole bunch of other people that are interested in programming. But until he gets to that point, it makes sense to at least try to understand his surroundings rather than trying to make them understand him. Ask him if he thinks its easier to understand him, or its easier for him to understand his class mates?
Unfortunately not very many of us can surround ourselves completely with people who all share our interests, but I've certainly had a lot more fun socially trying to get engage myself in what my friends and aquaintances are interested in, instead of just trying to engage them in my interests.
If he needs a jump start, try getting him to do something, whether its cutting his hair, wearing jeans (or not wearing jeans), etc, that you know he thinks *I* can't do that, but is actually something minor... He's probably backed himself into a niche with his peers that will take him some effort to widen.
Outperform it in terms of Mathematical sophistication?
Whats that supposed to mean? Answer a question Mathematica can't? Nearly every college math student I know can do that.
Or perhaps answer every question that Mathematica CAN answer? I highly doubt that there's a mathematician in the world today that can do that.
Mathematica is a tool, the results you get out are only as useful as your understanding of them.
Oh, and intellectually stunted generation? Intellectually stunted because students of today no longer learn several dozen arbitrary tricks to manually solve differential equations? Pullease. Mathematica (or Matlab, or Maple etc) gives students a tool to investigate problems that were previously inaccessible--that hardly makes them intellectually stunted.