If you've got so much data and still can't find differences, it's because your measurements are not informative enough. The only thing the author can conclude is that his way of measuring code quality cannot pick up differences in the part he has labeled "process".
There are some applications, mainly for disabled people, where people can control a cursor with a BCI (after training). It's slow and not very accurate.
Are you telling me that you can't do enterprisey architectures in ADA? But where would we get our daily dose of humor (http://thedailywtf.com/ from then? This has to be stopped!
I remember my professor returning in disgust after teaching 3rd year students. It turned out they didn't know what a standard deviation was. Average, ok, they knew that, but anything beyond that was a mystery to them, after two years of statistics classes. And these were students specializing in experimental psychology.
And don't get me started on a colleague who did a presentation on his research and was proud that he could tell that he finally had a obtained a significant result.
Well, I can't see surprise (or surprisal as it is also known) as a major breakthrough, at least not in cognitive psychology. All it does it tell you how far something (e.g. a word) deviates from some (corpus based) expectation. It does provide another way of looking at things like reading times and difficulty of sentences, but it is just another measure of something like relevance, probability (actually, surprisal is defined as the -log of the conditional probability of an event, just like in information theory).
The traffic example in the link is a cute application, but not really what something that will have a dramatic impact on computing. It's probably in the list because they got one example from 10 different disciplines, so nobody would feel left behind...
That's bollocks. You can't move brain tissue around. I'm a cognitive neuroscientist, not a neurosurgeon, so I can't think of all obstacles, but attaching blood vessels and synapses is not feasible. If you move the speech area (which one is that exactly? where did they move it to? was there a hole or something?) you're cutting billions of neurons with hundreds of billions of connections, all in dire need of oxygen. You'll need to reattach everything in five minutes or so to (already open?) blood vessels, otherwise the neurons die. I'm afraid it doesn't make any sense. You should give a citation or provide more details.
1. The method needs training per person, so you can't set up a pattern that will fit everyone, but you can get quite a few patterns after training. 2. The method is only observing the neuronal activity in your (first) visual processing areas (V1, V2 and V3 to be precise), so any association with a dog that bit you is not seen. 3. The activity in V1 is supposed to be a decent copy of the image projected onto the retina, although it is split up in different components. So retrieving the image from V1 is possible, but requires a resolution that is beyond fMRI. 4. The method is far from perfect. It can distinguish quite a few pictures, but (quote from the article):
Inspection revealed that identification errors tended to occur when the selected image was visually similar to the correct image. This suggests that noise in measured voxel activity patterns causes the identification algorithm to confuse images that have similar features. And these were perfectly controlled images, so no, like this nobody can distinguish which of two similar dogs you're looking at.
Not completely true. I once got modded up and down over and over again (really from +1 to +5 to -1 to +4 to 0 to... etc.) for critizing Linux' inability to appeal to Joe Average (installation, UI, software with all the features you don't want but none of the features you happen to know, that kind of thing). The result was that I couldn't post to Slashdot for a couple of weeks!
Can you image? I couldn't... Well, wait. Perhaps it wasn't that terrible. Anyway, not only critizing Apple will get you modded into oblivion, but critizing Linux too and the M$-fanbois will mod you back into conciousness, after which some automated system just blocks you.
I (personally, etc.) do not like Python. I'm from the school of language design that thinks it's not good that white space is significant. It reminds of FORTRAN and punch cards. However, I have my admiration for obscure languages as well (Caml is great according to me, but not really everyone's cup of tea). And there are good programmers that work almost exclusively in one language, some of them might even work in Java...
And don't get me started on Perl: while ( && ) anyone? That aside, Perl's superficial ugliness is no problem to you, yet exactly opposite to the reason your friend chooses Python. Still you laud his choice. Weird. Conclusion 2:
And strangely enough, your conclusion looks valid to me: such things probably set apart the real programmers from the merely "not more than 1 bug per 100 lines" types.
I'm not particularly fond of the spelling/grammar wars, but... Although your spelling is fine, punctuation and grammar are lacking. Here is a rewrite:
- As one of the elite few programmers who does... "Elite few programmers" is ungrammatical and superfluous (an elite is a small group by definition); "Who does" refers to a single person, but the phrase and context make clear that you are speaking about programmers in general, so "As one of the few programmers who do know how to spell" is better. - Add "I" before "thank you", or write "A thank you... from one of the few". - Add a comma before "I suppose". - The hyphen after "However" should also be a comma. - you should put hyphens or commas around "I think". - Remove the comma before "and" or add one before "but". - "- at a time..." is written in a way that is normally used to add a contrast ("At a time when foreign tourism to the U.S. is heavily down, the State Department has just raised the incoming visa fee to $131"). However, it just supports the context here, so "but this was at a time" would have been better.
And second, the signal path length for adding memory capacity to a device (any device) doesn't increase by n, it increases by log n. That's what I wrote: if you go from 2^n addresses to (2^n)^2, execution time doubles: log(2^n^2) / log(2^n) = 2.
If quanta are the building blocks of our universe and they display nondeterministic properties even though everything larger appears thoroughly deterministic... Does that imply anything about the hypothetical ur-universe? It implies our current universe is not computable/predictable by our mathematics. It would also seem to imply that if you build a universe with a TM, you will lose something. However, that is speculation, since the real nature of quantum indeterminacy is unknown. If it is pure randomness, you would not necessarily have to lose it in the simulation.
In your studied opinion, that new vantage would definitely not offer another step towards either finding a polynomial-time algorithm that works in our universe, nor offer a step towards a proof that no such algorithm exists?
In my studied opinion, it it would do precisely that, namely prove the relation between P and NP in our universe, and therefore not be a proof of the reality or simulated nature of our universe.
The TCAM example is faulted: it is also O(log n), as is shown in another proof. If you're going to extend it from an n byte address (i.e. 2^n addresses) to a 2n byte address (2^n^2 addresses), the signal path increases by a factor 2 and thus execution time. The fact that it can complete the operation in a single cycle does not make it O(1), since it is not a Turing Machine.
Now, you have to admit that if we could design an alternate universe by changing several basic properties of our own in which we could prove that a whole slew of NP math problems have easy solutions, that would be a remarkably interesting result.
That could be interesting: there are already ways known to do it in P, namely one in which a TM can duplicate itself for each choice and both new instances proceed along a different path. However, that's the very definition of NP, so not all universes in which an NP-hard problem is solved in P-time will shed a new light on the relation between the two classes...
find a mathematical problem which is conceptually simple (e.g. factoring the product of large primes) but which we know to be hard ( O(x^n) ) and then construct a simulation of a finite ur-universe in which the problem is easy.
Wow, hold it there cowboy! You're making two reasoning errors here.
1. Problems we know to be hard are problems in NP. However, we still don't know if NP problems can be solved in polynomial time or not. So if you could build such a universe, you would have a "P=NP proof", not a proof of the reality of our universe. 2. Of course you can circumvent this by looking at another property, X, for which there is a proof. However, your simulation is going to run on a Turing machine, which follows the laws of mathematics. So property X will hold in your simulated universe as well, unless you manage to uncover a way of computing the uncomputable...
Well, I'm a mac user and I also spoof the agent string (in Camino with the still working CaminoTools), for the same reason. Now not many people will do that, but it adds uncertainty.
On an related note: from what I've seen, Linux desktop market share is pretty low. It's reserved for computer geeks and a few anti-globalists that hate Microsoft because it's a big corporation and has bad aesthetics...
Holy Guacamoly! Since when is Phys.Rev.A open to vandalism? On the other hand, quite a few of Wikipedia's more technical entries seem to have been edited rather badly, and I've seen someone used one of them for the introduction of an article, without checking the f*kcing original source. And no, I'm not going to correct Wikipedia, since that will only give a false sense of correctness. Students and researchers should consult authoritative resources, not one that's just good enough for a quick curiosity search.
The real shame of it is that the Dillo project is on hold now, even though with the tiniest fraction of the resources of the Mozilla project, it could very quickly become an absolutely amazing web browser.
Just pray that they're not going to throw Mozilla's resources to this project, because then Mozilla will perform just as well as Dillo...
That was ironic, but I do think that having many people work on a project can lead to "certain" design decisions under pressure (e.g. in an attempt to make it safe to run code that's not been thoroughly checked or to make the interface easier for new-comers) that in the beginning have a modest performance penalty, but in the end bog everything down and can almost not be removed without the whole system collapsing. Anyway, it has Mythical Man Month written all over it, I guess.
For someone who criticizes spelling, the first word in your post contains a remarkable oversight. I guess it's the spelling nazi's doom.
Anyway, plasticity is there, but there is a "vaster" body of evidence that the older you get, the harder it is to learn. Of course, you still can remember things, but e.g. these memories have to compete with existing ones. Consequently, perhaps our brain's plasticity is not enough to accomodate a huge amount of memories. Then again, plasticity does not seem to apply equally in all regions or functions. Language (my topic) learning gets significantly (not in the p whatever sense) more difficult with age.
About the neurological effect: I think it has been shown that plasticity in the retina is very limited in time. If I remember properly, the first connections form very quickly and do not change any more. Therefore, different brain regions might have different plasticity patterns with different implications.
"Our" culture might not treat the elderly properly, but then you're speaking about people over 70. The effects I'm speaking about manifest themselves from the early 20s onwards (and for language, there is quite a large group of researchers that claim the breaking point is around 6 to 8).
As a case in point, take classical musicians. Quite a few of the top performers started playing very, very young. There is not one I know of that started playing at the age of 25. Or take the so-called wolf children. They can't seem to acquire fluent speech when found too late. Blind or deaf people that gain vision/hearing at later age also perform rather below normal levels.
So perhaps the medical doctors are presumptuous with their assumptions on plasticity, but that doesn't show that our brains can keep up with age.
That kind of conclusion is totally unwarranted. To begin with, the mice were not 70 years old. No, don't laugh! Either mouse neurons age as fast as the mice themselves do, which implies that (the processes in) their neurons differ fundamentally from ours, or these neurons age the way we do, but then they were studying two year old neurons, which I thought used to be considered pretty young.
Second, the observation that learning and memorizing becomes more difficult with age is pretty solid. If our neurons maintain their plasticity, these people should explain how a plastic brain stops learning.
Concluding: the observations are probably true, the conclusions were just made to draw attention and get more funding (aging is a big topic for funds these days). Such is the sad state of science.
Go and tell that. "Hey, it's your own fault, you know! You should have driven another car."
You sir, are an idiot and I feel sorry for the thousands of people you claim to have acquainted. Your arguments consist of anecdotes and your argument goes along the lines of "well, I shot him, but if he would have worn a bullet-proof vest, he wouldn't have been dead." It's no excuse and you know it.
6 mph, that's nearly 10km/h. That's not little. And still a 17-20% increase, accounting for 37-44% more kinetic energy. And it does matter, since that happens to have a relation to the length of the braking path (in case I chose the wrong word: the length of the path the vehicle traverses from the moment of braking until complete stop). And don't forget that at a higher speed, you cross a larger distance during the reaction time. So the car definitely hit with more energy than it would have had, had the drivers obeyed the rules.
I don't know about the qualification "speed demon", but irresponsible comes to mind, as does senseless death.
At least in the country where I live, running a stop sign does not make you free game. Drivers should still pay attention of people crossing the street at any moment and speeding is not going to help. Anyway, 6 miles is nearly 10%, and I wonder at which point that was measured. The heart attack was clearly induced by accident and the wife's death can also be attributed.
We're gonna hang 'm all, we're gonna hang 'm all. We don't know who did it but we're gonna hang 'm all.
If the earth were managed properly,... That's like saying "if pigs could fly": you know it's not going to happen.
Look at our brains. They were designed to learn forever. Ok, I'm a post-doc in neuro-cognition and I can say that although our brains might learn as long as they live, that does go at the cost of something else. Learn A, forget B (although that's a bit simplistic). And the speed with which we learn decreases over time. And our cells stop renewing, and the myellin insulation of our synapses deteriorates over time, and... So learning forever is a bit of a misnomer here; changing as long as they live is more appropriate, and there is a clear upper limit to our intellectual capacity: at this moment, you can't beat a computer any longer in draughts, and soon no-one will be able to win a chess match against a machine. There are systems we cannot understand: look at how long we are struggling with the definition of "life"!
Slashdot Mirror
If you've got so much data and still can't find differences, it's because your measurements are not informative enough. The only thing the author can conclude is that his way of measuring code quality cannot pick up differences in the part he has labeled "process".
There are some applications, mainly for disabled people, where people can control a cursor with a BCI (after training). It's slow and not very accurate.
Are you telling me that you can't do enterprisey architectures in ADA? But where would we get our daily dose of humor (http://thedailywtf.com/ from then? This has to be stopped!
I remember my professor returning in disgust after teaching 3rd year students. It turned out they didn't know what a standard deviation was. Average, ok, they knew that, but anything beyond that was a mystery to them, after two years of statistics classes. And these were students specializing in experimental psychology.
And don't get me started on a colleague who did a presentation on his research and was proud that he could tell that he finally had a obtained a significant result.
After 19 failures...
Well, I can't see surprise (or surprisal as it is also known) as a major breakthrough, at least not in cognitive psychology. All it does it tell you how far something (e.g. a word) deviates from some (corpus based) expectation. It does provide another way of looking at things like reading times and difficulty of sentences, but it is just another measure of something like relevance, probability (actually, surprisal is defined as the -log of the conditional probability of an event, just like in information theory).
The traffic example in the link is a cute application, but not really what something that will have a dramatic impact on computing. It's probably in the list because they got one example from 10 different disciplines, so nobody would feel left behind...
That's bollocks. You can't move brain tissue around. I'm a cognitive neuroscientist, not a neurosurgeon, so I can't think of all obstacles, but attaching blood vessels and synapses is not feasible. If you move the speech area (which one is that exactly? where did they move it to? was there a hole or something?) you're cutting billions of neurons with hundreds of billions of connections, all in dire need of oxygen. You'll need to reattach everything in five minutes or so to (already open?) blood vessels, otherwise the neurons die. I'm afraid it doesn't make any sense. You should give a citation or provide more details.
2. The method is only observing the neuronal activity in your (first) visual processing areas (V1, V2 and V3 to be precise), so any association with a dog that bit you is not seen.
3. The activity in V1 is supposed to be a decent copy of the image projected onto the retina, although it is split up in different components. So retrieving the image from V1 is possible, but requires a resolution that is beyond fMRI.
4. The method is far from perfect. It can distinguish quite a few pictures, but (quote from the article): Inspection revealed that identification errors tended to occur when the selected image was visually similar to the correct image. This suggests that noise in measured voxel activity patterns causes the identification algorithm to confuse images that have similar features. And these were perfectly controlled images, so no, like this nobody can distinguish which of two similar dogs you're looking at.
Try movielens (http://www.movielens.org/): it does offer a search with specific languages.
You might also try to copy my recommendations, since it recommends quite a few Japanese movies based on that...
Not completely true. I once got modded up and down over and over again (really from +1 to +5 to -1 to +4 to 0 to ... etc.) for critizing Linux' inability to appeal to Joe Average (installation, UI, software with all the features you don't want but none of the features you happen to know, that kind of thing). The result was that I couldn't post to Slashdot for a couple of weeks!
... Well, wait. Perhaps it wasn't that terrible. Anyway, not only critizing Apple will get you modded into oblivion, but critizing Linux too and the M$-fanbois will mod you back into conciousness, after which some automated system just blocks you.
Can you image? I couldn't
I should have gotten the thanks-for-pointing-out-the-GP-is-not-a-good-summary-in-a=nerdy=way karma, I guess.
But you agree I should ask for a modest part, let's say 10%, right?
ALTER TABLE fingerprints ADD BLOB;
Wrong reasoning, right conclusion?
I (personally, etc.) do not like Python. I'm from the school of language design that thinks it's not good that white space is significant. It reminds of FORTRAN and punch cards. However, I have my admiration for obscure languages as well (Caml is great according to me, but not really everyone's cup of tea). And there are good programmers that work almost exclusively in one language, some of them might even work in Java...
And don't get me started on Perl: while ( && ) anyone? That aside, Perl's superficial ugliness is no problem to you, yet exactly opposite to the reason your friend chooses Python. Still you laud his choice. Weird. Conclusion 2:
And strangely enough, your conclusion looks valid to me: such things probably set apart the real programmers from the merely "not more than 1 bug per 100 lines" types.
I think it is now time we all read "Real programmer" again: http://www.pbm.com/~lindahl/real.programmers.html.
I'm not particularly fond of the spelling/grammar wars, but ... Although your spelling is fine, punctuation and grammar are lacking. Here is a rewrite:
... ... from one of the few". ..." is written in a way that is normally used to add a contrast ("At a time when foreign tourism to the U.S. is heavily down, the State Department has just raised the incoming visa fee to $131"). However, it just supports the context here, so "but this was at a time" would have been better.
- As one of the elite few programmers who does
"Elite few programmers" is ungrammatical and superfluous (an elite is a small group by definition); "Who does" refers to a single person, but the phrase and context make clear that you are speaking about programmers in general, so "As one of the few programmers who do know how to spell" is better.
- Add "I" before "thank you", or write "A thank you
- Add a comma before "I suppose".
- The hyphen after "However" should also be a comma.
- you should put hyphens or commas around "I think".
- Remove the comma before "and" or add one before "but".
- "- at a time
If quanta are the building blocks of our universe and they display nondeterministic properties even though everything larger appears thoroughly deterministic... Does that imply anything about the hypothetical ur-universe? It implies our current universe is not computable/predictable by our mathematics. It would also seem to imply that if you build a universe with a TM, you will lose something. However, that is speculation, since the real nature of quantum indeterminacy is unknown. If it is pure randomness, you would not necessarily have to lose it in the simulation.
In my studied opinion, it it would do precisely that, namely prove the relation between P and NP in our universe, and therefore not be a proof of the reality or simulated nature of our universe.
Now, you have to admit that if we could design an alternate universe by changing several basic properties of our own in which we could prove that a whole slew of NP math problems have easy solutions, that would be a remarkably interesting result.The TCAM example is faulted: it is also O(log n), as is shown in another proof. If you're going to extend it from an n byte address (i.e. 2^n addresses) to a 2n byte address (2^n^2 addresses), the signal path increases by a factor 2 and thus execution time. The fact that it can complete the operation in a single cycle does not make it O(1), since it is not a Turing Machine.
That could be interesting: there are already ways known to do it in P, namely one in which a TM can duplicate itself for each choice and both new instances proceed along a different path. However, that's the very definition of NP, so not all universes in which an NP-hard problem is solved in P-time will shed a new light on the relation between the two classes...
Wow, hold it there cowboy! You're making two reasoning errors here.
1. Problems we know to be hard are problems in NP. However, we still don't know if NP problems can be solved in polynomial time or not. So if you could build such a universe, you would have a "P=NP proof", not a proof of the reality of our universe.
2. Of course you can circumvent this by looking at another property, X, for which there is a proof. However, your simulation is going to run on a Turing machine, which follows the laws of mathematics. So property X will hold in your simulated universe as well, unless you manage to uncover a way of computing the uncomputable...
Conclusion: the whole idea is logically invalid.
Well, I'm a mac user and I also spoof the agent string (in Camino with the still working CaminoTools), for the same reason. Now not many people will do that, but it adds uncertainty.
On an related note: from what I've seen, Linux desktop market share is pretty low. It's reserved for computer geeks and a few anti-globalists that hate Microsoft because it's a big corporation and has bad aesthetics...
Holy Guacamoly! Since when is Phys.Rev.A open to vandalism? On the other hand, quite a few of Wikipedia's more technical entries seem to have been edited rather badly, and I've seen someone used one of them for the introduction of an article, without checking the f*kcing original source. And no, I'm not going to correct Wikipedia, since that will only give a false sense of correctness. Students and researchers should consult authoritative resources, not one that's just good enough for a quick curiosity search.
Just pray that they're not going to throw Mozilla's resources to this project, because then Mozilla will perform just as well as Dillo...
That was ironic, but I do think that having many people work on a project can lead to "certain" design decisions under pressure (e.g. in an attempt to make it safe to run code that's not been thoroughly checked or to make the interface easier for new-comers) that in the beginning have a modest performance penalty, but in the end bog everything down and can almost not be removed without the whole system collapsing. Anyway, it has Mythical Man Month written all over it, I guess.
For someone who criticizes spelling, the first word in your post contains a remarkable oversight. I guess it's the spelling nazi's doom.
Anyway, plasticity is there, but there is a "vaster" body of evidence that the older you get, the harder it is to learn. Of course, you still can remember things, but e.g. these memories have to compete with existing ones. Consequently, perhaps our brain's plasticity is not enough to accomodate a huge amount of memories. Then again, plasticity does not seem to apply equally in all regions or functions. Language (my topic) learning gets significantly (not in the p whatever sense) more difficult with age.
About the neurological effect: I think it has been shown that plasticity in the retina is very limited in time. If I remember properly, the first connections form very quickly and do not change any more. Therefore, different brain regions might have different plasticity patterns with different implications.
"Our" culture might not treat the elderly properly, but then you're speaking about people over 70. The effects I'm speaking about manifest themselves from the early 20s onwards (and for language, there is quite a large group of researchers that claim the breaking point is around 6 to 8).
As a case in point, take classical musicians. Quite a few of the top performers started playing very, very young. There is not one I know of that started playing at the age of 25. Or take the so-called wolf children. They can't seem to acquire fluent speech when found too late. Blind or deaf people that gain vision/hearing at later age also perform rather below normal levels.
So perhaps the medical doctors are presumptuous with their assumptions on plasticity, but that doesn't show that our brains can keep up with age.
That kind of conclusion is totally unwarranted. To begin with, the mice were not 70 years old. No, don't laugh! Either mouse neurons age as fast as the mice themselves do, which implies that (the processes in) their neurons differ fundamentally from ours, or these neurons age the way we do, but then they were studying two year old neurons, which I thought used to be considered pretty young.
Second, the observation that learning and memorizing becomes more difficult with age is pretty solid. If our neurons maintain their plasticity, these people should explain how a plastic brain stops learning.
Concluding: the observations are probably true, the conclusions were just made to draw attention and get more funding (aging is a big topic for funds these days). Such is the sad state of science.
PS I hold a post-doc in neurocognition.
Go and tell that. "Hey, it's your own fault, you know! You should have driven another car."
You sir, are an idiot and I feel sorry for the thousands of people you claim to have acquainted. Your arguments consist of anecdotes and your argument goes along the lines of "well, I shot him, but if he would have worn a bullet-proof vest, he wouldn't have been dead." It's no excuse and you know it.
6 mph, that's nearly 10km/h. That's not little. And still a 17-20% increase, accounting for 37-44% more kinetic energy. And it does matter, since that happens to have a relation to the length of the braking path (in case I chose the wrong word: the length of the path the vehicle traverses from the moment of braking until complete stop). And don't forget that at a higher speed, you cross a larger distance during the reaction time. So the car definitely hit with more energy than it would have had, had the drivers obeyed the rules.
I don't know about the qualification "speed demon", but irresponsible comes to mind, as does senseless death.
At least in the country where I live, running a stop sign does not make you free game. Drivers should still pay attention of people crossing the street at any moment and speeding is not going to help. Anyway, 6 miles is nearly 10%, and I wonder at which point that was measured. The heart attack was clearly induced by accident and the wife's death can also be attributed.
We're gonna hang 'm all, we're gonna hang 'm all. We don't know who did it but we're gonna hang 'm all.
BTW, I don't think our brains were designed.