From Wikipedia, "OpenCL (Open Computing Language) is a framework for writing programs that execute across heterogeneous platforms consisting of CPUs, GPUs, and other processors." In other words, write your massively parallel programs using OpenCL and then run them on the device (or combination of devices) that executes your program the fastest.
Hopefully, OpenCL will have the same cataylzing effect on HPC that OpenGL had on computer graphics, but time will tell.
Word of warning to Intel: Almost nobody wants to hand-code assembly to run your SIMD instructions. People doing HPC (at least the ones using CUDA) are scientists and engineers who typically have better things to worry about than reading through detailed tomes on the i7 architecture. Make it more convenient (i.e. via OpenCL) or continue to lose market share in this area.
There's really only one group that has an unquestionable claim to be concerned primarily with education and that's the parents. They're not consulted because they might ask uncomfortable questions like "Will the computer do anything worthwhile?" Neither the activists nor the politicians are interested in having to answer questions like that.
I would also like to add that parenting in general is a major issue in education. If a child`s parents read them at night when they`re young, take an interest in their children`s school work (but don`t hand them the answers), foster a sense of curiosity, and most of all, set a good example, their children are far more likely to succeed.
On the other hand, if their parents don`t give a rat`s ass, set horrible example, spoil their children with material goods, and/or expect nothing, then outcome is usually quite predictable.
Agreed. Not to mention the fact that timed tests are actually a really bad way of measuring one's understanding of
anything but the most trivial facts, and routine procedures, given the time constraints and pressure (creative
minds don't always function at optimum levels under duress). During my undergrad many of the mathematics professors
understood this was the case. The tests were still quite challenging, but the grading was fairly forgiving in the
sense that if nobody got the problem it was struck from the test. It was on the assignments that our professors
really tortured us. Some of the problems took in excess of three hours to solve, some I thought about for days. The
moment of inspiration could come at anytime - for some reason the shower always seemed (and still is) particularly
fruitful...
Deep understanding is a very difficult thing to assess. It is not tied to knowing specific facts or procedures.
Don't get me wrong, you definitely need to know the fundamentals of your field inside and out. It's more like a
kind of "sixth-sense" that one develops after solving enough problems and knowing enough facts that enables an
individual to reduce the search space of the solution down to something more manageable. Even then, the subconscious still has a great deal to do with finding the solution. Its same sort of insight that chess masters have; those who
can simply glance at board already in play and make a cunning move. A great deal of the pursuit of solving non-trivial
problems is also sheer dedication to the task.
In summary, I think the US school system has lost its way. It will not develop deep and curious minds by elevating test scores above all else. That is the easy way out, the real challenge of education is much harder. In the end, it is the both students and thus the entire country that will suffer.
Without a computer you have to learn how to think.
Indeed. It is also likely that a computer provides an additional distraction. Instead of actually doing school work or using one's imagination, the computer provides instant gratification in the form of games, the web, and social networking. Less time is spent using those parts of the brain that govern creativity and problem solving. Those kids' spatial abilities and manual dexterity are probably unreal though.
The concepts of "Turning Complete" and the "Halting Problem" are distinct.
Where did I say otherwise? This was exactly the point I was trying to make. I was not arguing against with the fact that a Turing machine cannot solve the Halting Problem. I was arguing with the GP's implied assertion that the fact that a language cannot describe a solution to the Halting Problem is part of the definition of what it means to be Turing-complete. Yes, it was an extremely pedantic and nitpicky thing to do - I'll admit that - but, I am correct. One must prove that Turing machines cannot solve the Halting Problem in order to state it as fact.
I believe you are just trying to regurgitate concepts that your professors lectured about to appear educated. You fail.
I hold a BS in mathematics (with honors) and a minor in computer science from a good university. I have won numerous awards and national scholarships on the basis of my demonstrated ability. I consider multivariate calculus bedtime reading. If that is failure, what exactly do I have to do in order to succeed?
So you what sort of things do you model in predictive analytics? What kind of methods do you use? How reliable are your predictions?
I majored in mathematics with an emphasis in probability theory, and I have an interest in machine learning, so please don't refrain from any technical details/jargon you might offer in helping me to understand the nature of this field.
Turing-complete means that it is able to perform all of the functions of a universal Turing machine
Correct, that is the definition of what it means to be Turing-complete.
a Turing-complete language (or system) by definition is unable to solve the halting problem expressed within that system.
Wrong, assuming a system can perform all the functions of a Turing machine we can prove that said system is unable to solve the Halting Problem by assuming that it does, and proceeding to obtain a contradiction. You are confusing theorems with definitions - a subtle, but important distinction exists.
Scientists really have to do a better job at communicating clearly with less jargon
Most of us try to do exactly as you suggest when speaking with laymen, but there are two problems with this approach:
1. We are so used to thinking "in jargon" that we use it just as we use everyday language. We unconsciously assume people understand certain words or ideas, because with think about them all the time.
2. Scientists are inventive. We invent new ideas, and when our vocabulary fails to describe those ideas adequately or efficiently, we invent new words. This process has really accelerated over the last few centuries to the point where scientists really speak a very different language than the rest of the population in a professional context. It usually takes between 4-6 years to learn the language, and become creative with it. Thus, we can't always explain certain things in a reasonable amount of time without resorting to either gross simplification or what you call jargon.
Your comment is essentially correct, but school is not industry. The point of education is the pursuit of knowledge and personal excellence. How does copying code from a website in order to satisfy the requirements of an assignment fit in with those objectives?
Most universities (at least in Canada), have a software engineering stream, or a software design stream. There are plenty of opportunities to learn about engineering over the course of an undergraduate career: internships, project courses (many inexperienced students avoid these like the plague), summer research projects, open source projects, or invent your own! Ultimately, it is the student's responsibility to get involved and find opportunities to practice engineering, no matter which programme they are in - you simply can't expect to learn everything in a classroom.
Having worked on projects with both CS and engineering students, I've found a roughly equal percentage are completely clueless. Very few have a firm grasp of how to go about working on a large non-trivial project, because they've never done it themselves. How most of these people managed to make it into upper-year courses is beyond me...
The bigger portion of computer science goes into more advanced topics - networking, software engineering, UI, graphics, graph theory, algorithms, compilers, NLP, etc. - that you wouldn't really need for coding...
You know it really really irks me when someone says something like this, because it simply isn't true - not by a long shot. If all you want to do all day is make high-level API calls, write web applications, and play buzzword bingo, then no you don't need anything but exposure to the relevant language and some frameworks. However, if you want to do anything remotely interesting or non-trivial, then most (if not all) of these topics become important. An excellent example, and by no means the only one, is game programming. If you want to do game programming you must be familiar with ALL of the above topics (with the possible exception of NLP) including graph theory. You don't need to know how to engineer software if you're writing it? Seriously? No wonder there's so much shit out there...
Sorry, but if you want to work in industry or academia in a computer science related field, you must be a strong individual coder. If you are not, well too bad, the position will likely go to someone who is... and if you do happen to get the position and still don't perform, I bet you'll be out the door in O(1) time.
People need downtime. Getting rid of weekends will burn people out. People need time to be creative, to relax, to socialize. If you eliminate those elements of a person's life, you lose an important part of that individual.
That said, I am not against incentives that push people to succeed. I say, make cash prizes available for students that demonstrate potential; that will get people working hard!
You will not be scanning books with an iPhone... at least not with this technique. This system uses structured light to deduce the curvature of the page, which requires a synchronized and calibrated camera/projector system - i.e. very specialized hardware and not something likely to show up on a piece of consumer electronics.
We are also talking about MIT students looking to prove themselves, which makes for a smart and highly motivated group - by no means your average group of code monkeys.
Slashdot Mirror
From Wikipedia, "OpenCL (Open Computing Language) is a framework for writing programs that execute across heterogeneous platforms consisting of CPUs, GPUs, and other processors." In other words, write your massively parallel programs using OpenCL and then run them on the device (or combination of devices) that executes your program the fastest.
Hopefully, OpenCL will have the same cataylzing effect on HPC that OpenGL had on computer graphics, but time will tell.
Word of warning to Intel: Almost nobody wants to hand-code assembly to run your SIMD instructions. People doing HPC (at least the ones using CUDA) are scientists and engineers who typically have better things to worry about than reading through detailed tomes on the i7 architecture. Make it more convenient (i.e. via OpenCL) or continue to lose market share in this area.
Unfortunately the evidence doesn't support your hypothesis: http://www.timesonline.co.uk/tol/news/uk/article571206.ece
How about we try teaching people to be rational whilst supplying them with good reasons to behave - that seems logical to me.
There's really only one group that has an unquestionable claim to be concerned primarily with education and that's the parents. They're not consulted because they might ask uncomfortable questions like "Will the computer do anything worthwhile?" Neither the activists nor the politicians are interested in having to answer questions like that.
I would also like to add that parenting in general is a major issue in education. If a child`s parents read them at night when they`re young, take an interest in their children`s school work (but don`t hand them the answers), foster a sense of curiosity, and most of all, set a good example, their children are far more likely to succeed.
On the other hand, if their parents don`t give a rat`s ass, set horrible example, spoil their children with material goods, and/or expect nothing, then outcome is usually quite predictable.
Agreed. Not to mention the fact that timed tests are actually a really bad way of measuring one's understanding of anything but the most trivial facts, and routine procedures, given the time constraints and pressure (creative minds don't always function at optimum levels under duress). During my undergrad many of the mathematics professors understood this was the case. The tests were still quite challenging, but the grading was fairly forgiving in the sense that if nobody got the problem it was struck from the test. It was on the assignments that our professors really tortured us. Some of the problems took in excess of three hours to solve, some I thought about for days. The moment of inspiration could come at anytime - for some reason the shower always seemed (and still is) particularly fruitful...
Deep understanding is a very difficult thing to assess. It is not tied to knowing specific facts or procedures. Don't get me wrong, you definitely need to know the fundamentals of your field inside and out. It's more like a kind of "sixth-sense" that one develops after solving enough problems and knowing enough facts that enables an individual to reduce the search space of the solution down to something more manageable. Even then, the subconscious still has a great deal to do with finding the solution. Its same sort of insight that chess masters have; those who can simply glance at board already in play and make a cunning move. A great deal of the pursuit of solving non-trivial problems is also sheer dedication to the task.
In summary, I think the US school system has lost its way. It will not develop deep and curious minds by elevating test scores above all else. That is the easy way out, the real challenge of education is much harder. In the end, it is the both students and thus the entire country that will suffer.
Without a computer you have to learn how to think.
Indeed. It is also likely that a computer provides an additional distraction. Instead of actually doing school work or using one's imagination, the computer provides instant gratification in the form of games, the web, and social networking. Less time is spent using those parts of the brain that govern creativity and problem solving. Those kids' spatial abilities and manual dexterity are probably unreal though.
As Einstein famously pointed out, "If we knew what we were doing, it wouldn't be called research, would it?"
Dawkins is a BIOLOGIST
And therefore couldn't possibly be knowledgeable about anything other than biology.
The concepts of "Turning Complete" and the "Halting Problem" are distinct.
Where did I say otherwise? This was exactly the point I was trying to make. I was not arguing against with the fact that a Turing machine cannot solve the Halting Problem. I was arguing with the GP's implied assertion that the fact that a language cannot describe a solution to the Halting Problem is part of the definition of what it means to be Turing-complete. Yes, it was an extremely pedantic and nitpicky thing to do - I'll admit that - but, I am correct. One must prove that Turing machines cannot solve the Halting Problem in order to state it as fact.
I believe you are just trying to regurgitate concepts that your professors lectured about to appear educated. You fail.
I hold a BS in mathematics (with honors) and a minor in computer science from a good university. I have won numerous awards and national scholarships on the basis of my demonstrated ability. I consider multivariate calculus bedtime reading. If that is failure, what exactly do I have to do in order to succeed?
So you what sort of things do you model in predictive analytics? What kind of methods do you use? How reliable are your predictions? I majored in mathematics with an emphasis in probability theory, and I have an interest in machine learning, so please don't refrain from any technical details/jargon you might offer in helping me to understand the nature of this field.
Turing-complete means that it is able to perform all of the functions of a universal Turing machine
Correct, that is the definition of what it means to be Turing-complete.
a Turing-complete language (or system) by definition is unable to solve the halting problem expressed within that system.
Wrong, assuming a system can perform all the functions of a Turing machine we can prove that said system is unable to solve the Halting Problem by assuming that it does, and proceeding to obtain a contradiction. You are confusing theorems with definitions - a subtle, but important distinction exists.
Scientists really have to do a better job at communicating clearly with less jargon
Most of us try to do exactly as you suggest when speaking with laymen, but there are two problems with this approach: 1. We are so used to thinking "in jargon" that we use it just as we use everyday language. We unconsciously assume people understand certain words or ideas, because with think about them all the time. 2. Scientists are inventive. We invent new ideas, and when our vocabulary fails to describe those ideas adequately or efficiently, we invent new words. This process has really accelerated over the last few centuries to the point where scientists really speak a very different language than the rest of the population in a professional context. It usually takes between 4-6 years to learn the language, and become creative with it. Thus, we can't always explain certain things in a reasonable amount of time without resorting to either gross simplification or what you call jargon.
Your comment is essentially correct, but school is not industry. The point of education is the pursuit of knowledge and personal excellence. How does copying code from a website in order to satisfy the requirements of an assignment fit in with those objectives?
Most universities (at least in Canada), have a software engineering stream, or a software design stream. There are plenty of opportunities to learn about engineering over the course of an undergraduate career: internships, project courses (many inexperienced students avoid these like the plague), summer research projects, open source projects, or invent your own! Ultimately, it is the student's responsibility to get involved and find opportunities to practice engineering, no matter which programme they are in - you simply can't expect to learn everything in a classroom. Having worked on projects with both CS and engineering students, I've found a roughly equal percentage are completely clueless. Very few have a firm grasp of how to go about working on a large non-trivial project, because they've never done it themselves. How most of these people managed to make it into upper-year courses is beyond me...
The bigger portion of computer science goes into more advanced topics - networking, software engineering, UI, graphics, graph theory, algorithms, compilers, NLP, etc. - that you wouldn't really need for coding...
You know it really really irks me when someone says something like this, because it simply isn't true - not by a long shot. If all you want to do all day is make high-level API calls, write web applications, and play buzzword bingo, then no you don't need anything but exposure to the relevant language and some frameworks. However, if you want to do anything remotely interesting or non-trivial, then most (if not all) of these topics become important. An excellent example, and by no means the only one, is game programming. If you want to do game programming you must be familiar with ALL of the above topics (with the possible exception of NLP) including graph theory. You don't need to know how to engineer software if you're writing it? Seriously? No wonder there's so much shit out there...
Sorry, but if you want to work in industry or academia in a computer science related field, you must be a strong individual coder. If you are not, well too bad, the position will likely go to someone who is... and if you do happen to get the position and still don't perform, I bet you'll be out the door in O(1) time.
Corporations are persons under the law, so it applies to them as well.
Sorry, a lot of us end up going to graduate school. See you in two years =)
No, it should be called an autobot!
People need downtime. Getting rid of weekends will burn people out. People need time to be creative, to relax, to socialize. If you eliminate those elements of a person's life, you lose an important part of that individual. That said, I am not against incentives that push people to succeed. I say, make cash prizes available for students that demonstrate potential; that will get people working hard!
Indeed! Wikileaks should not tell me what to think! I'm a grown up, I can make up my own mind tyvm!
Thank you! It's about time someone said that! On the other hand, firing on the vehicle coming to rescue the wounded man is definitely questionable.
I mean, imagine if you worked on a popular OS and my boss told me to put a back-door in, saying the NSA required it of us. what would you do? :)
I would contact the RCMP and CSIS immediately!
Indeed... Did you see the lens on the prototype? Good luck fitting that on a smartphone =P
You will not be scanning books with an iPhone... at least not with this technique. This system uses structured light to deduce the curvature of the page, which requires a synchronized and calibrated camera/projector system - i.e. very specialized hardware and not something likely to show up on a piece of consumer electronics.
We are also talking about MIT students looking to prove themselves, which makes for a smart and highly motivated group - by no means your average group of code monkeys.