It's pretty much necessary, though, given limited time. Generally, you want to be teaching concepts, not fiddling with software details, and that's easiest if you just pick one piece of software for the purposes of teaching, and assume people can learn the details of other software on their own. So, for example, if you're teaching C++, you might want to be able to just assume everyone has access to g++ and GNU make, preferably all in the same version, instead of also dealing with XCode and Visual Studio and gcc under Cygwin and god knows what else. The easiest way to do that is just to have a Unix computer lab students have access to.
I agree; student laptops are useful for generic computer usage, but not that great for assuming a particular set of software, unless you're going to go the extra step and mandate that students buy a particular computer with a particular OS and software environment. If you aren't going to do that, you're stuck with some of your students running Windows, some OS X, a handful Linux, and very little you can assume about what they can install and run.
If you have a computer lab with some known software install, you know that if you want to use some Mac-only app in the curriculum, for example, you can send them over to the Mac lab to use it (likewise for Unix- or Windows-only stuff).
If the only claim is that science converges in the limit to truth in the mathematical sense, then the statement is about as good as not saying anything at all about science, since it literally says nothing at all about science's accuracy within the next 5 million years. I assumed the great-grandparent poster was making some sort of claim about the convergence of science towards truth on human timescales.
I assumed he meant something stronger about things on human time scales, roughly that science accumulates truth with relatively minor setbacks. If you take asymptotic convergence literally, then yes you're correct, but it's also a useless statement, because it's possible for science to asymptotically converge to the truth without *any* science in the present day or for the next 5 million years to be true.
Presumably he didn't mean a statement as weak as, "science will one day converge to the truth, even though it might be the case that all of science that we've ever done is wrong, and all science for the next several million years might also be wrong".
What sort of timescale? To quote a bit of widely held scientific method, the problem is that that claim is unfalsifiable: any example of a long-held theory that turned out to be wrong, no matter how long it was held, can be rationalized as "well, over a longer timescale, science converges to the truth". The only way to disprove that would be to wait an infinitely long time for science to fail to converge to the truth.
Asimov may have had a PhD in science, but he did nothing of note in science, spending virtually his entire adult life writing science fiction and popular-science works, and virtually none of it writing peer-reviewed journal articles actually on scientific topics. His writings, unsurprisingly, therefore tend to take the mythologized view of science common in sci-fi and pop-science.
Most scientists don't understand science, outside their tiny provincial field; I'm a scientist so I see this all the time. Most have very fairy-tale notions of the scientific method and knowledge production in particular.
You might want to read up on some of the people (scientists especially) who have taken the time to understand how science works, and written on the philosophy and sociology of science.
In particular, it is certainly not true that science converges asymptotically to the truth. It oven diverges substantially, sometimes for hundreds of years, before entire fields (like "racial hygiene") are thrown out as failed experiments. We're currently in the middle of a debate over whether string theory should be placed in that dustbin or not, for example.
Very few scientists in other fields consider economics a legitimate science. In its partial defense, it's studying very complex phenomena--- considerably more complex than understanding the weather, for example, which is itself no cakewalk. On the other hand, economics doesn't seem to really understand that it's dealing with complex dynamical systems, and has been extremely slow to import the tools now standard in all other areas that deal with complex dynamical systems (including weather). Instead they seem to rely mainly on equilibrium assumptions that are unlikely to ever be even approximately true.
The reason you're finding curricula that don't include things you consider necessary for CS is that the number of things that a significant proportion of people consider necessary for CS has ballooned to the point where it doesn't fit in 4 years. Since there's strong resistance to turning CS into a 5-year degree (my alma mater actually suggested it and got shot down quickly), something has to go.
Different schools make different choices about what, treat the rest as block boxes, and people agree with those choices to varying extents. You seem to be mainly focused on what's sometimes called "old-school core CS", the low-level bit-banging and compilers/OS level on top of that. Many schools do focus on that, but the tradeoff is that they treat other things as black boxes; e.g. instead of knowing how to design an algorithm with good asymptotic running time, you look up an algorithm in a book, and focus on how to implement it efficiently on your current hardware. There are lots of other elements; a big one is user interfaces, which are a core part of building apps that anyone can actually use (some people argue that the entire field of CS is about communication, between humans and machines), but treated poorly in low-level-focused CS programs.
I agree keyboards with small key travel are worse, but there is a considerable range of key travel you can get in non-buckling-spring keyboards, too. I'm perfectly happy with my Microsoft Ergonomic 4000 keyboard, and find the larger travel does make it a lot better than both my laptop's very shallow keyboard, and an old Dell desktop keyboard I used to use. (All 3 are rubber-dome.)
For example, maybe you prefer black paint, but are willing to offset the effect by driving a smaller, and more fuel-efficient car. Why shouldn't driving a black 90-hp econobox be an option, when that's actually considerably more environmentally friendly than a white Hummer? Why is California banning the first but not the second?
Raising their gas tax significantly is the most straightforward way to tax inefficiency in cars, without micromanaging where exactly the inefficiency comes from. Use more gas, pay more tax.
Maybe our legislators who are always so worried about sexual exploitation of children as an excuse to censor the internet and everything else, might want to look into whether prohibiting the government from forcibly stripping children naked shouldn't be a higher priority.
At least some of the founders believed that corporations could, if left unchecked, become quasi-governments and similar threats to individual liberty. Thus Thomas Jefferson, for example, expressed the opinion that powerful multi-national corporations should be brought in check. That doesn't seem to be part of the modern libertarian agenda.
Insofar as the language proper is defined by the language standards, the I/O libraries are part of C, because they're specified in the ANSI C and C99 reports. Any conforming C implementation must have the standard I/O functions, and they must behave in the way the standard specifies. That differs quite a bit from the situation with networking libraries, which are third-party and not covered by the C standard.
The problem is still the efficiency, though. There are lots of ways to mark units of computation as "this could be done separately, but depends on Y"--- OpenMP provides a bunch of them, for example, and there's been proposals dating back to the 80s, probably earlier. The problem is figuring out how to implement that efficiently, though, so that the synchronization overhead doesn't dominate the parallelization gains. Does the system spawn new threads? Maintain a pool of worker threads and feed thunks to them? Some hybrid approach? How does it determine when it's worth the effort of doing anything for a particular bit of computation versus just doing it inline and saving the overhead? Etc.
Basically Grand Central is yet another in the decades-long line of proposals for specifying parallelizable computations. What's still an open question is whether they've solved the harder part, a way to, as you say, "[route] computation packets to wherever they can go, and then [receive] the results", without that routing and receiving taking inordinate overhead.
Sure, but then you lost a lot of it again: from 1937 to 1942, the market dropped 50%, so your $450,000 was back to $225,000. If you can perfectly market time, then sure, you can always make money (even in down markets--- just time your short-selling well). Most investors likely bought either too early or two late (most of the gains were very soon after the bottom), or held on too late past 1937.
And without a resolution to the current political crisis in Pakistan, Pakistanis won't be able to get their tech industry going again. Nearly everything is related if you really stretch.
I'm not saying these things aren't worth discussing, just that it isn't really Slashdot's core competence. I participate at multiple discussion forums for precisely the reason that I do find many things valuable to discuss. But I don't really debate AIG bailouts at Slashdot, and I don't really debate the merits of OpenOffice at DailyKos.
Majority owners can't actually do "whatever they want" unless they actually own 100% of the company. The trouble here is that the 20% minority owners are something of an annoying fiction--- the company is by rights bankrupt and the shares worthless, but the government has unwound this mess somewhat ineptly and as a result on paper the company still exists and has value, and 20% of that value is owned in the private sector by people who technically the company has a fiduciary duty to.
Perhaps we should discuss rising income inequality in the US? Or the wars in Darfur? Lots of stuff is worth discussing, but I think having some focus on forums tends to make them a bit better.
So I've been around long enough (insert UID bragging here) to know that Slashdot has a fairly wide definition of "tech news", which is fine. Tech influences lots of things, and vice versa. Plenty of political news is relevant to a tech-news site, from DMCA-related things, to voting machines, to internet-speech restrictions (or just speech restrictions that touch online speech). Discussing that here seems fine.
But really, how does this fit in even if you squint? A tax dispute between AIG and the US government, tangled up with a financial bailout package. A mess, sure, but not an even-slightly-tech-related one.
Sorry, but that's a cop-out. You have a Labor-lead government who wants to remove porn and anything they disagree with from the internet, and is using the Christian loony as cover.
Are those two votes really worth having their party tarred as the party of censorship? It seems like it'll just play into the hands of right-wing critics who consider leftist parties to be the parties of authoritarianism and the nanny state---because in this case they really are.
Slashdot Mirror
It's pretty much necessary, though, given limited time. Generally, you want to be teaching concepts, not fiddling with software details, and that's easiest if you just pick one piece of software for the purposes of teaching, and assume people can learn the details of other software on their own. So, for example, if you're teaching C++, you might want to be able to just assume everyone has access to g++ and GNU make, preferably all in the same version, instead of also dealing with XCode and Visual Studio and gcc under Cygwin and god knows what else. The easiest way to do that is just to have a Unix computer lab students have access to.
I agree; student laptops are useful for generic computer usage, but not that great for assuming a particular set of software, unless you're going to go the extra step and mandate that students buy a particular computer with a particular OS and software environment. If you aren't going to do that, you're stuck with some of your students running Windows, some OS X, a handful Linux, and very little you can assume about what they can install and run.
If you have a computer lab with some known software install, you know that if you want to use some Mac-only app in the curriculum, for example, you can send them over to the Mac lab to use it (likewise for Unix- or Windows-only stuff).
If the only claim is that science converges in the limit to truth in the mathematical sense, then the statement is about as good as not saying anything at all about science, since it literally says nothing at all about science's accuracy within the next 5 million years. I assumed the great-grandparent poster was making some sort of claim about the convergence of science towards truth on human timescales.
I assumed he meant something stronger about things on human time scales, roughly that science accumulates truth with relatively minor setbacks. If you take asymptotic convergence literally, then yes you're correct, but it's also a useless statement, because it's possible for science to asymptotically converge to the truth without *any* science in the present day or for the next 5 million years to be true.
Presumably he didn't mean a statement as weak as, "science will one day converge to the truth, even though it might be the case that all of science that we've ever done is wrong, and all science for the next several million years might also be wrong".
What sort of timescale? To quote a bit of widely held scientific method, the problem is that that claim is unfalsifiable: any example of a long-held theory that turned out to be wrong, no matter how long it was held, can be rationalized as "well, over a longer timescale, science converges to the truth". The only way to disprove that would be to wait an infinitely long time for science to fail to converge to the truth.
Asimov may have had a PhD in science, but he did nothing of note in science, spending virtually his entire adult life writing science fiction and popular-science works, and virtually none of it writing peer-reviewed journal articles actually on scientific topics. His writings, unsurprisingly, therefore tend to take the mythologized view of science common in sci-fi and pop-science.
Most scientists don't understand science, outside their tiny provincial field; I'm a scientist so I see this all the time. Most have very fairy-tale notions of the scientific method and knowledge production in particular.
You might want to read up on some of the people (scientists especially) who have taken the time to understand how science works, and written on the philosophy and sociology of science.
In particular, it is certainly not true that science converges asymptotically to the truth. It oven diverges substantially, sometimes for hundreds of years, before entire fields (like "racial hygiene") are thrown out as failed experiments. We're currently in the middle of a debate over whether string theory should be placed in that dustbin or not, for example.
Very few scientists in other fields consider economics a legitimate science. In its partial defense, it's studying very complex phenomena--- considerably more complex than understanding the weather, for example, which is itself no cakewalk. On the other hand, economics doesn't seem to really understand that it's dealing with complex dynamical systems, and has been extremely slow to import the tools now standard in all other areas that deal with complex dynamical systems (including weather). Instead they seem to rely mainly on equilibrium assumptions that are unlikely to ever be even approximately true.
The reason you're finding curricula that don't include things you consider necessary for CS is that the number of things that a significant proportion of people consider necessary for CS has ballooned to the point where it doesn't fit in 4 years. Since there's strong resistance to turning CS into a 5-year degree (my alma mater actually suggested it and got shot down quickly), something has to go.
Different schools make different choices about what, treat the rest as block boxes, and people agree with those choices to varying extents. You seem to be mainly focused on what's sometimes called "old-school core CS", the low-level bit-banging and compilers/OS level on top of that. Many schools do focus on that, but the tradeoff is that they treat other things as black boxes; e.g. instead of knowing how to design an algorithm with good asymptotic running time, you look up an algorithm in a book, and focus on how to implement it efficiently on your current hardware. There are lots of other elements; a big one is user interfaces, which are a core part of building apps that anyone can actually use (some people argue that the entire field of CS is about communication, between humans and machines), but treated poorly in low-level-focused CS programs.
I agree keyboards with small key travel are worse, but there is a considerable range of key travel you can get in non-buckling-spring keyboards, too. I'm perfectly happy with my Microsoft Ergonomic 4000 keyboard, and find the larger travel does make it a lot better than both my laptop's very shallow keyboard, and an old Dell desktop keyboard I used to use. (All 3 are rubber-dome.)
For example, maybe you prefer black paint, but are willing to offset the effect by driving a smaller, and more fuel-efficient car. Why shouldn't driving a black 90-hp econobox be an option, when that's actually considerably more environmentally friendly than a white Hummer? Why is California banning the first but not the second?
Raising their gas tax significantly is the most straightforward way to tax inefficiency in cars, without micromanaging where exactly the inefficiency comes from. Use more gas, pay more tax.
Maybe our legislators who are always so worried about sexual exploitation of children as an excuse to censor the internet and everything else, might want to look into whether prohibiting the government from forcibly stripping children naked shouldn't be a higher priority.
At least some of the founders believed that corporations could, if left unchecked, become quasi-governments and similar threats to individual liberty. Thus Thomas Jefferson, for example, expressed the opinion that powerful multi-national corporations should be brought in check. That doesn't seem to be part of the modern libertarian agenda.
Insofar as the language proper is defined by the language standards, the I/O libraries are part of C, because they're specified in the ANSI C and C99 reports. Any conforming C implementation must have the standard I/O functions, and they must behave in the way the standard specifies. That differs quite a bit from the situation with networking libraries, which are third-party and not covered by the C standard.
The problem is still the efficiency, though. There are lots of ways to mark units of computation as "this could be done separately, but depends on Y"--- OpenMP provides a bunch of them, for example, and there's been proposals dating back to the 80s, probably earlier. The problem is figuring out how to implement that efficiently, though, so that the synchronization overhead doesn't dominate the parallelization gains. Does the system spawn new threads? Maintain a pool of worker threads and feed thunks to them? Some hybrid approach? How does it determine when it's worth the effort of doing anything for a particular bit of computation versus just doing it inline and saving the overhead? Etc.
Basically Grand Central is yet another in the decades-long line of proposals for specifying parallelizable computations. What's still an open question is whether they've solved the harder part, a way to, as you say, "[route] computation packets to wherever they can go, and then [receive] the results", without that routing and receiving taking inordinate overhead.
Sure, but then you lost a lot of it again: from 1937 to 1942, the market dropped 50%, so your $450,000 was back to $225,000. If you can perfectly market time, then sure, you can always make money (even in down markets--- just time your short-selling well). Most investors likely bought either too early or two late (most of the gains were very soon after the bottom), or held on too late past 1937.
And without a resolution to the current political crisis in Pakistan, Pakistanis won't be able to get their tech industry going again. Nearly everything is related if you really stretch.
I'm not saying these things aren't worth discussing, just that it isn't really Slashdot's core competence. I participate at multiple discussion forums for precisely the reason that I do find many things valuable to discuss. But I don't really debate AIG bailouts at Slashdot, and I don't really debate the merits of OpenOffice at DailyKos.
The 1929 stock market peak wasn't reached again until about 1955, and a consistent upwards trend didn't even start until 1942.
Majority owners can't actually do "whatever they want" unless they actually own 100% of the company. The trouble here is that the 20% minority owners are something of an annoying fiction--- the company is by rights bankrupt and the shares worthless, but the government has unwound this mess somewhat ineptly and as a result on paper the company still exists and has value, and 20% of that value is owned in the private sector by people who technically the company has a fiduciary duty to.
Perhaps we should discuss rising income inequality in the US? Or the wars in Darfur? Lots of stuff is worth discussing, but I think having some focus on forums tends to make them a bit better.
So I've been around long enough (insert UID bragging here) to know that Slashdot has a fairly wide definition of "tech news", which is fine. Tech influences lots of things, and vice versa. Plenty of political news is relevant to a tech-news site, from DMCA-related things, to voting machines, to internet-speech restrictions (or just speech restrictions that touch online speech). Discussing that here seems fine.
But really, how does this fit in even if you squint? A tax dispute between AIG and the US government, tangled up with a financial bailout package. A mess, sure, but not an even-slightly-tech-related one.
Sorry, but that's a cop-out. You have a Labor-lead government who wants to remove porn and anything they disagree with from the internet, and is using the Christian loony as cover.
Speaking of becoming America, you don't actually have a "Labour Party" at all--- they spell their name "Labor".
Are those two votes really worth having their party tarred as the party of censorship? It seems like it'll just play into the hands of right-wing critics who consider leftist parties to be the parties of authoritarianism and the nanny state---because in this case they really are.