I said there are 50 different PhD programmes here. One of them is "Women's Studies and Feminist Research", I then made a quip about people doing good work in 48 of those programmes, but specifically not journalism or women's studies. Nothing more than that. And honestly, overall, that had nothing to do with what I was getting at.
Let us not forget that democracy gives the people the ability to choose things contrary to what other people choose. It is much easier (and cheaper) to sway a politician than to sway the masses. Germany and france didn't come around to their current borders until about 1956/57 when the french gave up on taking over the Saar. That is, after 900 years of stabbing, shooting and occupying each other, the recent total occupation of cosmopolitan france, and all of a (a newly defined) germany, the killing of millions of people - they were still squabbling over who gets to keep what for themselves for a decade.
People, as a whole, can, and will choose what benefits them, even if it as at the expense of someone else. If we give people democracy a hell of a lot of them aren't going to go the nelson mandella truth and reconciliation route, they are going to demand territories which cannot be given voluntarily. And who do you side with? How do you even define what is a legitimate democratic outcome or not, is a majority of people in the middle east a legitimate democratic outcome, or does it need to be done country by country? If the world votes against the US existing and decides to carve it up and redistrict it back to mexico, spain the UK and various native inhabitants, is that democracy we want to support?
Democracy is a dangerous, and deeply flawed idea. It is suitable in conjunction with other systems but by itself it is a path to a very dark place, albeit rarely, but those places are very dark. The challenge the world faces is building systems which both represent the best interests of the people, including taking their opinion into account, and resolving when those two things (best interests and desires) do not align. But if people will vote for less taxes, more spending, conquest at the expense of others and so on, then democracy is unsustainable, and must be balanced by control from people who actually have some sense. The people who are in control, are, in turn, hopefully balanced against being nuts and can be removed if they fail that test. But democracy has a tendency to form a feedback loop of corruption and incompetence. I'm sure there's ways to deal with that, but not in a/. post.
Just as there is the fields medal in mathematics (and the new, perhaps more appropriate Abel prize), there is the ACM A.M. Turing award for computing.
The problem with making more nobel prizes is where do you draw the line? Why isn't there one for astronomy and astrophysics, separate from the one for physics (these guys really do complain about being lumped together alot), or organic, and inorganic chemistry. How about splitting the nobel prize in medicine into a 'procedures' and a 'biochemistry' category.
Why not a Nobel prize in business, as separate from a Nobel prize in economics? Or different sub branches of economics.
Hell, there are, at just the school I am at, (exactly) 50 different PhD programmes offered. Why doesn't each of those get a nobel prize? Women's studies and feminist research, history, music etc. There are people who do great work in all of those 50 programmes, well, ok, maybe not journalism or women's studies, but the other 48 anyway,
Nobel prizes are an odd tool. They are largely awarded, in the sciences at least, well after the work is done, and in many cases awarded clearly in a sequence (so that they can award both the discoverer of something really cool *and* all the people who made that discovery possible). Computing doesn't quite seem to be ready for that yet. All of the big work, especially on the hardware side, is done by corporations, with huge arrays of people involved, and as much as there are a lot of people who develop a lot of really neat and powerful novel algorithms they get Turing awards already... It would seem kinda silly to be rewarding Intel, or IBM or the like for their fundamental computer research. They do a lot of it, and they deserve industry recognition, (which they get), but I'm not sure it makes much sense to be handing them a nobel prize.
It's probably floating around because it doesn't matter anymore. The only likely candidates for nVidia aquisition that anyone would care about are intel and Sony. Qualcomm, HP, IBM, other PCB/IC outfits or big tech companies could buy nVidia without changing much in the marketplace. But MS isn't about to offer up any sort of meaningful counter offer to prevent a nvidia buyout unless it's Sony offering (sony uses nvidia parts after all). If qualcomm or IBM or HP or any of those guys offered up 3 billion dollars (or 4 billion or whatever it would cost), MS would say 'good luck!'. And if intel offered, well then you have DoJ lawyers more than MS meddling with things.
Nvidia is a fabless semiconductor company, who happens to design reference parts consumers have actually heard of. The only people likely to be interested in them are other semi conductor companies (at least for an acquisition), or a big electronics company. They have a very cool software suite and computing tools for semiconductor simulation, but that's not MS's business at all. I'm not even sure the semi conductor fab guys would want it.
Of course, if journalists were smart enough to be scientists they wouldn't be journalists.
At best a journalist is a noble spy, at worst a corporate lackey, neither of those are people who actually know how to actually do analysis. Journalists like to think, because the process is superficially similar, that they understand science. Research, theory, investigate theory, analyse results, reveal results. But their theories are necessarily simple because they lack the ability to do proper analysis. Blogs are destroying traditional journalism because people who actually know something on a topic can do proper analysis.
Sometimes (120 of the last 140 years of biology for example) has been 'measure a lot and hope you figure out what the hell it means'.
A *lack* of correlation is a result too. And enough sets of correlation eventually leads someone to form a model of what it means.
You can measure to test a model, or measure to build up enough information to construct a meaningful model. If you find out the current model is invalid, through experimental testing, you may not have a broader result in mind than 'see how age affects driving'.
Real science is quantitative analysis of, for example, exactly how much worse drivers get with age. The specific mechanics of what things they get worse at, etc.
The media takes that, and takes the conclusion: they get worse with age/disease, and leave out the details. The details are for, well, people who actually build cars, or systems or the like. The researcher usually isn't trying to prove a 'duh' point, they're trying to quantify a 'duh' point.
Beautiful women are distracting. Ok. By how much? How do you quantify that? How do you study that? If the presence of beautiful women reduce men's productivity by 0.5% that's very different than 25% - the trend, and effect, may be the same (assuming you can quantify to that scale) to the media. But one is good science, one isn't (and no, you can't even express good science in 2 sentences).
If Microsoft built MSE out of the box into windows they would find themselves in front of a court before it could run its first AV scan.
And MSE has updates every day. How long has it taken Apple to roll out an update? Oh, and it's self updating.
How is a process that runs invisibly ever a good thing? What do you do if that AV has a bug in it, or otherwise breaks things? How do you turn it off if it accidentally keeps frying something important?
I'm not sure 'built into the system' means anything. Calculator apps are 'built into the system' and I can live without them. It matters what AV products can hook into, which both seem to be adequate at.
Because one is an evolution of the other, and the producers and designers ultimately decide the direction of products. You can do your job absolutely perfectly as a programmer, and still have the project be a complete disaster.
In many cases (in fact virtually all programming places I've been) managers - as in the guys with business degrees and nothing else, are paid less than programmers with equivalent experience. I have a good friend who is an office manager for a team of doctors, and I assure you is paid no where near what they are. Some managers are good and move up. Most are terrible and serve in a support role where they may organize who gets what office and that sort of thing, and make sure you're actually doing your job, but you get paid more.
But technical directors, producers, lead design people are all evolutions of other jobs into management. A good technical director might have started as a programmer - and knows how to task out programming jobs to 100 other programmers that can actually be done. Being good at that, and not wasting 50% of the labour under your supervision is a big deal, and you're paying for people who don't suck.
I suppose one could distinguish between management and leadership. Leaders are paid well because it's hard to find good ones. Management is a support role (just as IT is a support role), but it is the logical path to leadership, and directly shares a lot of skill overlap. But there are lots of routes to leadership that aren't getting a business degree and then shrinking everyone's cube's but 5% so you can squeeze 4% more employees on the same floor.
I know a few guys who a decade ago were in the million a year range, and are now in the 250-300k.
If you're working with people databases, financials, or just on a product that happens to do crazy awesome (minecraft) you can make a pile of money. But expect 100-200k range if you're really good these days it seems like.
That's not to say you can't make money in direct offshoots of programming, for example becoming a producer on a project, where you may touch some programming still but are now more managerial, or design.
Right now those 'fusion' type products are pretty terrible, I'm not sure decent RAM would be enough to save them, at least if you want decent gaming performance. Also, once you stick the CPU and GPU together you can pipe data directly from one to the other, so the RAM side of things probably changes a bit, but I'm not sure how. I've migrated to AI temporarily from the world of GPGPU so I'm not fully up on how, in detail, they need to talk to each other.
AMD may be betting the farm on a strategy that involves buying more expensive memory, that can, say, do both fast in order and fast out of order reads. Even if that's a 50% increase in price (on the RAM) you're still saving big on the GPU side of things, and they might just figure it's a problem technology will pretty easily solve fairly soon. They aren't really all in yet (in terms of products on shelves), so this might be a 4 years from now sort of problem, when the hardware landscape may be very different than today, I'm not sure, and right now that RAM wouldn't get you much, and there would be no reason to manufacture in volume, so no one would use it.
I could certainly see a future of CPU cores, GPU cores, FPGA's etc. That would make programming *very* interesting for high performance applications.
Ya, it seems like a readyboost, but for SSD's would be the way to go.
My problem is that, while I have a pile of data, even the OS has a lot of data I don't want (language packs for example). Because regular disk space is so cheap that wasn't a problem. But caching would seem to solve, rather directly, the problem of having a lot of crap I usually don't want, for the relatively commonly used data I do want.
Even the jumplist in Windows 7 would a viable half assed solution to this problem.
That problem is that they didn't say 1: we're jumping ship and 2: we're launching WP7 devices within the next 30 days.
Letting symbian die a long slow death is monumentally stupid. They've had time to both think about this decision, and build the partnership with MS (or google), and actually implement solutions. Symbian should be dead already, and the only Nokia phone you should be able to buy should have WP7 (or whatever else they could have gone with).
Nokia is the (old) GM of the phone business. They had a phone for everyone. That's not, in and of itself, a bad strategy. But the 'computer' part of the phone (which apple and google are doing well with) sucked, they had great 'phone' parts (call quality, voice dialing etc.). Whether or not WP7, and MS cash can save them remains to be seen. But announcing your current product lineup is dead out the door, and *not* having a replacement ready to launch is begging for trouble.
July 2011 is the 3 year anniversary of the iPhone 3g, which, IMO was the first good computer with a phone attached. Before that, Nokia had better phones, browsers etc. as did RIM. Now my iPhone 3g is, in many ways, a piece of junk (lack of replacable battery for example, general lack of build quality etc.). But on July 11, 2011 (3rd anniversary of the 3g), Nokia won't have an offering on par with the iPhone 3g. That says a lot about what they've been doing, or not doing, for the last 4 years. Don't get me wrong, the N8 isn't a terrible product, but it has no where near the software umph that the iphone does. Better camera, replacable battery, good, terrible software, bad.
By now Nokia should have either an ARM or x86 WP7 on the market, even a single core roughly on par with a nexus S/iphone 4 sort of level. Windows phones might be behind the curve, but with a big backer they might kick their ass in gear. Nothing of the sort has happened, and they're getting nothing done.
I don't think he means OpenCl specifically. OpenCl is a tool that connects you to GPU hardware. GPU hardware is designed for a different problem than the CPU, so they have different performance characteristics, in the not too distant future heterogenous multi core chips that do both the CPU and GPU calculations of today will be mainstream, and there will general purpose computing tools (which OpenCl is a relatively early generation of, along with CUDA) to access that hardware.
While I don't agree with the idea that this is the entire future, it's certainly part of it. Right now you can have 1200mm^2 of top tier parts in a computer, roughly split half and half CPU/GPU - but not every machine needs that, and it's hard to cool much more than that. So long as there's a market which maximizes performance and uses all of that, CPU/GPU integration will not be total. But there will be, especially in mobile and not top end machines 'enough' performance in 600-800 mm^2 of space, which can be a single IC package which will be a combined CPU-GPU.
It is, I suppose, a bit like the integration of the math co-processor into the CPU a decade ago. GPU's are basically just really big co-processors, and eventually all that streaming, floating point mathy stuff will belong in the CPU. That transition doesn't even have to be painful, a 'cheap' fusion product could be 4 cpu cores and 4 GPU cores, whereas an expensive product might be a 8 core CPU in one package, and 8 cores of GPU power on a separate card, but otherwise the same parts (with the same programming API). The unified memory will eventually obsolete the dedicated GPU probably, but GPU RAM is designed for streaming, in order operations, whereas CPU ram is for out of order random memory block grabs, ram that does either in order or out of order equally well would solve that problem (or as long as it does it well enough), but architecturally I would have GPU ram as a *copy* of the piece of memory that the gpu portion of a fusion part will talk to.
As to what the huge market is: OpenCL gives you easier access to the whole rendering subsystem for non rendering purposes. So your 'killer' apps, are laptops, tablets, mobile phones, low powered desktops, really, anything anyone does any sort of 3D on (games, windows 7, that sort of thing), so basically everything, all your drawing tools.
The strategy is poorly articulated with OpenCl, but I see where they're going. I'm not sure what Intel is doing in this direction though, which will probably be the deciding factor, and nVIDIA, rather than positioning for a buyout (by Intel), seems to be ready to jump ship to SoC/ARM type products. Intel doesn't seem to have the GPU know how to make a good combined product, but they can certainly try and fix that.
I have no doubt that if a similar situation happened in the US or canada an adequate number of people could be found who would volunteer. Not everyone eligible would volunteer of course, but you don't need a million 70 year olds to try and fix 5 nuclear reactors at 1 facility. If you don't have enough free volunteers, offer cash, which of course increases the number of volunteers.
In Japan they have healthcare, they have confidence that the state will take care of them if, in 10 years, they *do* get cancer. In Canada and the US we have systems that (for 70/80 year olds at least) would cover them, but people don't have confidence in their governments the same way, though, ironically, the japanese government is in far worse shape than the US.
except that the reactor was built in the 1960's and 70's. The people who made the fundamental design decisions about how large and earthquake and tsunami to build for are the ones at the post retirement age that would be volunteering.
How is pushing data to the machines approved by the vendors any different than pushing data to the machines regardless of whether it's approved by the vendor.
In this day and age how many of them actually want to mail you a physical disk, and even if they do, what good does that do if there's an exploit potentially in the wild already?
hence 'exploit'. If someone discovers and exploit and uses that to slip on a attached storage device - through an exploit - you may have no easy mechanism for logging it.
In a real time connected system you could notice some weird data transfer or disk access if you're lucky, if it's not connected, you're SOL.
Notice the iran stuxnet attack that specifically targetted control systems in siemens equipment.
Even then, it is, in many cases, easier to connect specialized hardware controllers and software to regular operating systems so you aren't writing say all of the GUI tools yourself so that your operators can actually see what's going on.
I haven't gotten a virus in the 15 or so years I've been using a computer. But I don't see why I shouldn't have insurance on my house- sure it probably won't burn down. Even if that only effects 1/1000 people. But it can be a real pain to actually fix a virus infected machine and preserve all the data on it. Oh, and did I mention anti virus is free? When you had to pay for it, or had long, painful downloads that was another matter. When it's free, legal, and no noticeable performance hit what is the harm done to you by having it?
and if it's not connected to a network it becomes a very labour intensive task to push out updates to the systems to prevent against the viruses.
Even if there is a whole internal network, that isn't connected to the internet all the modern computer security holes remain, and you either have to keep them all standalone - and update them all manually, or network them internally, update them all internally (as in, download updates by hand, transfer them to the appropriate internal network), you still need to get the updates out ASAP because you could have security problems.
If anything, at this point, you may be worse off not being connected. Because by the time your IT guy gets around to developing and rolling out images for a dozen different types of regular windows/linux/mac machines, and then all the custom hardware, you may have already been compromised, and you lack a lot of the intrusion detection tools that rely on well, the network, to work.
Imagine you have a computer (for sake of argument lets make it a generic windows 7 PC), that you manually update on the 2nd wednesday of every month (the day after patch tuesday) - it isn't internet connected. Now, this computer has some super important stuff on it. And you want to know it hasn't been accessed via USB or someone just plugging in a network connection to it. How does it alert you if someone *is* trying to compromise it (or doing anything untoward)? By the time you look at it again how do you know if a USB drive has been connected - especially if it exploited a 0 day vulnerability that cropped up in the month gap between patch tuesdays. If you want to update the intrusion detection system to keep it up to date every day, you're going to have to go to *every* computer that has anything important on it, every day to upload virus signatures etc. The internal network faces essentially the same issue, you might have a single point of copy over - which is a single point of failure.
Lets say I'm really into Strategy games. So I search for strategy games once or twice. Find some awesome strategy games sites, bookmark them, and then visit the bookmarks directly. I'm not generating search traffic for strategy games. But I will generate a lot for FPS, games and maybe sports games, because I'm not to into those, and when I do want to find something on them, I have to search for it.
Porn is, in that sense, no different that a series of specialized niche markets. If you're really into something and, through a successful search find that 'thing', well...then you don't search for it anymore. Differentiating between traffic and search is probably not trivial however. Search to me represents traffic that is under represented, or that is advertised badly (imagine if I did a search for 'news for nerds' and didn't find/. that would not say much about interest in news for nerds, only that one of the biggest sources of news for nerds wasn't providing good results).
Slashdot Mirror
I said there are 50 different PhD programmes here. One of them is "Women's Studies and Feminist Research", I then made a quip about people doing good work in 48 of those programmes, but specifically not journalism or women's studies. Nothing more than that. And honestly, overall, that had nothing to do with what I was getting at.
Let us not forget that democracy gives the people the ability to choose things contrary to what other people choose. It is much easier (and cheaper) to sway a politician than to sway the masses. Germany and france didn't come around to their current borders until about 1956/57 when the french gave up on taking over the Saar. That is, after 900 years of stabbing, shooting and occupying each other, the recent total occupation of cosmopolitan france, and all of a (a newly defined) germany, the killing of millions of people - they were still squabbling over who gets to keep what for themselves for a decade.
People, as a whole, can, and will choose what benefits them, even if it as at the expense of someone else. If we give people democracy a hell of a lot of them aren't going to go the nelson mandella truth and reconciliation route, they are going to demand territories which cannot be given voluntarily. And who do you side with? How do you even define what is a legitimate democratic outcome or not, is a majority of people in the middle east a legitimate democratic outcome, or does it need to be done country by country? If the world votes against the US existing and decides to carve it up and redistrict it back to mexico, spain the UK and various native inhabitants, is that democracy we want to support?
Democracy is a dangerous, and deeply flawed idea. It is suitable in conjunction with other systems but by itself it is a path to a very dark place, albeit rarely, but those places are very dark. The challenge the world faces is building systems which both represent the best interests of the people, including taking their opinion into account, and resolving when those two things (best interests and desires) do not align. But if people will vote for less taxes, more spending, conquest at the expense of others and so on, then democracy is unsustainable, and must be balanced by control from people who actually have some sense. The people who are in control, are, in turn, hopefully balanced against being nuts and can be removed if they fail that test. But democracy has a tendency to form a feedback loop of corruption and incompetence. I'm sure there's ways to deal with that, but not in a /. post.
Just as there is the fields medal in mathematics (and the new, perhaps more appropriate Abel prize), there is the ACM A.M. Turing award for computing.
The problem with making more nobel prizes is where do you draw the line? Why isn't there one for astronomy and astrophysics, separate from the one for physics (these guys really do complain about being lumped together alot), or organic, and inorganic chemistry. How about splitting the nobel prize in medicine into a 'procedures' and a 'biochemistry' category.
Why not a Nobel prize in business, as separate from a Nobel prize in economics? Or different sub branches of economics.
Hell, there are, at just the school I am at, (exactly) 50 different PhD programmes offered. Why doesn't each of those get a nobel prize? Women's studies and feminist research, history, music etc. There are people who do great work in all of those 50 programmes, well, ok, maybe not journalism or women's studies, but the other 48 anyway,
Nobel prizes are an odd tool. They are largely awarded, in the sciences at least, well after the work is done, and in many cases awarded clearly in a sequence (so that they can award both the discoverer of something really cool *and* all the people who made that discovery possible). Computing doesn't quite seem to be ready for that yet. All of the big work, especially on the hardware side, is done by corporations, with huge arrays of people involved, and as much as there are a lot of people who develop a lot of really neat and powerful novel algorithms they get Turing awards already... It would seem kinda silly to be rewarding Intel, or IBM or the like for their fundamental computer research. They do a lot of it, and they deserve industry recognition, (which they get), but I'm not sure it makes much sense to be handing them a nobel prize.
It's probably floating around because it doesn't matter anymore. The only likely candidates for nVidia aquisition that anyone would care about are intel and Sony. Qualcomm, HP, IBM, other PCB/IC outfits or big tech companies could buy nVidia without changing much in the marketplace. But MS isn't about to offer up any sort of meaningful counter offer to prevent a nvidia buyout unless it's Sony offering (sony uses nvidia parts after all). If qualcomm or IBM or HP or any of those guys offered up 3 billion dollars (or 4 billion or whatever it would cost), MS would say 'good luck!'. And if intel offered, well then you have DoJ lawyers more than MS meddling with things.
Nvidia is a fabless semiconductor company, who happens to design reference parts consumers have actually heard of. The only people likely to be interested in them are other semi conductor companies (at least for an acquisition), or a big electronics company. They have a very cool software suite and computing tools for semiconductor simulation, but that's not MS's business at all. I'm not even sure the semi conductor fab guys would want it.
Of course, if journalists were smart enough to be scientists they wouldn't be journalists.
At best a journalist is a noble spy, at worst a corporate lackey, neither of those are people who actually know how to actually do analysis. Journalists like to think, because the process is superficially similar, that they understand science. Research, theory, investigate theory, analyse results, reveal results. But their theories are necessarily simple because they lack the ability to do proper analysis. Blogs are destroying traditional journalism because people who actually know something on a topic can do proper analysis.
Sometimes (120 of the last 140 years of biology for example) has been 'measure a lot and hope you figure out what the hell it means'.
A *lack* of correlation is a result too. And enough sets of correlation eventually leads someone to form a model of what it means.
You can measure to test a model, or measure to build up enough information to construct a meaningful model. If you find out the current model is invalid, through experimental testing, you may not have a broader result in mind than 'see how age affects driving'.
Non linear relationships in things people presume are linear is a very interesting result. A still implies B, but not in the way we thought.
Real science is quantitative analysis of, for example, exactly how much worse drivers get with age. The specific mechanics of what things they get worse at, etc.
The media takes that, and takes the conclusion: they get worse with age/disease, and leave out the details. The details are for, well, people who actually build cars, or systems or the like. The researcher usually isn't trying to prove a 'duh' point, they're trying to quantify a 'duh' point.
Beautiful women are distracting. Ok. By how much? How do you quantify that? How do you study that? If the presence of beautiful women reduce men's productivity by 0.5% that's very different than 25% - the trend, and effect, may be the same (assuming you can quantify to that scale) to the media. But one is good science, one isn't (and no, you can't even express good science in 2 sentences).
If Microsoft built MSE out of the box into windows they would find themselves in front of a court before it could run its first AV scan.
And MSE has updates every day. How long has it taken Apple to roll out an update? Oh, and it's self updating.
How is a process that runs invisibly ever a good thing? What do you do if that AV has a bug in it, or otherwise breaks things? How do you turn it off if it accidentally keeps frying something important?
I'm not sure 'built into the system' means anything. Calculator apps are 'built into the system' and I can live without them. It matters what AV products can hook into, which both seem to be adequate at.
Because one is an evolution of the other, and the producers and designers ultimately decide the direction of products. You can do your job absolutely perfectly as a programmer, and still have the project be a complete disaster.
In many cases (in fact virtually all programming places I've been) managers - as in the guys with business degrees and nothing else, are paid less than programmers with equivalent experience. I have a good friend who is an office manager for a team of doctors, and I assure you is paid no where near what they are. Some managers are good and move up. Most are terrible and serve in a support role where they may organize who gets what office and that sort of thing, and make sure you're actually doing your job, but you get paid more.
But technical directors, producers, lead design people are all evolutions of other jobs into management. A good technical director might have started as a programmer - and knows how to task out programming jobs to 100 other programmers that can actually be done. Being good at that, and not wasting 50% of the labour under your supervision is a big deal, and you're paying for people who don't suck.
I suppose one could distinguish between management and leadership. Leaders are paid well because it's hard to find good ones. Management is a support role (just as IT is a support role), but it is the logical path to leadership, and directly shares a lot of skill overlap. But there are lots of routes to leadership that aren't getting a business degree and then shrinking everyone's cube's but 5% so you can squeeze 4% more employees on the same floor.
I know a few guys who a decade ago were in the million a year range, and are now in the 250-300k.
If you're working with people databases, financials, or just on a product that happens to do crazy awesome (minecraft) you can make a pile of money. But expect 100-200k range if you're really good these days it seems like.
That's not to say you can't make money in direct offshoots of programming, for example becoming a producer on a project, where you may touch some programming still but are now more managerial, or design.
Right now those 'fusion' type products are pretty terrible, I'm not sure decent RAM would be enough to save them, at least if you want decent gaming performance. Also, once you stick the CPU and GPU together you can pipe data directly from one to the other, so the RAM side of things probably changes a bit, but I'm not sure how. I've migrated to AI temporarily from the world of GPGPU so I'm not fully up on how, in detail, they need to talk to each other.
AMD may be betting the farm on a strategy that involves buying more expensive memory, that can, say, do both fast in order and fast out of order reads. Even if that's a 50% increase in price (on the RAM) you're still saving big on the GPU side of things, and they might just figure it's a problem technology will pretty easily solve fairly soon. They aren't really all in yet (in terms of products on shelves), so this might be a 4 years from now sort of problem, when the hardware landscape may be very different than today, I'm not sure, and right now that RAM wouldn't get you much, and there would be no reason to manufacture in volume, so no one would use it.
I could certainly see a future of CPU cores, GPU cores, FPGA's etc. That would make programming *very* interesting for high performance applications.
Ya, it seems like a readyboost, but for SSD's would be the way to go.
My problem is that, while I have a pile of data, even the OS has a lot of data I don't want (language packs for example). Because regular disk space is so cheap that wasn't a problem. But caching would seem to solve, rather directly, the problem of having a lot of crap I usually don't want, for the relatively commonly used data I do want.
Even the jumplist in Windows 7 would a viable half assed solution to this problem.
That problem is that they didn't say 1: we're jumping ship and 2: we're launching WP7 devices within the next 30 days.
Letting symbian die a long slow death is monumentally stupid. They've had time to both think about this decision, and build the partnership with MS (or google), and actually implement solutions. Symbian should be dead already, and the only Nokia phone you should be able to buy should have WP7 (or whatever else they could have gone with).
Nokia is the (old) GM of the phone business. They had a phone for everyone. That's not, in and of itself, a bad strategy. But the 'computer' part of the phone (which apple and google are doing well with) sucked, they had great 'phone' parts (call quality, voice dialing etc.). Whether or not WP7, and MS cash can save them remains to be seen. But announcing your current product lineup is dead out the door, and *not* having a replacement ready to launch is begging for trouble.
July 2011 is the 3 year anniversary of the iPhone 3g, which, IMO was the first good computer with a phone attached. Before that, Nokia had better phones, browsers etc. as did RIM. Now my iPhone 3g is, in many ways, a piece of junk (lack of replacable battery for example, general lack of build quality etc.). But on July 11, 2011 (3rd anniversary of the 3g), Nokia won't have an offering on par with the iPhone 3g. That says a lot about what they've been doing, or not doing, for the last 4 years. Don't get me wrong, the N8 isn't a terrible product, but it has no where near the software umph that the iphone does. Better camera, replacable battery, good, terrible software, bad.
By now Nokia should have either an ARM or x86 WP7 on the market, even a single core roughly on par with a nexus S/iphone 4 sort of level. Windows phones might be behind the curve, but with a big backer they might kick their ass in gear. Nothing of the sort has happened, and they're getting nothing done.
I don't think he means OpenCl specifically. OpenCl is a tool that connects you to GPU hardware. GPU hardware is designed for a different problem than the CPU, so they have different performance characteristics, in the not too distant future heterogenous multi core chips that do both the CPU and GPU calculations of today will be mainstream, and there will general purpose computing tools (which OpenCl is a relatively early generation of, along with CUDA) to access that hardware.
While I don't agree with the idea that this is the entire future, it's certainly part of it. Right now you can have 1200mm^2 of top tier parts in a computer, roughly split half and half CPU/GPU - but not every machine needs that, and it's hard to cool much more than that. So long as there's a market which maximizes performance and uses all of that, CPU/GPU integration will not be total. But there will be, especially in mobile and not top end machines 'enough' performance in 600-800 mm^2 of space, which can be a single IC package which will be a combined CPU-GPU.
It is, I suppose, a bit like the integration of the math co-processor into the CPU a decade ago. GPU's are basically just really big co-processors, and eventually all that streaming, floating point mathy stuff will belong in the CPU. That transition doesn't even have to be painful, a 'cheap' fusion product could be 4 cpu cores and 4 GPU cores, whereas an expensive product might be a 8 core CPU in one package, and 8 cores of GPU power on a separate card, but otherwise the same parts (with the same programming API). The unified memory will eventually obsolete the dedicated GPU probably, but GPU RAM is designed for streaming, in order operations, whereas CPU ram is for out of order random memory block grabs, ram that does either in order or out of order equally well would solve that problem (or as long as it does it well enough), but architecturally I would have GPU ram as a *copy* of the piece of memory that the gpu portion of a fusion part will talk to.
As to what the huge market is: OpenCL gives you easier access to the whole rendering subsystem for non rendering purposes. So your 'killer' apps, are laptops, tablets, mobile phones, low powered desktops, really, anything anyone does any sort of 3D on (games, windows 7, that sort of thing), so basically everything, all your drawing tools.
The strategy is poorly articulated with OpenCl, but I see where they're going. I'm not sure what Intel is doing in this direction though, which will probably be the deciding factor, and nVIDIA, rather than positioning for a buyout (by Intel), seems to be ready to jump ship to SoC/ARM type products. Intel doesn't seem to have the GPU know how to make a good combined product, but they can certainly try and fix that.
I have no doubt that if a similar situation happened in the US or canada an adequate number of people could be found who would volunteer. Not everyone eligible would volunteer of course, but you don't need a million 70 year olds to try and fix 5 nuclear reactors at 1 facility. If you don't have enough free volunteers, offer cash, which of course increases the number of volunteers.
In Japan they have healthcare, they have confidence that the state will take care of them if, in 10 years, they *do* get cancer. In Canada and the US we have systems that (for 70/80 year olds at least) would cover them, but people don't have confidence in their governments the same way, though, ironically, the japanese government is in far worse shape than the US.
except that the reactor was built in the 1960's and 70's. The people who made the fundamental design decisions about how large and earthquake and tsunami to build for are the ones at the post retirement age that would be volunteering.
How is pushing data to the machines approved by the vendors any different than pushing data to the machines regardless of whether it's approved by the vendor.
In this day and age how many of them actually want to mail you a physical disk, and even if they do, what good does that do if there's an exploit potentially in the wild already?
hence 'exploit'. If someone discovers and exploit and uses that to slip on a attached storage device - through an exploit - you may have no easy mechanism for logging it.
In a real time connected system you could notice some weird data transfer or disk access if you're lucky, if it's not connected, you're SOL.
Which, I believe, I distinguished between.
They don't. But all the same problems remain.
Notice the iran stuxnet attack that specifically targetted control systems in siemens equipment.
Even then, it is, in many cases, easier to connect specialized hardware controllers and software to regular operating systems so you aren't writing say all of the GUI tools yourself so that your operators can actually see what's going on.
I haven't gotten a virus in the 15 or so years I've been using a computer. But I don't see why I shouldn't have insurance on my house- sure it probably won't burn down. Even if that only effects 1/1000 people. But it can be a real pain to actually fix a virus infected machine and preserve all the data on it. Oh, and did I mention anti virus is free? When you had to pay for it, or had long, painful downloads that was another matter. When it's free, legal, and no noticeable performance hit what is the harm done to you by having it?
what would be the performance hit of avast or MSE, in whatever quantitative metric you have used to determine that it slows your PC down too much.
and if it's not connected to a network it becomes a very labour intensive task to push out updates to the systems to prevent against the viruses.
Even if there is a whole internal network, that isn't connected to the internet all the modern computer security holes remain, and you either have to keep them all standalone - and update them all manually, or network them internally, update them all internally (as in, download updates by hand, transfer them to the appropriate internal network), you still need to get the updates out ASAP because you could have security problems.
If anything, at this point, you may be worse off not being connected. Because by the time your IT guy gets around to developing and rolling out images for a dozen different types of regular windows/linux/mac machines, and then all the custom hardware, you may have already been compromised, and you lack a lot of the intrusion detection tools that rely on well, the network, to work.
Imagine you have a computer (for sake of argument lets make it a generic windows 7 PC), that you manually update on the 2nd wednesday of every month (the day after patch tuesday) - it isn't internet connected. Now, this computer has some super important stuff on it. And you want to know it hasn't been accessed via USB or someone just plugging in a network connection to it. How does it alert you if someone *is* trying to compromise it (or doing anything untoward)? By the time you look at it again how do you know if a USB drive has been connected - especially if it exploited a 0 day vulnerability that cropped up in the month gap between patch tuesdays. If you want to update the intrusion detection system to keep it up to date every day, you're going to have to go to *every* computer that has anything important on it, every day to upload virus signatures etc. The internal network faces essentially the same issue, you might have a single point of copy over - which is a single point of failure.
Lets say I'm really into Strategy games. So I search for strategy games once or twice. Find some awesome strategy games sites, bookmark them, and then visit the bookmarks directly. I'm not generating search traffic for strategy games. But I will generate a lot for FPS, games and maybe sports games, because I'm not to into those, and when I do want to find something on them, I have to search for it.
Porn is, in that sense, no different that a series of specialized niche markets. If you're really into something and, through a successful search find that 'thing', well...then you don't search for it anymore. Differentiating between traffic and search is probably not trivial however. Search to me represents traffic that is under represented, or that is advertised badly (imagine if I did a search for 'news for nerds' and didn't find /. that would not say much about interest in news for nerds, only that one of the biggest sources of news for nerds wasn't providing good results).