These tests are total BS, I have never needed more than one question to determine human vs bot: "3OOO + 1 = ?" (using O's for zeros). I don't understand why people feel computers are getting "more human", they're simply not, they're just getting more and more programming. Some bot writer will eventually get burned by this question and add the concept to his bot's programming, but has that really made it more human? I'll just switch to homophones, puns or other language devices, the possibilities are endless. In my opinion creative, original, human behavior from a program is so far off, it's depressing to think about. Also, we humans are getting "smarter" (better at solving problems), so it's not even clear to me that computers are getting "more human" faster than we're learning and differentiating ourselves from programmed behavior.
I think you're preaching to the choir here on killing x86. The x86 ops get translated to RISC ops anyway. What I wonder is why they haven't attempted to release two versions: an x86 version, and a stripped down RISC version without the x86 decoder. Obviously this would be monumental task at all levels of the design, but it would seem they could get similar performance on the RISC version without as much effort as needed for the x86 version since that overhead is removed. I would guess(and hope) that most of their design effort goes into optimizing the design in the RISC world after the instructions are translated anyway. This will never happen though because windows == x86 only. Being able to compile most of the needed applications from source gives hardware designers the freedom to shed legacy interfaces every 5 years instead of every 30. It would be a glorious future if hardware producers started realizing that open source software == greater hardware design flexibility == better performance/cost. Hopefully this is already happening with the shift from x86 to ARM on netbooks.
Yes, I wish this existed. The problem might be that without modification you need a new JVM instance for each new applet, but it could probably spawn a new one each time one is used so it's ready immediately. There are some projects to run multiple programs within one JVM so that might work too. It's odd to an outsider that there's so much work going into optimizing javascript and flash (and from the crazies, silverlight) when java could fill all those roles. It strikes me (a java fanatic) that java is the unix of the web, it's secure and scalable and to those that use it, it seems perfect for the job, but no one uses it, instead everyone uses a solution cobbled together from different sources (adobe, microsoft, google). Again, I'm just an outsider looking in...
The summary cleverly omits that solving a linear equation is neither NP complete nor NP hard, the speed up is from O(n) to O(log n). I think you'd need a ginormous matrix for this to be useful depending on the constants and such (Of course it'd be crime for me to read paper instead of the abstract to actually find out the details). They already have the applications for quantum computing, it will be much more interesting when they actually figure out a way to build the damn things.
I'm sure it's still a significant result and there's a good chance they did something new that can be used in other applications.
I agree. Are there even any universities that don't make you sign this sort of thing? I just assumed it was like that everywhere. I don't understand how this is any different from trying to do this to a company after they paid you to write some code. If you keep the rights to the code, what are they keeping that they paid you for all that time?
Yep I'm happy with the potential. I didn't mean to be so negative on openmoko for releasing more than almost any other hardware manufacturer has. It'd be great though if they came out and said the reason they're not releasing the schematic and pcb files is that they don't want anyone else manufacturing the device. Really all my ranting could have been summarized as "it would take $25K of work to translate those pdfs into a device." (sorry about that)
It's great that there are free tools available. I have used them and they are great for most projects. When you're dealing with a major project such as open moko I have a feeling it'd be a lot like emptying a swimming pool with a spoon though.
I'm not talking about free speech, I'm talking about free beer. The programs I'm talking about (I won't mention any names) just aren't available in _any_ form for > $1000. If someone handed me a copy of the software I'd be more than happy to use it, regardless of the source, it's just not available. But let's assume Cadence or someone else was offering these programs for free, or you just happen to have access to them. If I showed up at work on monday and my boss handed me those pdfs and wanted me to add a resistor somewhere or fix a short and wanted it done by the end of the day (it would take an hour or two with the correct files), that would truely be a laughable scenario. You'd have to re-enter those schematics all over again, then track down the footprints for all the components (and redraw some of them I'm sure), then layout the components again, and hope the program you're using can route it as well as it was done originally (that's assuming there's no signal integrity issues with the board!). Then you'd have to send for a debugging prototype or two because you're going to make a mistake somewhere. All told this is going to cost at least $25,000 including labor. My point is that there's at least a $25,000 gap between those pdf files and a set of gerber files and a list of component suppliers that you know will give you a functional board.
Don't take my word for it, send those pdfs to one of the companies that does this and get an estimate. All I'm saying is that some of these designs (such as openmoko) are really really far away from full disclosure. In my opinion those pdfs should be called open documentation, not open hardware.
Currently almost all "open hardware" is only open in the sense that the information you need to write software for it is available. You need more than the schematics to replicate it. You need files showing how the wires are routed on the pcb (assuming you're going to modify it somehow, why else would you go to all this trouble?). Oh and by the way you need > $1000 software to generate the files you'll send to manufacture the pcb. And after you do that you'll have to locate and buy each of those components on the board, assuming they'll even sell you > 10 of them at a time. And after that you'll have to solder fine pitch leads and bga parts (latter is impossible without some kind of oven and solder paste). You're also assuming there's no roms anywhere on the board. It's not that hard to do all this if there's 10 components on the board, but that is really not the case with a cellphone. A lot of people see schematics and say "oh I could make that thing." It's really not that simple.
Has any desktop version of windows been ported to any other architecture? Methinks not, it would seem porting something as complex as windows to a completely different architecture would be an insurmountable task. I'm not knowledgable of their tools, but I think they'd have to write a new backend for their compilers, and that's just the tip of the iceberg...
Have you met any humanities graduate students lately? Saying professors are productive depends on what metric you're using for productivity. If you can make your own goals(which I would argue many professors can), it's much more easy to be motivated (and to say you've achieved your goal), no?
"apt-get install" isn't the major advantage, it's "apt-get remove" and dpkg -l (list installed packages? i use gentoo...). One of the points of the package manager to keep track of where all those files are hiding so they can be removed later. I'm not an expert on how it works in windows, but I think you have to trust the competence and motives of whoever writes the uninstaller, I don't think the os tracks anything.
Using a package manager, you could theoretically (and i'm sure there's a million reasons why it doesn't work in practice) go back to a clean install by specifying the correct options and removing all the packages you've added since the installation, if you install all of them with the package manager.
One of the reasons why I personally think linux does this better is because I've tried to get back to something resembling a clean install in windows, without any success. Sometimes it's a challenge just to determine what programs are installed (but I've never used vista, so I can't fairly judge if the situation has improved since xp).
What are the chances it lands on its side or back? It would be very embarrassing to land the thing on its side or back, with everything otherwise functional and intact, but still completely useless.
I mostly agree, though I was speculating that most hybrid designs are more unreliable since they require both system to be functional. I would guess the exception is the design with an electric drive and a small gas generator used to charge the batteries (chevy volt for instance). In my opinion all the other hybrid designs are of less value because 1) they are more complicated to design and build, 2) they don't motivate people towards getting a larger percentage of the energy from the electric grid.
They make, and people buy, hybrid suvs for example. As far as I'm concerned this is even worse than buying the gas powered suv because people assume they're "doing their part" because it's a hybrid, and so they'll keep buying this design as long as they're available. I don't think making them plug-in is going to help because they're so heavy that they'd require a huge battery stack and a larger electric motor to make that work, to the point where there's no gain in efficiency.
I think the plug-in vehicles with an electric drivetrain and small gas generator is vastly superior because it makes people aware of how far they can go on just the batteries before they start burning gas, while still giving them the assurance that they'll never get stuck anywhere. If the owners of one of these realize they haven't bought any gas in 2 years, it's highly likely they'll look for a model without a generator for their next purchase.
I don't think consumers are making illogical emotional associations, I think they're just using an incorrect information in making decision. For instance, if they think there's nothing wrong with burning fossil fuels and the cost of fuel is low, why would they ever buy a more expensive electric vehicle or a hybrid?
The problem is that a lot of people associate 0 benefit to protecting the environment, and 0 penalty to any consequences more than 20 years in the future. Under these assumptions, why would they ever go out of their way to pay more for designs that address these issues they don't care about?
So what happens when I run my three year old.exe on them? Then wouldn't this "unfair benchmark" be a fair indication of performance? This is why things need to be recompiled per microarchitecture, not isa.
If you're just using software (and just mainstream stuff), there's not much point. However, if you're a developer or like to poke around in the sources, you can't ask for a much nicer system right out of a box. You have the source for every package on your system, as well as pretty much every compiler known to man, and they're all configured reasonably without much fuss. If you find some random piece of software on the internet and it's not in your package manager, it can be a painful experience to compile it on some binary based distros, but you probably have much better odds of getting it to compile on gentoo, because your whole system is configured for compiling everything from scratch. Of course, there's plenty things that are quite painful with gentoo, but usually it's up front and well documented.
Slashdot Mirror
They're relying on "we tried it with a few in the same area and nothing bad happened" from the DARPA urban challenge at this point.
These tests are total BS, I have never needed more than one question to determine human vs bot: "3OOO + 1 = ?" (using O's for zeros). I don't understand why people feel computers are getting "more human", they're simply not, they're just getting more and more programming. Some bot writer will eventually get burned by this question and add the concept to his bot's programming, but has that really made it more human? I'll just switch to homophones, puns or other language devices, the possibilities are endless. In my opinion creative, original, human behavior from a program is so far off, it's depressing to think about. Also, we humans are getting "smarter" (better at solving problems), so it's not even clear to me that computers are getting "more human" faster than we're learning and differentiating ourselves from programmed behavior.
I think you're preaching to the choir here on killing x86. The x86 ops get translated to RISC ops anyway. What I wonder is why they haven't attempted to release two versions: an x86 version, and a stripped down RISC version without the x86 decoder. Obviously this would be monumental task at all levels of the design, but it would seem they could get similar performance on the RISC version without as much effort as needed for the x86 version since that overhead is removed. I would guess(and hope) that most of their design effort goes into optimizing the design in the RISC world after the instructions are translated anyway. This will never happen though because windows == x86 only. Being able to compile most of the needed applications from source gives hardware designers the freedom to shed legacy interfaces every 5 years instead of every 30. It would be a glorious future if hardware producers started realizing that open source software == greater hardware design flexibility == better performance/cost. Hopefully this is already happening with the shift from x86 to ARM on netbooks.
Call me ignorant or incompetent, but 'sudo su' gets the job done...
Yes, I wish this existed. The problem might be that without modification you need a new JVM instance for each new applet, but it could probably spawn a new one each time one is used so it's ready immediately. There are some projects to run multiple programs within one JVM so that might work too. It's odd to an outsider that there's so much work going into optimizing javascript and flash (and from the crazies, silverlight) when java could fill all those roles. It strikes me (a java fanatic) that java is the unix of the web, it's secure and scalable and to those that use it, it seems perfect for the job, but no one uses it, instead everyone uses a solution cobbled together from different sources (adobe, microsoft, google). Again, I'm just an outsider looking in...
The summary cleverly omits that solving a linear equation is neither NP complete nor NP hard, the speed up is from O(n) to O(log n). I think you'd need a ginormous matrix for this to be useful depending on the constants and such (Of course it'd be crime for me to read paper instead of the abstract to actually find out the details). They already have the applications for quantum computing, it will be much more interesting when they actually figure out a way to build the damn things.
I'm sure it's still a significant result and there's a good chance they did something new that can be used in other applications.
I agree. Are there even any universities that don't make you sign this sort of thing? I just assumed it was like that everywhere. I don't understand how this is any different from trying to do this to a company after they paid you to write some code. If you keep the rights to the code, what are they keeping that they paid you for all that time?
I wonder if any students take out ads themselves containing key equations or formulas. Or for the evil-minded, ads with incorrect formulas...
Yep I'm happy with the potential. I didn't mean to be so negative on openmoko for releasing more than almost any other hardware manufacturer has. It'd be great though if they came out and said the reason they're not releasing the schematic and pcb files is that they don't want anyone else manufacturing the device. Really all my ranting could have been summarized as "it would take $25K of work to translate those pdfs into a device." (sorry about that)
It's great that there are free tools available. I have used them and they are great for most projects. When you're dealing with a major project such as open moko I have a feeling it'd be a lot like emptying a swimming pool with a spoon though.
Dude, I do this for a living.
eaglecad: tools for signal integrity?
digikey: some parts only come in reels of 1000 or more
smds: what's your solutions for bgas?
I'm not talking about free speech, I'm talking about free beer. The programs I'm talking about (I won't mention any names) just aren't available in _any_ form for > $1000. If someone handed me a copy of the software I'd be more than happy to use it, regardless of the source, it's just not available. But let's assume Cadence or someone else was offering these programs for free, or you just happen to have access to them. If I showed up at work on monday and my boss handed me those pdfs and wanted me to add a resistor somewhere or fix a short and wanted it done by the end of the day (it would take an hour or two with the correct files), that would truely be a laughable scenario. You'd have to re-enter those schematics all over again, then track down the footprints for all the components (and redraw some of them I'm sure), then layout the components again, and hope the program you're using can route it as well as it was done originally (that's assuming there's no signal integrity issues with the board!). Then you'd have to send for a debugging prototype or two because you're going to make a mistake somewhere. All told this is going to cost at least $25,000 including labor. My point is that there's at least a $25,000 gap between those pdf files and a set of gerber files and a list of component suppliers that you know will give you a functional board.
Don't take my word for it, send those pdfs to one of the companies that does this and get an estimate. All I'm saying is that some of these designs (such as openmoko) are really really far away from full disclosure. In my opinion those pdfs should be called open documentation, not open hardware.
Currently almost all "open hardware" is only open in the sense that the information you need to write software for it is available. You need more than the schematics to replicate it. You need files showing how the wires are routed on the pcb (assuming you're going to modify it somehow, why else would you go to all this trouble?). Oh and by the way you need > $1000 software to generate the files you'll send to manufacture the pcb. And after you do that you'll have to locate and buy each of those components on the board, assuming they'll even sell you > 10 of them at a time. And after that you'll have to solder fine pitch leads and bga parts (latter is impossible without some kind of oven and solder paste). You're also assuming there's no roms anywhere on the board. It's not that hard to do all this if there's 10 components on the board, but that is really not the case with a cellphone. A lot of people see schematics and say "oh I could make that thing." It's really not that simple.
as long as it's 90 degrees counter clockwise
Octave is a free version of Matlab, practically all your Matlab code will work in Octave.
Has any desktop version of windows been ported to any other architecture? Methinks not, it would seem porting something as complex as windows to a completely different architecture would be an insurmountable task. I'm not knowledgable of their tools, but I think they'd have to write a new backend for their compilers, and that's just the tip of the iceberg...
Have you met any humanities graduate students lately? Saying professors are productive depends on what metric you're using for productivity. If you can make your own goals(which I would argue many professors can), it's much more easy to be motivated (and to say you've achieved your goal), no?
"apt-get install" isn't the major advantage, it's "apt-get remove" and dpkg -l (list installed packages? i use gentoo...). One of the points of the package manager to keep track of where all those files are hiding so they can be removed later. I'm not an expert on how it works in windows, but I think you have to trust the competence and motives of whoever writes the uninstaller, I don't think the os tracks anything.
Using a package manager, you could theoretically (and i'm sure there's a million reasons why it doesn't work in practice) go back to a clean install by specifying the correct options and removing all the packages you've added since the installation, if you install all of them with the package manager.
One of the reasons why I personally think linux does this better is because I've tried to get back to something resembling a clean install in windows, without any success. Sometimes it's a challenge just to determine what programs are installed (but I've never used vista, so I can't fairly judge if the situation has improved since xp).
What are the chances it lands on its side or back? It would be very embarrassing to land the thing on its side or back, with everything otherwise functional and intact, but still completely useless.
I mostly agree, though I was speculating that most hybrid designs are more unreliable since they require both system to be functional. I would guess the exception is the design with an electric drive and a small gas generator used to charge the batteries (chevy volt for instance). In my opinion all the other hybrid designs are of less value because 1) they are more complicated to design and build, 2) they don't motivate people towards getting a larger percentage of the energy from the electric grid.
They make, and people buy, hybrid suvs for example. As far as I'm concerned this is even worse than buying the gas powered suv because people assume they're "doing their part" because it's a hybrid, and so they'll keep buying this design as long as they're available. I don't think making them plug-in is going to help because they're so heavy that they'd require a huge battery stack and a larger electric motor to make that work, to the point where there's no gain in efficiency.
I think the plug-in vehicles with an electric drivetrain and small gas generator is vastly superior because it makes people aware of how far they can go on just the batteries before they start burning gas, while still giving them the assurance that they'll never get stuck anywhere. If the owners of one of these realize they haven't bought any gas in 2 years, it's highly likely they'll look for a model without a generator for their next purchase.
I don't think consumers are making illogical emotional associations, I think they're just using an incorrect information in making decision. For instance, if they think there's nothing wrong with burning fossil fuels and the cost of fuel is low, why would they ever buy a more expensive electric vehicle or a hybrid?
The problem is that a lot of people associate 0 benefit to protecting the environment, and 0 penalty to any consequences more than 20 years in the future. Under these assumptions, why would they ever go out of their way to pay more for designs that address these issues they don't care about?
Do you have a source for pure electric cars being more unreliable? I'd agree that hybrids are, but I think that's because they have both systems...
you might try pulseaudio and then mute firefox.
So what happens when I run my three year old .exe on them? Then wouldn't this "unfair benchmark" be a fair indication of performance? This is why things need to be recompiled per microarchitecture, not isa.
enron?
If you're just using software (and just mainstream stuff), there's not much point. However, if you're a developer or like to poke around in the sources, you can't ask for a much nicer system right out of a box. You have the source for every package on your system, as well as pretty much every compiler known to man, and they're all configured reasonably without much fuss. If you find some random piece of software on the internet and it's not in your package manager, it can be a painful experience to compile it on some binary based distros, but you probably have much better odds of getting it to compile on gentoo, because your whole system is configured for compiling everything from scratch. Of course, there's plenty things that are quite painful with gentoo, but usually it's up front and well documented.
I think there's bigger problems than memory throughput, keeping the memory hierarchy coherent for 1000 L1 caches would be prohibitively difficult ...