And how easy is it for extension writers to ensure that they are not holding too many or too old references to objects in their code?
Given how popular it is to say Firefox consumes a lot of memory, I cannot quiet see how the Mozilla Foundation hasn't found time or will to, for example, write a tool that will track the memory usage per extension. Shouldn't be that hard, just keep track which extension created each object and then use GC to find out the number of bytes reachable from those objects. At the cost of some memory;). Not sure if the competition does that, but I can easily imagine how having memory management per-tab eases this issue a lot.
My 'solution' to the problem is use use ulimit -v 2000000 before Firefox and start it when it dies.
Apparently I'm one of those running Firefox with little RAM, measly 6 gigabytes of it. But it still annoys the hell out of me when Firefox takes more than 4 gigabytes of memory.
Yes, I never shut down Firefox unless it dies or takes all the memory. Yes, it's probably some extension. No, there isn't a good way to track it down other than divide-and-conquer which is manual labor.
While that may be true, would having the keys of all existing iPhone devices be a sample large enough? Or maybe you could link to research that can successfully predict the keys OpenSSL generates. No, Debian OpenSSL doesn't count..
This will help, but from the security POV, only the devices which have not been rooted; after that, incoming firmwares could be intercepted and applied either not at all or only partially.
I suppose they could have sign the firmware for the boot loader to check, but given the root password 'newsys', this doen't seem to go with their style..
However, should they ever choose to replace all checkouts with self-checkouts, I very much suspect that the time savings (for clients) would be gone, because it only needs a few people per checkout to stall the queue.
You know, there are other costs involved than just manufacturing the chip. For example recouping the R&D costs. I would be very surprised if the price difference between dual-core and quad-core CPU's reflected the difference between the manufacturing costs of those two.
Internally it encodes the keyboard as MIDI and then again decodes them in the synthesizer. So basically the portable synthesizer you want is a subproject of this. I guess he figured he might want to use the synthesizer separately at times (much more portable).
Hello? You can buy Bluetooth-USB-modules for as little as $3 from Amazon. And that is the price for a single item sold to a consumer. The prices of development-kit-level items is hardly comparable to the actual price to implement Bluetooth in a mass-produced gizmo.
I wonder how come you didn't find that the 433MHz wireless modems cost $40 at Sparkfun as well..
I believe a large reason, if not the largest, for not using Bluetooth in simple wireless gadgets is the amount of electricity it takes compared to a simpler custom-protocol device on the 433MHz band.
In that case I believe you must often whip out the disassembler to weed out bugs. After all, object files are where the real code lies in and they are merely generated from some (for example).c-files.
..did it configure the DNS for you as well? No? How did you configure that then?
I'm guessing you had already IPv4 configured (manually or via DHCP) and used its DNS settings. I suppose that's a decent solution for the time being, as most hosts already are configured to IPv4. But I suppose at the same time they could have their IPv6 address assigned over DHCP too, even if via a DHCP extension..
(Although I run radvd also, and it's great, until you have accidentally ran it on a host that used to do teredo, and you're no longer running it: happy times removing the IP from all hosts.)
The article isn't loading for me, but: can't they simply measure the amount of CPU used during the benchmark and use that information in the benchmark? I don't think it's basically evil to perform that kind of offloading (except in this case when the rules of 3DMark forbid using empirical data on it to optimize performance; but then again, I would imagine many other pieces of software also get this treatment without bad effects on quality or game experience), but dynamically detecting the situation would definitely be complicated; and it might even sometimes give the wrong answer.
One pretty useful heuristic for this kind of optimization would however be "is the CPU usage high without offloading GPU work to CPU: if so, don't do it". Hey, maybe the drivers could have a 'profiling'-mode, which would perhaps slow the performance but figure out the optimal parameters for running the program.
However, Infiniband latency (according to Wikipedia) is in the range of 200 nanoseconds, which is around 100-1000 faster than a gigabit network. That can make a huge difference in certain operations. Also it seems to me the Infiniband protocol is more suitable for clustering than just shoveling data over local network, with operations such as direct memory access and transactions.
Also I wouldn't say the 8Gbit Infiniband latency is anything to sneeze at, after all, 10Gbit Ethernet gear is just coming.
Well, of course RIAA can also look for certain titles of interest on the web, but it's more difficult than simply searching for generic pirated content.
But, you mention posting links to torrents instead of uploading them as images. Where do you put the actual data in that case?
A program that will clear out meta data of the file (not difficult to expect on a image hosting site, even if only to remove unused thumbnail data) will also eliminate such content. Also, it is easy to automatically detect images that have (significant amount of) content other than image data. Detecting these hid.im-files would require a specific algorithm written for it - although, not a very difficult one. (But prone to fail should the format be changed.)
There's a difference between hacking a device you own and you use, and a system that's used by others.
It's not just IT people. People all times have fiddled with devices they've had, and I believe this is in part how many inventions have been made. It just is that some people do it with computing devices and other people do it with, say, cars. Software-based devices are much more malleable than devices built out of screws and bolts; just changing a few lines can go a long way making the device do what you like, instead of doing what the device builder thought would be good.
It is quite difficult to turn a toaster into a waffle iron. But how about making your mobile phone location-awarely be silent automatically while you are in the movie theater? How about interfacing your car GPS with your mobile phone's GPS to mark the spot where your car is at the moment? It can actually be fun to do those things, which is actually the biggest reason why people do it! Don't you have any spare time? All work and no play?
Regarding the "unpredictability" - well, if they're doing it for themselves, it's just them that need to worry about it. It's not like that all commercial devices behave fully predictably. Why would "an ordinary user" make use of my device? It is each person's personal choice to make use of a customized - or self-built - device.
I too can see that there are interests to not make systems hackable. But where is the downside? History has shown that it has been an excellent way to make a device or platform more popular by giving people the possibility to play around with it. Perhaps an open source project about D.C. metro signalling would indeed allow people to find issues in the system and fix them? Of course, it's not likely to attract a lot of eyes, because many have a metro system they can run at home - not that fun.
It appears you don't have a clear idea of what is being discussed here. How is a PDA-class device with navigation going to cause a car crash? And what would be lost if someone indeed managed to change software on my device without permission, I can always just reflash the firmware and I'm set again?
Having hackable devices means that when you find something that doesn't work or doesn't work the way you like, instead of bitching and whining to the vendor, you are actually in the position to do something about it. Possibly even share the solution with someone else. Which also means that the solution might already be out there.. You might actually learn something while fixing the software, too.
If everything just worked Just The Way I Wanted(TM), there would be no need - or atleast much less need - for hacking. Unfortunately, until the world is perfect, that's not going to happen. I prefer hackable devices over non-hackable just for the capability, even if I never did do any alterations to it, though - why take the chance?
Well, it seems to me N810 has the development side much more official and open: for one, it isn't called hacking;-). Also I doubt the mapping software itself on the TomTom unit is GPL.
But on the other hand, the built-in GPS of N810 is not very sensitive. You also need to pay extra for the navigation features (obviously no source available for that component). The navigation software doesn't seem as nice as the older TomTom software I've tried on a S80 phone. I haven't tried Navit (just read about it on this thread) so no idea if it's any good.
You know what's more ridiculous? Blasting a megawatt of power and have no use of it at all!
Oh yeah, well, someone might be using it for receiving..
I don't think it has yet been suggested - by others than Tesla, or atleast by Nokia - that we should be building separate broadcasters just for broadcasting power to these devices with their abysmal power-efficiency.
Make the user registration form (mandatory to access said services) ask a mandatory piece of information: location. City, town, or what ever your target population is described by. If the answer is not what you want, give the account, but make the website just as defunct as it used to be before making the account.:-)
Slashdot Mirror
And how easy is it for extension writers to ensure that they are not holding too many or too old references to objects in their code?
Given how popular it is to say Firefox consumes a lot of memory, I cannot quiet see how the Mozilla Foundation hasn't found time or will to, for example, write a tool that will track the memory usage per extension. Shouldn't be that hard, just keep track which extension created each object and then use GC to find out the number of bytes reachable from those objects. At the cost of some memory ;). Not sure if the competition does that, but I can easily imagine how having memory management per-tab eases this issue a lot.
My 'solution' to the problem is use use ulimit -v 2000000 before Firefox and start it when it dies.
Apparently I'm one of those running Firefox with little RAM, measly 6 gigabytes of it. But it still annoys the hell out of me when Firefox takes more than 4 gigabytes of memory.
Yes, I never shut down Firefox unless it dies or takes all the memory. Yes, it's probably some extension. No, there isn't a good way to track it down other than divide-and-conquer which is manual labor.
While that may be true, would having the keys of all existing iPhone devices be a sample large enough? Or maybe you could link to research that can successfully predict the keys OpenSSL generates. No, Debian OpenSSL doesn't count..
Well, here's a pro-tip that enhanced my GIMP experience tremendously: try the tab key!
This will help, but from the security POV, only the devices which have not been rooted; after that, incoming firmwares could be intercepted and applied either not at all or only partially.
I suppose they could have sign the firmware for the boot loader to check, but given the root password 'newsys', this doen't seem to go with their style..
"LulzSec" already earned waiting spots in prison with what they've already done. Not doing anymore doesn't constitute "take back".
Sure, but stopping the attacks is a good way to reduce the risk of getting caught in the future, if they aren't caught yet.
Yes, I also prefer self-checkouts.
However, should they ever choose to replace all checkouts with self-checkouts, I very much suspect that the time savings (for clients) would be gone, because it only needs a few people per checkout to stall the queue.
You know, there are other costs involved than just manufacturing the chip. For example recouping the R&D costs. I would be very surprised if the price difference between dual-core and quad-core CPU's reflected the difference between the manufacturing costs of those two.
Until we get superconducting interconnections, there will be massive efficiency losses.
Perhaps advancements in superconductivity is what will finally kick green power sources into a gear?
Internally it encodes the keyboard as MIDI and then again decodes them in the synthesizer. So basically the portable synthesizer you want is a subproject of this. I guess he figured he might want to use the synthesizer separately at times (much more portable).
Hello? You can buy Bluetooth-USB-modules for as little as $3 from Amazon. And that is the price for a single item sold to a consumer. The prices of development-kit-level items is hardly comparable to the actual price to implement Bluetooth in a mass-produced gizmo.
I wonder how come you didn't find that the 433MHz wireless modems cost $40 at Sparkfun as well..
I believe a large reason, if not the largest, for not using Bluetooth in simple wireless gadgets is the amount of electricity it takes compared to a simpler custom-protocol device on the 433MHz band.
In that case I believe you must often whip out the disassembler to weed out bugs. After all, object files are where the real code lies in and they are merely generated from some (for example) .c-files.
..did it configure the DNS for you as well? No? How did you configure that then?
I'm guessing you had already IPv4 configured (manually or via DHCP) and used its DNS settings. I suppose that's a decent solution for the time being, as most hosts already are configured to IPv4. But I suppose at the same time they could have their IPv6 address assigned over DHCP too, even if via a DHCP extension..
(Although I run radvd also, and it's great, until you have accidentally ran it on a host that used to do teredo, and you're no longer running it: happy times removing the IP from all hosts.)
The article isn't loading for me, but: can't they simply measure the amount of CPU used during the benchmark and use that information in the benchmark? I don't think it's basically evil to perform that kind of offloading (except in this case when the rules of 3DMark forbid using empirical data on it to optimize performance; but then again, I would imagine many other pieces of software also get this treatment without bad effects on quality or game experience), but dynamically detecting the situation would definitely be complicated; and it might even sometimes give the wrong answer.
One pretty useful heuristic for this kind of optimization would however be "is the CPU usage high without offloading GPU work to CPU: if so, don't do it". Hey, maybe the drivers could have a 'profiling'-mode, which would perhaps slow the performance but figure out the optimal parameters for running the program.
However, Infiniband latency (according to Wikipedia) is in the range of 200 nanoseconds, which is around 100-1000 faster than a gigabit network. That can make a huge difference in certain operations. Also it seems to me the Infiniband protocol is more suitable for clustering than just shoveling data over local network, with operations such as direct memory access and transactions.
Also I wouldn't say the 8Gbit Infiniband latency is anything to sneeze at, after all, 10Gbit Ethernet gear is just coming.
Sites are more likely to accept the same amount of data in a picture than in text.
Well, of course RIAA can also look for certain titles of interest on the web, but it's more difficult than simply searching for generic pirated content.
But, you mention posting links to torrents instead of uploading them as images. Where do you put the actual data in that case?
A program that will clear out meta data of the file (not difficult to expect on a image hosting site, even if only to remove unused thumbnail data) will also eliminate such content. Also, it is easy to automatically detect images that have (significant amount of) content other than image data. Detecting these hid.im-files would require a specific algorithm written for it - although, not a very difficult one. (But prone to fail should the format be changed.)
Nope. Those 'packets' don't need to be remembered by the router. Flows, which are created by upon receiving packets, do.
There's a difference between hacking a device you own and you use, and a system that's used by others.
It's not just IT people. People all times have fiddled with devices they've had, and I believe this is in part how many inventions have been made. It just is that some people do it with computing devices and other people do it with, say, cars. Software-based devices are much more malleable than devices built out of screws and bolts; just changing a few lines can go a long way making the device do what you like, instead of doing what the device builder thought would be good.
It is quite difficult to turn a toaster into a waffle iron. But how about making your mobile phone location-awarely be silent automatically while you are in the movie theater? How about interfacing your car GPS with your mobile phone's GPS to mark the spot where your car is at the moment? It can actually be fun to do those things, which is actually the biggest reason why people do it! Don't you have any spare time? All work and no play?
Regarding the "unpredictability" - well, if they're doing it for themselves, it's just them that need to worry about it. It's not like that all commercial devices behave fully predictably. Why would "an ordinary user" make use of my device? It is each person's personal choice to make use of a customized - or self-built - device.
I too can see that there are interests to not make systems hackable. But where is the downside? History has shown that it has been an excellent way to make a device or platform more popular by giving people the possibility to play around with it. Perhaps an open source project about D.C. metro signalling would indeed allow people to find issues in the system and fix them? Of course, it's not likely to attract a lot of eyes, because many have a metro system they can run at home - not that fun.
It appears you don't have a clear idea of what is being discussed here. How is a PDA-class device with navigation going to cause a car crash? And what would be lost if someone indeed managed to change software on my device without permission, I can always just reflash the firmware and I'm set again?
Having hackable devices means that when you find something that doesn't work or doesn't work the way you like, instead of bitching and whining to the vendor, you are actually in the position to do something about it. Possibly even share the solution with someone else. Which also means that the solution might already be out there.. You might actually learn something while fixing the software, too.
If everything just worked Just The Way I Wanted(TM), there would be no need - or atleast much less need - for hacking. Unfortunately, until the world is perfect, that's not going to happen. I prefer hackable devices over non-hackable just for the capability, even if I never did do any alterations to it, though - why take the chance?
Well, it seems to me N810 has the development side much more official and open: for one, it isn't called hacking ;-). Also I doubt the mapping software itself on the TomTom unit is GPL.
But on the other hand, the built-in GPS of N810 is not very sensitive. You also need to pay extra for the navigation features (obviously no source available for that component). The navigation software doesn't seem as nice as the older TomTom software I've tried on a S80 phone. I haven't tried Navit (just read about it on this thread) so no idea if it's any good.
You know what's more ridiculous? Blasting a megawatt of power and have no use of it at all!
Oh yeah, well, someone might be using it for receiving..
I don't think it has yet been suggested - by others than Tesla, or atleast by Nokia - that we should be building separate broadcasters just for broadcasting power to these devices with their abysmal power-efficiency.
I found this some time ago:
http://directory.fsf.org/project/cveyetracker/
which is apparently abandoned. But only 4 years ago :).
Make the user registration form (mandatory to access said services) ask a mandatory piece of information: location. City, town, or what ever your target population is described by. If the answer is not what you want, give the account, but make the website just as defunct as it used to be before making the account. :-)