Seriously, remind me why I need privacy. I forget sometimes.
Well, even if you aren't concerned about someone blackmailing you using data gathered from in part from vehicle GPS data, your life is affected by many other people (politicians, businessmen, etc..) who could be blackmailed in ways that may be (perhaps indirectly) detrimental to yourself or society in general.
It's not science fiction, but since you mention CS Lewis, I thought I'd recommend Till we have Faces. It's not as well known as some of his other books (the space trilogy, Chronicles of Narnia, etc...) but is very good. It is a retelling of the story of Psyche and Cupid from greek mythology, from the point of view of Psyche's sister.
"books by rating" at iblist
on
Top 20 Geek Novels
·
· Score: 4, Insightful
This is not quite the same thing, but iblist maintains a list of top books by rating. Geeks are disproportionally represented in their user base, so this is a not entirely unlike a "favorite geek books" list.
That is a good question, but I myself am wondering about the conspicuous absense of John Ronald Reuel Tolkien. Is the Lord of the Rings not geeky enough?
The Personal Telco Project also has a good list of prism 2 cards, which are well supported in Linux by several different drivers (orinoco, wvlan_cs, host_ap). HostAP allows you to use your computer as an access point, though 'till recently you had to compile it separately.
From the article summary (opting out of the free registration, I didn't read the article), the average household wastes 100 watts continuously on devices that are off. That's 2.4 kwh per day, or 876 kilowatt hours per year. Assuming electricity costs 10 cents per kilowatt hour, that would be $87.60 per year. Assuming there are 100 million households in the US, that would amount to 8.76 billion dollars. Does "over $1 billion" in this context actually mean "about $9 billion"?
That is true, a very short duty cycle could be used. However, I think in this case the author of the article simply assumed that "wireless" in the paper automatically means "wifi", which it certainly does not. (The paper does not specify any particular protocol.)
Good point, then the only people who can screw with the system are corrupt poll workers and ballot counters rather than corrupt software programmers.
Fortunately, individual poll workers have very little influence over the outcome of an election. Given the safeguards that polling places tend to employ, skewing results of a paper election by a significant margin would require a large conspiracy with significant cooperation between many people -- such a thing would be hard to keep secret. On the other hand, bad software can easily skew an election if there are no safeguards in place to verify the count.
I think they were reffering to IP6 containing MAC addresses.
Perhaps, but in that case it was a content-free post, since it only repeated what had already been said (not that that is an unheard-of occurance here on slashdot). Therefore, I assumed the poster was refering to IPv4.
No, they aren't. IP packets are incapsulated in ethernet packets for local hops. Ethernet packets contain the mac address in the header, but these aren't delivered end-to-end unless both ends of a connection are in the same subnet.
In IPv6, it is envisioned that machines could use their mac address for the last 48 bits of their IP address so that they can claim a unique address within a subnet without a dhcp request, but this is only one possible convention. The truly paranoid could use a randomly generated number instead.
It's a bad idea because what happens when the driver ABI changes? You have to wait umpteen months for the company to get off thier asses and fix it - like nVidia.
It seems unlikely at this point that Nvidia will ever open-source their video card drivers. (They might not even be able to legally - it's not uncommon for commercial software to contain code from third parties.) Assuming that this is the case, a stable ABI would make Nvidia's task much easier and would probably result in higher quality Linux drivers.
It also precludes anyone else from fixing bugs in the broken, half assed crap most corporates spit out these days.
Nvidia is no exception to this. Several recent Nvidia driver releases have not worked at all on my GeForce4. When that happens, it is nice to have the option of switching to an earlier driver release. Unfortunately, the way things are now, that may mean installing an earlier kernel as well. About a year ago, "installing" a somewhat older Nvidia driver entailed downloading kernel source so the installer would have something to compile its Linux-to-binary-Nvidia-module interface layer against, then applying a patch to Nvidia's source code to fix some bug, then fixing the Linux kernel source to add an EXPORT_SYMBOL statement that someone had taken out (breaking source compatibility during the middle of a stable kernel series). After recompiling everything, it worked, but it took me several days to figure it out. Most users would have given up.
This is the sort of thing that drives users away from Linux. I would like to live in a world where all the best hardware was fully supported by open source drivers, but it isn't going to happen any time soon. The current model, in which closed source drivers are tolerated but deliberately rendered difficult to support and maintain, doesn't work very well. It would probably be better in the long run to either disallow binary-only modules (and give up on accellerated opengl on most graphics cards, among other things) or to accept them and at least make a basic effort not to break things.
I think if someone really wants to build an ipod killer, they should add a wireless interface. Add software such that any two such devices, when near each other, automatically start copying files back and forth. File transfer priority could be determined by a recommender system.
The company could insulate themselves from an RIAA lawsuit by allowing open source developers write the system for them, and distribute it as an unofficial firmware modification...
Perhaps it is only a matter of time before someone does this.
www.shield.org maintains a database of sources of malicious network traffic. Many organizations submit firewall logs to dshield, so they have a pretty good global view of who the bad apples are on the network. For anyone who administers network connected machines, it's a good idea to periodically look up your IP(s) or subnet(s), and see if anyone has generated any complaints about any of your own boxes.
Caveat: This will probably only identify the most aggregious zombies, and only the ones that are doing things that firewalls can identify as malicious. Just because your IPs don't show up on dshield, doesn't mean they aren't zombies.
Mynetwatchman is a similar service, there may be others as well.
I'm tired of hearing about all these advances that we will NEVER see.
What do you mean? Using DWDM, there have been systems available to run a singlemode fiber pair at multi-terabit speeds for years. If they were able to achieve 1tbps at one wavelength, that's an impressive accomplishment, but the article doesn't say.
Perhaps a subtler question would we whether they used one wavelength or multiple wavelengths. Modern long-distance fiber optic systems can run several hundred OC-192s over one fiber pair (one strand in each direction) using dense wavelength division multiplexing (DWDM), and can easily exceed 1tbps.
I'm sorry, you must be correct. Galileo was put under arrest for heresy for espousing views that conflict with nothing that the bible says. Thanks for clearing that one up!
Not everything done in the name of religion is actually in accordance with the teachings of that religion. Galileo's ideas conflicted with an (at the time) commonly accepted interpretation of the Bible, not the actual content of the Bible. Jesus was arrested as a heretic as well for similar reasons.
Now please point out to be the verse in the bible that states that I should have only a single wife!
I don't believe polygamy is anywhere specifically endorsed or condemned (I could be wrong), however there are numerous instances in the old testament where having more than one wife caused a great deal of conflict, and for that reason one could reasonably conclude that it's a bad idea.
3D raster hardware is fast because it's pipelined. For high-quality indirect illumination, you need to store the whole damn model in RAM and bounce rays all over on different geometry and textures - blowing out your cache and stalling the pipeline.
I never said global illumination was easy... As for storing the whole model in RAM, ray tracers can represent the model using more general primitives than triangles (spheres, cones, quadrics, fractals, etc...), so the RAM requirement might not be such a big issue for some models. And you can always apply the same tricks that polygonal renderers use, like pretend parts of the model you can't see don't exist if they don't contribute much to the global illumination calculation.
The trouble with "photorealistic" games is that they really aren't actually photorealistic at all, because they don't accont for indirect illumination. Games keep adding more and more triangles, yet that approach has reached a point of diminishing returns. Graphics cards have become faster and more programmable, but they're using the wrong algorithms. Should we be surprised when gamers complain that new games are starting to look more and more like old games?
Unfortunately, the alternatives have historically been computationally too difficult for real time use (though one could make the case that ray tracing is actually faster than rasterization once one passes a certain threshold of scene complexity), though that may change as hardware gets faster and algorithms get better.
Here are some animations of the sorts of effects that should be present in games if they are to truly be called "photorealistic" (and no, I don't know how long it took to render those).
Of course, this is orthogonal to the argument of whether games should be photorealistic at all, but I suspect a non-photorealistic renderer based on ray tracing and photon mapping could still look much better in general than a non-photorealistic renderer based on triangle rasterization (assuming you have enough cpu power to do the former).
Anything that needs to be scanned or plugged in to a reader controlled by the merchant is at a bit of risk unless you can absolutely trust that reader.
Properly designed, an RSA smart card or usb token will not divulge its private key to anything you plug it into. The card has the capability to compute a signature of a given message uploaded from the reader. There is a risk of the reader uploading an agreement different than the one you think you are signing (like "I agree to pay $1.00" vs "I agree to pay $1000.00"), but that's not insurmountable if the card/token has some sort of basic confirmation interface.
Plus, numbers that can be typed in are probably a lot easier to shoe-horn into e-commerce sites than USB fobs.
That's just a matter of having the right software.
Slashdot Mirror
It's not science fiction, but since you mention CS Lewis, I thought I'd recommend Till we have Faces. It's not as well known as some of his other books (the space trilogy, Chronicles of Narnia, etc...) but is very good. It is a retelling of the story of Psyche and Cupid from greek mythology, from the point of view of Psyche's sister.
This is not quite the same thing, but iblist maintains a list of top books by rating. Geeks are disproportionally represented in their user base, so this is a not entirely unlike a "favorite geek books" list.
The Personal Telco Project also has a good list of prism 2 cards, which are well supported in Linux by several different drivers (orinoco, wvlan_cs, host_ap). HostAP allows you to use your computer as an access point, though 'till recently you had to compile it separately.
From the article summary (opting out of the free registration, I didn't read the article), the average household wastes 100 watts continuously on devices that are off. That's 2.4 kwh per day, or 876 kilowatt hours per year. Assuming electricity costs 10 cents per kilowatt hour, that would be $87.60 per year. Assuming there are 100 million households in the US, that would amount to 8.76 billion dollars. Does "over $1 billion" in this context actually mean "about $9 billion"?
That is true, a very short duty cycle could be used. However, I think in this case the author of the article simply assumed that "wireless" in the paper automatically means "wifi", which it certainly does not. (The paper does not specify any particular protocol.)
802.11 cards typically consume around 1 or 2 watts. They are probably targeting much simpler radios, like those used in motes.
Perhaps, but in that case it was a content-free post, since it only repeated what had already been said (not that that is an unheard-of occurance here on slashdot). Therefore, I assumed the poster was refering to IPv4.
No, they aren't. IP packets are incapsulated in ethernet packets for local hops. Ethernet packets contain the mac address in the header, but these aren't delivered end-to-end unless both ends of a connection are in the same subnet.
In IPv6, it is envisioned that machines could use their mac address for the last 48 bits of their IP address so that they can claim a unique address within a subnet without a dhcp request, but this is only one possible convention. The truly paranoid could use a randomly generated number instead.
It seems unlikely at this point that Nvidia will ever open-source their video card drivers. (They might not even be able to legally - it's not uncommon for commercial software to contain code from third parties.) Assuming that this is the case, a stable ABI would make Nvidia's task much easier and would probably result in higher quality Linux drivers.
Nvidia is no exception to this. Several recent Nvidia driver releases have not worked at all on my GeForce4. When that happens, it is nice to have the option of switching to an earlier driver release. Unfortunately, the way things are now, that may mean installing an earlier kernel as well. About a year ago, "installing" a somewhat older Nvidia driver entailed downloading kernel source so the installer would have something to compile its Linux-to-binary-Nvidia-module interface layer against, then applying a patch to Nvidia's source code to fix some bug, then fixing the Linux kernel source to add an EXPORT_SYMBOL statement that someone had taken out (breaking source compatibility during the middle of a stable kernel series). After recompiling everything, it worked, but it took me several days to figure it out. Most users would have given up.
This is the sort of thing that drives users away from Linux. I would like to live in a world where all the best hardware was fully supported by open source drivers, but it isn't going to happen any time soon. The current model, in which closed source drivers are tolerated but deliberately rendered difficult to support and maintain, doesn't work very well. It would probably be better in the long run to either disallow binary-only modules (and give up on accellerated opengl on most graphics cards, among other things) or to accept them and at least make a basic effort not to break things.
I think if someone really wants to build an ipod killer, they should add a wireless interface. Add software such that any two such devices, when near each other, automatically start copying files back and forth. File transfer priority could be determined by a recommender system.
The company could insulate themselves from an RIAA lawsuit by allowing open source developers write the system for them, and distribute it as an unofficial firmware modification...
Perhaps it is only a matter of time before someone does this.
www.shield.org maintains a database of sources of malicious network traffic. Many organizations submit firewall logs to dshield, so they have a pretty good global view of who the bad apples are on the network. For anyone who administers network connected machines, it's a good idea to periodically look up your IP(s) or subnet(s), and see if anyone has generated any complaints about any of your own boxes.
Caveat: This will probably only identify the most aggregious zombies, and only the ones that are doing things that firewalls can identify as malicious. Just because your IPs don't show up on dshield, doesn't mean they aren't zombies.
Mynetwatchman is a similar service, there may be others as well.
Perhaps a subtler question would we whether they used one wavelength or multiple wavelengths. Modern long-distance fiber optic systems can run several hundred OC-192s over one fiber pair (one strand in each direction) using dense wavelength division multiplexing (DWDM), and can easily exceed 1tbps.
Unfortunately, the alternatives have historically been computationally too difficult for real time use (though one could make the case that ray tracing is actually faster than rasterization once one passes a certain threshold of scene complexity), though that may change as hardware gets faster and algorithms get better.
Here are some animations of the sorts of effects that should be present in games if they are to truly be called "photorealistic" (and no, I don't know how long it took to render those).
Of course, this is orthogonal to the argument of whether games should be photorealistic at all, but I suspect a non-photorealistic renderer based on ray tracing and photon mapping could still look much better in general than a non-photorealistic renderer based on triangle rasterization (assuming you have enough cpu power to do the former).