There are relatively few quasars that are observable. Probably a lot fewer that are observable at the same time by two locations, if the two locations are geographically diverse. It is possible for a third party to monitor these discrete locations. Noise would be different to the two observation locations, which could be overcome using sufficient error coding in the plaintext at of course the loss of plaintext entropy, making it easier for a third party with perhaps a noisier signal (due to being slightly out-of-bound, etc) to obtain the plaintext.
The fundimental problem is that the data is not fully random -- it is mostly deterministic based on the key of what quasar, what frequency and bandwidth, and what time. So an outside person could recover the plaintext by obtaining the observable behavior and trying all keys, or if the outside person could somehow obtain the key.
This is a very similar situation to a "good" pseudorandom number generator. You can transmit the seed for the pseudorandom number generator and generate a one-time pad from the pseudorandom number generator. I guess the difference is that quasar behavior is not observable after the fact, but if it is feasable for the data to be logged then they reduce to similar solutions: find all the pads within the keyspace, xor with the cipher text, and watch for the entropy to drop or visibility of known plaintext.
1554 is, quite possibly, the greatest. Beer. Ever. Here in Austin there are some fine pubs that have it on tap.
Though over the hot Texas summer I've become quite fond of the other end of the spectrum: strong IPA's. Dogfish head 90 minute IPA. Stone Ruination IPA. I'm making myself thirsty.
Nukes are complex and can't detonate without some sort of computer running the show.
Incorrect. There was no computer in either of the nuclear weapons used in the field. Most of what you need is a supercritical mass of highly fissile material. The two nuclear weapons used against Japan in WWII used a small, regular explosion to combine two sub-critical masses of fissile material together.
The nice thing about CDMA is that everyone elses' communication looks like noise, you get frequency diversity to reduce narrowband fade, and you can find all of your multiple paths using a rake receiver. So you can probably get decent SNR even with multipath effects.
And, as we all know, information <= bandwith * log2(1+SNR)
SNR is highly correlated to power output at a given distance, and I think you'll find that telecommunications base stations have a higher power allowance than your 802.11b WAN access point. There are other clever tricks to get more SNR out of a base station, like beamforming.
The UMTS/HSDPA bands are pretty huge... 5 Mhz. So if you're getting SNR of 4db, your theoretical limit is about 5*4/3, or between 6 and 7 mbits/sec, if my calculations are correct.
Sure, the spectrum is shared. So is the bandwith available on wired connections on most of your paths around the internet. My cable modem is shared with the folks in my neighborhood, but I still get great speeds. Globbing together bursty traffic is often a good thing.
MediaFLO is also being rolled out across the country. The MediaFLO distributer has Channel 55 all across the country, and can boradcast at fairly high power.
I've actually seen a MediaFLO handset... and the TV quality actually looks really, really good. Suprisingly good. And the handset I saw had 3 hours of battery life while watching TV. Channel switching times are on par with my DirecTV.
I think both DVB-H and MediaFLO transmit at 30fps @ QVGA. QVGA is about the same size as CIF/D1, which is
very passable TV resolution. QVGA screens are also becoming more common on handsets. And, if you think about it, a 3" diagonal screen at book distance (let's say, 18 inches) is similar to a 40" TV across the room (at, say, 20 feet away).. not a huge TV (I have a projector) but not unuseable.
I read somewhere that MediaFLO was coming to verizon next year!
Cell phones are getting higher quality video playback *and* video recording. In a couple of years, your average cell phone will be a passable digital camera and digital video recorder. What I think is interesting is the formatting. Most phones have their screens oriented vertically... I wonder if in five years we will see more content in "tallscreen" instead of "widescreen" as people use their phones for more and more media...
In a nutshell,
Europe settled on GSM and codified it into law. The USA allowed carriers to use the spectrum they paied for pretty much however they like.
Some carriers chose a TDMA scenario that fit into the existing analog channelization (old TDMA), some chose to rechannelize and use a different TDMA scheme (GSM), and some went with Qualcomm's crazy CDMA. As it turns out, the CDMA carriers could fit more users on the same spectrum, which can mean more money. Not having governmentally-forced standards made this possible.
And, of course, all the 3G stuff is based on some variant of CDMA. So maybe the US solution wasn't so crazy after all.
And it's not always just US v. Europe-and-the-rest-of-the-world. Many things vary around the world. What side of the road do you drive on? Electricity isn't just 110@60hz or 220@50... Japan has 100 @ 50. CDMA 1x is popular in Korea, not just the USA. How many countries use PDC cell phones? Which countries use u-law and which use a-law?
MIMO increases SNR mainly by adding transmit and receive diversity.
A good example of receive diversity is the "rabbit ears" on a TV antenna. Signal that is faded to one antenna is likely to not be faded on the other.
Also, SNR on a wireless LAN is much better than SNR on a WAN like a cell. Which means that it is easier to acheive higher data rates on a LAN. This is why 802.11[abg] has good data rates and also why you can't use it over long distances very well.
With multiple antennas you can also do beamforming, which changes a signal for a user to be more directional. This decreases interference for other folks.
But, of course, you don't have to worry too much about other's signals, their walsh coded signals won't correlate with yours.:-)
So far, nobody has been able to break Shannon's Law. There is no magic wand.
Provisional patents are granted nearly universally and without analysis, they exist so people get some early protection and not have to wait 8 years for the patent to go through the system. They also buy some time for the claimant to actually create the full application. They don't have to be ``real'' patent applications, just some notes attached to the proper forms are sufficient.
It's sort of like calling "dibs" if you are buying a house. It just tells the world that you are doing something, but nothing is really legally binding until a real contract is in place, and even then it's not very binding for the buyer until closing.
So there's really nothing interesting here, yet, other than someone applied for a provisional patent for a story line. If the person doesn't apply for a real patent in a short amount of time (1 year I think) then nothing really happens. And if a real patent is submitted, it will be subject to actual scrutiny. We can discuss how good or bad that scrutiny is, but it's infinitely more scrutinous[0] than what is given for a provisional patent.
As usual, IANAL, but I have been an author on several patents.
Not that slashdot folks would care these days. "The patent office is so lame! OMGWTFSQLBBQ!" On the other hand, I guess the author of the book is pretty smart, by applying for a patent you get Slashdot to give you free advertising for your novel.
Actually, I think that there might be some existing basis for this kind of application. For instance, you can patent the decorative pattern of a set of silverware, preventing from others using your cool new pattern... for a while. Which is the whole theory behind patents... you get a monopoly on your idea for a while, but in exchange you release the idea to the world.
I also think that it might be better for Free Software folks to work with the system rather than against it, by building a patent portfolio of patents under a GPL-like licensing agreement, something that uses patent licensing in a similar way to copyright licensing. It does take money, though.
[0] Scrutinous may not actually be a word. I am not a philologist either.
This is like a host of other academic projects. They all start out with the premise "Suppose I have this grid of CPUs/ALUs/whatever". Then they use an army of grad students to hand code for the grid. You get some interesting SPEC results, publish some papers, and get more research money. This is not new, this has been the case for a long, long time.
But often the ideas don't pan out in real life. With TRIPS, you get inflated IPC results from inflated instruction counts from huge superblock schedules. The TRIPS compiler (last I saw) was not suitable for real life applications. The fact that they can fab things like TRIPS chips and boards only shows that we have so many transistors on a chip these days, any crazy-ass idea you have can be produced.
With modern out-of-order pipelines, you get instruction issues in dataflow order. You have to be very careful when trying to encode dataflow in the instruction set. If you obey it, you can have problems when instructions don't have the latency you expect (like cache misses).
If you want my opinion on where the interesting ideas that will get used in the future are, look at what people are going to do with PS3/Cell. Look at languages and operating environments that facilitate parallel structuring of code. Once you get the program into independent threads of execution with low synchronization, the amount of independant instructions you can present to the machine skyrockets.
I dunno, maybe I'm too critical. The guys at UT Austin are doing interesting things, and getting interesting results. If nobody was doing research, nobody would come up with the clever ideas we need in the industry. Work on the compiler, especially, has usefulness outside of a TRIPS ALU grid. I just think this "grid of ALUs" idea is getting old, and since UT and UW like these ideas, it's most of what you see in ISCA and MICRO.
On the other hand, I guess if you have a billion transistors, why not throw a big grid of ALUs in there?
I guess we'll see what pans out in the industry in the upcoming years. I'd place my money on more threads/CPUs, and not so much in the "sea of ALUs" approach. But I know the company I left last year was thinking of this kind of idea in a real product. If you can make it work for real, you can make some bucks. If not... well you can publish some papers I guess.:-)
I think you'll start to see slow adoption Real Soon Now (next few years). However, there's a big question about spectrum. WiMax vendor folks can buy up spectrum like wireless carriers, but that is expensive. They can use bands that don't require licensing (like your 802.11 devices), but (potentially) you'll get lots of interference.
Also, there is really no unity on spectrum for WiMax stuff yet. For 802.11b, for instance, most devices today work in that 2.4Ghz band, so devices are all compatible. Not so much for 802.16, last I saw there were lots of frequencies that could be used, in both licensed and unlicensed spectrum spaces. And it's unlikely that a device you'll get will have antenna systems designed for every possible allowed 802.16 frequency... which I'd wager means that you will likely need to buy hardware that matches your vendor.
I think for the near term, you should see if you have either WCDMA or CDMA 1xEV-DO rev A data coverage in your area. EV-DO has decent bandwith, and DO rev A really reduces latency and increases reverse link bandwith. As a bonus, you should be able to use the service in most major populated areas... You might have to shell out bucks though. For DO rev A, Sprint and Verizon already own the spectrum, and are starting to roll out these services. The GSM folks are switching to WCDMA, but I don't know the state of their data services. My experience is that GPRS/EDGE doesn't have very good data rates in real life... youll want to stick with the 3G data standards.
Or, if you are lucky, you might find a smaller service provider that uses directional 802.11 in your area.. that might work reasonably well.
Even gcc has a built in quicksort, and gcc has almost nothing built-in.
qsort(3) and bsearch(3) are part of the "stdlib" part of the C library.
They are not on the list of compiler-implementable builtins in C90 nor C99.
So I think you mean, "Even the C library has quicksort".
Remember that C is the True, Pure language, and that the library should be distinct from the language. Just ignore the fancy builtins allowed in C90/C99, they are the insurgance of the heretics.:-)
They don't care so much about lunches, or 20% time, or even pay.
Great engineers want to do interesting things. Google is doing interesting things, if you are in that line of work. Many other companies do interesting work for other fields... if you are an aerospace major, Google probably isn't the place you want to work. If you are skilled in the art of circuit design, Google probably isn't the place you want to work... maybe an ASIC design company would fit you better.
I've been an engineer who is unhappy with half of his group, sometimes frustrated by his manager, and believing that the management of the company was a bunch of blathering idiots. But I enjoyed working there, because I got do very interesting things.
Since then, I've moved to a better company and get to do even more interesting things. But the important part was the interesting things I got to do.
Engineering programs vary vastly based on specialization. University of Wisconsin and University of Texas rank very high in Computer Architecture. Georgia Tech is high in Human-Computer Interfaces, as I recall. Many of the major schools have a top-3 spot in one corner of the research or another.
Slashdot Mirror
And, again, how many of those are observable by both sites at the same time with enough fidelity to be able to share the signal for use as a pad?
Most SRAM is 4T or 6T. Most DRAM is a capacitor with a transistor for control of charge/discharge.
The fundimental problem is that the data is not fully random -- it is mostly deterministic based on the key of what quasar, what frequency and bandwidth, and what time. So an outside person could recover the plaintext by obtaining the observable behavior and trying all keys, or if the outside person could somehow obtain the key.
This is a very similar situation to a "good" pseudorandom number generator. You can transmit the seed for the pseudorandom number generator and generate a one-time pad from the pseudorandom number generator. I guess the difference is that quasar behavior is not observable after the fact, but if it is feasable for the data to be logged then they reduce to similar solutions: find all the pads within the keyspace, xor with the cipher text, and watch for the entropy to drop or visibility of known plaintext.
So basically they have on-chip embedded DRAM, and external DRAM.
So... why do they have embedded DRAM (which isn't as good as DRAM in a process optimized for it) if the external is just as fast (according to TFA)?
Though over the hot Texas summer I've become quite fond of the other end of the spectrum: strong IPA's. Dogfish head 90 minute IPA. Stone Ruination IPA. I'm making myself thirsty.
Beware the ides of March!
BEWARE!
Et tu, nintendo?
Sigh. Like anyone reading Slashdot these days finds this funny. Back in my day... et cetera.
Try koules
Incorrect. There was no computer in either of the nuclear weapons used in the field. Most of what you need is a supercritical mass of highly fissile material. The two nuclear weapons used against Japan in WWII used a small, regular explosion to combine two sub-critical masses of fissile material together.
EV-DO revision A has lower latency. CDMA 1x providers are upgrading.
And, as we all know, information <= bandwith * log2(1+SNR)
SNR is highly correlated to power output at a given distance, and I think you'll find that telecommunications base stations have a higher power allowance than your 802.11b WAN access point. There are other clever tricks to get more SNR out of a base station, like beamforming.
The UMTS/HSDPA bands are pretty huge... 5 Mhz. So if you're getting SNR of 4db, your theoretical limit is about 5*4/3, or between 6 and 7 mbits/sec, if my calculations are correct.
Sure, the spectrum is shared. So is the bandwith available on wired connections on most of your paths around the internet. My cable modem is shared with the folks in my neighborhood, but I still get great speeds. Globbing together bursty traffic is often a good thing.
I've actually seen a MediaFLO handset... and the TV quality actually looks really, really good. Suprisingly good. And the handset I saw had 3 hours of battery life while watching TV. Channel switching times are on par with my DirecTV.
I think both DVB-H and MediaFLO transmit at 30fps @ QVGA. QVGA is about the same size as CIF/D1, which is very passable TV resolution. QVGA screens are also becoming more common on handsets. And, if you think about it, a 3" diagonal screen at book distance (let's say, 18 inches) is similar to a 40" TV across the room (at, say, 20 feet away).. not a huge TV (I have a projector) but not unuseable.
I read somewhere that MediaFLO was coming to verizon next year!
Cell phones are getting higher quality video playback *and* video recording. In a couple of years, your average cell phone will be a passable digital camera and digital video recorder. What I think is interesting is the formatting. Most phones have their screens oriented vertically... I wonder if in five years we will see more content in "tallscreen" instead of "widescreen" as people use their phones for more and more media...
In a nutshell, Europe settled on GSM and codified it into law. The USA allowed carriers to use the spectrum they paied for pretty much however they like.
Some carriers chose a TDMA scenario that fit into the existing analog channelization (old TDMA), some chose to rechannelize and use a different TDMA scheme (GSM), and some went with Qualcomm's crazy CDMA. As it turns out, the CDMA carriers could fit more users on the same spectrum, which can mean more money. Not having governmentally-forced standards made this possible.
And, of course, all the 3G stuff is based on some variant of CDMA. So maybe the US solution wasn't so crazy after all.
And it's not always just US v. Europe-and-the-rest-of-the-world. Many things vary around the world. What side of the road do you drive on? Electricity isn't just 110@60hz or 220@50... Japan has 100 @ 50. CDMA 1x is popular in Korea, not just the USA. How many countries use PDC cell phones? Which countries use u-law and which use a-law?
A good example of receive diversity is the "rabbit ears" on a TV antenna. Signal that is faded to one antenna is likely to not be faded on the other.
Also, SNR on a wireless LAN is much better than SNR on a WAN like a cell. Which means that it is easier to acheive higher data rates on a LAN. This is why 802.11[abg] has good data rates and also why you can't use it over long distances very well.
With multiple antennas you can also do beamforming, which changes a signal for a user to be more directional. This decreases interference for other folks.
But, of course, you don't have to worry too much about other's signals, their walsh coded signals won't correlate with yours. :-)
So far, nobody has been able to break Shannon's Law. There is no magic wand.
It's sort of like calling "dibs" if you are buying a house. It just tells the world that you are doing something, but nothing is really legally binding until a real contract is in place, and even then it's not very binding for the buyer until closing.
So there's really nothing interesting here, yet, other than someone applied for a provisional patent for a story line. If the person doesn't apply for a real patent in a short amount of time (1 year I think) then nothing really happens. And if a real patent is submitted, it will be subject to actual scrutiny. We can discuss how good or bad that scrutiny is, but it's infinitely more scrutinous[0] than what is given for a provisional patent.
As usual, IANAL, but I have been an author on several patents.
Not that slashdot folks would care these days. "The patent office is so lame! OMGWTFSQLBBQ!" On the other hand, I guess the author of the book is pretty smart, by applying for a patent you get Slashdot to give you free advertising for your novel.
Actually, I think that there might be some existing basis for this kind of application. For instance, you can patent the decorative pattern of a set of silverware, preventing from others using your cool new pattern... for a while. Which is the whole theory behind patents... you get a monopoly on your idea for a while, but in exchange you release the idea to the world.
I also think that it might be better for Free Software folks to work with the system rather than against it, by building a patent portfolio of patents under a GPL-like licensing agreement, something that uses patent licensing in a similar way to copyright licensing. It does take money, though.
[0] Scrutinous may not actually be a word. I am not a philologist either.
Cook's Illustrated, my favorite magazine.
Temprorary values can be determined by the chip dynamically, or encoded into existing instruction sets.
But often the ideas don't pan out in real life. With TRIPS, you get inflated IPC results from inflated instruction counts from huge superblock schedules. The TRIPS compiler (last I saw) was not suitable for real life applications. The fact that they can fab things like TRIPS chips and boards only shows that we have so many transistors on a chip these days, any crazy-ass idea you have can be produced.
With modern out-of-order pipelines, you get instruction issues in dataflow order. You have to be very careful when trying to encode dataflow in the instruction set. If you obey it, you can have problems when instructions don't have the latency you expect (like cache misses).
If you want my opinion on where the interesting ideas that will get used in the future are, look at what people are going to do with PS3/Cell. Look at languages and operating environments that facilitate parallel structuring of code. Once you get the program into independent threads of execution with low synchronization, the amount of independant instructions you can present to the machine skyrockets.
I dunno, maybe I'm too critical. The guys at UT Austin are doing interesting things, and getting interesting results. If nobody was doing research, nobody would come up with the clever ideas we need in the industry. Work on the compiler, especially, has usefulness outside of a TRIPS ALU grid. I just think this "grid of ALUs" idea is getting old, and since UT and UW like these ideas, it's most of what you see in ISCA and MICRO.
On the other hand, I guess if you have a billion transistors, why not throw a big grid of ALUs in there?
I guess we'll see what pans out in the industry in the upcoming years. I'd place my money on more threads/CPUs, and not so much in the "sea of ALUs" approach. But I know the company I left last year was thinking of this kind of idea in a real product. If you can make it work for real, you can make some bucks. If not... well you can publish some papers I guess. :-)
Also, there is really no unity on spectrum for WiMax stuff yet. For 802.11b, for instance, most devices today work in that 2.4Ghz band, so devices are all compatible. Not so much for 802.16, last I saw there were lots of frequencies that could be used, in both licensed and unlicensed spectrum spaces. And it's unlikely that a device you'll get will have antenna systems designed for every possible allowed 802.16 frequency... which I'd wager means that you will likely need to buy hardware that matches your vendor.
I think for the near term, you should see if you have either WCDMA or CDMA 1xEV-DO rev A data coverage in your area. EV-DO has decent bandwith, and DO rev A really reduces latency and increases reverse link bandwith. As a bonus, you should be able to use the service in most major populated areas... You might have to shell out bucks though. For DO rev A, Sprint and Verizon already own the spectrum, and are starting to roll out these services. The GSM folks are switching to WCDMA, but I don't know the state of their data services. My experience is that GPRS/EDGE doesn't have very good data rates in real life... youll want to stick with the 3G data standards.
Or, if you are lucky, you might find a smaller service provider that uses directional 802.11 in your area.. that might work reasonably well.
They are not on the list of compiler-implementable builtins in C90 nor C99.
So I think you mean, "Even the C library has quicksort".
Remember that C is the True, Pure language, and that the library should be distinct from the language. Just ignore the fancy builtins allowed in C90/C99, they are the insurgance of the heretics. :-)
D'oh. Getting too late. Just over 95 percent?
Engineers know that 100 only has one significant digit. He could be anywhere from just over 50 percent to just under 150 percent confident.
They don't care so much about lunches, or 20% time, or even pay.
Great engineers want to do interesting things. Google is doing interesting things, if you are in that line of work. Many other companies do interesting work for other fields... if you are an aerospace major, Google probably isn't the place you want to work. If you are skilled in the art of circuit design, Google probably isn't the place you want to work... maybe an ASIC design company would fit you better.
I've been an engineer who is unhappy with half of his group, sometimes frustrated by his manager, and believing that the management of the company was a bunch of blathering idiots. But I enjoyed working there, because I got do very interesting things.
Since then, I've moved to a better company and get to do even more interesting things. But the important part was the interesting things I got to do.
Engineering programs vary vastly based on specialization. University of Wisconsin and University of Texas rank very high in Computer Architecture. Georgia Tech is high in Human-Computer Interfaces, as I recall. Many of the major schools have a top-3 spot in one corner of the research or another.
This is a microkernel.