You're assuming two things: perfect filters and infinite bit depth. As I argued in my previous post, higher bit depth is a Good Thing. Better filters are left as an exercise to the high-end CD player manufacturers.
More frequency range isn't going to be recorded, played, or heard by anyone.
First of all, things above 22kHz aren't picked up by ordinary mics... Even the ultra-high-end Neumann U87Ai only claims 20-20kHz frequency response (http://www.neumann.com/mics/u87ai.htm)
Secondly, most speakers won't crank out those high frequencies without a severe falloff in response: the high-end Genelec 1038A triamped monitor gets you 33-20k Hz (-3dB). (http://www.genelec.com/products/1038a/1038a.htm)
Finally, most people can't hear above 20kHz, especially those people who are incessantly blasting their ears out with loud music.
The best reason for Super CD (or DVD or whatever) is higher bit depth, NOT higher sampling rate; going from 16/44.1 (CD quality) to 24/44.1 takes just 50% more space, for nontrivially better quality, while going from 16/44.1 to 16/88.2 brings minimal benefit at a 100% space penalty.
NP does not stand for NON polynomial, it stands for nondeterministic polynomial.
Every problem in P is in NP.
Decryption brute force by definition could not be accomplished in polytime; the problem is if you can always guess right in polytime, then it's a different story (essentially how a nondeterministic turing machine works).
</rant
Re:First mistake found
on
Does P = NP?
·
· Score: 1
doesn't matter, since graph edges must be polynomial (actually O(n^2)) in the number of vertices.
You're being paranoid; there are too many amateurs in audio and video for analog to be phased out/made illegal. Quite frankly, if the semipro audio equipment was killed (and any semipro equipment has to allow copies either in digital or analog domain), the RIAA would be killing itself because it would have no way to listen to demo tapes of up and coming artists.
Besides, can you come up with a single piece of legacy equipment which was declared illegal by a US court? Even scanners that can pick up cell phone bands were grandfathered in when that ban was put in place.
Analog audio will always exist because it's easy to build. I can build a mic and speakers from raw components, and digitizing it can be done by any number of off-the-shelf parts, which will need to continue to exist in order for manufacturers to create microphones.
Frankly I don't care that much about video; I already have my APEX...
too expensive to put the amp/decoder in each speaker... there's a reason people use passive monitors (I don't think that's a good thing, but if you're dealing with sound reinforcement, you pretty much either shell out the bucks or go passive).
hacking into the cabinets isn't the only way to do it... my trusty MD player or WAV recorder can be hooked up to a pretty high quality mic,...
legacy will always be a problem. just like you can burn the IBM CD and then rip it to mp3.
We (ad hoc networking community, not necessarily IETF MANET WG community) do pretty well for small nets (Broch et al, Mobicom 98) and geographically routable large nets (Ko and Vaidya, Mobicom 98; Li et al, Karp and Kung, Mobicom 00)
Also, a silly plug, we (the monarch group) haven't updated our latest internet draft because we've been busy writing working code =) Another interesting protocol is DSR:
http://www.monarch.cs.cmu.edu/ietf.html
I'm a wireless nets guy, so I'll talk about your radio points. As for (1), you use some CSMA/CA protocol to share a channel among many users. You could even use TDMA to get some guarantees about latency and bandwidth (subject to an error-free channel)
(2): if you're CDMA, it's not easy. Also, it's kinda illegal, so if people actually used these things, the FCC people with their scopes will come knocking at your door.
Even if you encrypt in the reader and decrypt in the software, someone can reverse engineer the software and release the keys, as was done to DVD's CSS.
If you let software do some descrambling, and give that software to anyone, even as object code, you're already doomed; you may as well not scramble in the first place.
Bah, I disagree about English and German being close enough to consider "dialects;" as a mandarin speaker, I understand a nontrivial amount of Szechuan dialect (without having to learn any), and almost all phrases translate word-for-word. The written forms are the same as well. This is true for Taiwanese and Cantonese as well, AFAIK.
Typing Chinese turns out to be easier than writing Chinese and only marginally harder than reading Chinese if you know any of the standard phonetic alphabets.
It's not particularly complicated either; generally the only strange part is having to look at the screen to pick your character. But OTOH if you used zhuyin to type the previous sentence, you're going to type 20-25 characters, or about 40-75 keystrokes. Besides, if you've ever gone to an internet cafe in Taiwan you'll see people who type Chinese pretty fast.
So saying that you have to learn quite a lot to type chinese is like saying you need to learn quite a lot to type english... you need to learn the alphabet for english (you need to learn an alphabet for Chinese).
FreeBSD network stack is similar to other UN*Xs (much more so than Linux at the kernel level) so changes are easier to port to other OSes. One reason I use FreeBSD, at least.
Music is an expensive hobby. Some of us can pay for it because we have a day job, some of us have to depend on it for a living. Do you think people can afford to put together $100k studios for high-quality pristene sound and spend $1k/hr for recording if music is free? How about software? Software is relatively cheap to make, except for labor. Should all software be free?
10GHz? Don't you have serious propogation problems at that frequency? No one even does 5.8 GHz ISM in spite of the fact that you can put out 4W over a section of that band for two reasons: cost of electronics handling 5.8GHz, and reduced range per watt in that frequency band (vs 2.4GHz)
Realistically, you have to trade bitrate for power, or range for power, or range for bitrate. You can't have it all.
RSA equivalent to factoring (*) If I recall correctly, I took a crypto class where they proved that if you could get nontrivial advantage on the high log_2 n bits of an arbitrary message encrypted using an n-bit RSA key in polytime, you could factor in polytime. (that is, finding the high 10 bits of a message encrypted with RSA is roughly equivalent to factoring the key... so you can use 10 bits per 1024 bit block, and be pretty secure.)
Re:GSM Dual-Band (Re:Bad Business Model)
on
Iridium Saved?
·
· Score: 1
Yes, so even at $6 billion, the finance costs are ridiculous. But you're covering the whole planet. I don't know about your ballpark sales/marketing/incidentals numbers, those should be mostly the responsibility of the people who resell iridium. As I mentioned in my earlier post, O&M alone is killing Iridium. But that's not because of the market; partially Iridium should price discriminate against oceanic markets (where Inmarsat is it's only serious competitor) and slash prices over land (compete with analog roaming in the US, for example)... if they had designed in a less capacity for more power trade, they could sell competitively sized handsets, and they'd have a nice niche market. Besides, DoD is a nice customer; they doubtless have their own LEOs, but don't want to use them in anything but the most dire circumstances.
GSM Dual-Band (Re:Bad Business Model)
on
Iridium Saved?
·
· Score: 1
But all (I think) Iridium phones take SIM-Cards and roam on GSM networks... the point is to have a system that works when you're out of range of GSM networks. Of course, you have to be outdoors and all that, but...
So GSM doesn't cover oceans and such, and Iridium could have made a great market out of that, but I haven't seen that side of it's marketing. (In fact, given the alternatives in that space, Iridium is small and more convienient).
How can you not make money? Two letters: O&M.
on
Iridium Saved?
·
· Score: 1
Even if someone got rid of all of Iridium's debt and gave the constellation to you, you'd still lose money on the O&M (operations and maintenence) agreement with Motorola. MOT's been running the constellation free of charge so that IRID could attempt to make debt payments, but that didn't quite work, but even without debt payments, income < O&M costs, so.... you'd have to get a better O&M deal with MOT to even break even.
Basically what running the constellation means is making sure that the satellites don't fall out of their orbits due to perturbations caused by the sun and moon, and the fact that the Earth is not a perfect sphere. Also there's network maintence costs, and billing costs.
You could get a Motorola L7089, which works on all 3 GSM bands, and you can interface to with IR... Pretty sure you can get a global charger. You're going to be in pain when you see the cell phone bills though; if you stay in any one place long enough, you're advised to buy the local disposable SIMs. (eg Orange SIMs or Vodaphone SIMs in UK,...)
You still have worries (or at least I hope you do) of software going bonkers and such. Especially with an unmanned vehicle. Evil Linux hackers might h@x0r your blimp and run it into Air Force One. Seriously though, I'm not even comfortable with 100% computer controlled cars, much less aircraft.
As for covering the world, if you're ever in the middle of the ocean, you'll know why you need to cover the world with high availability. Of course, there are GEOs for such situations, but...
As far as monopoly goes, cellular (another McCaw thing) was expensive, but then competition came. Since Teledesic/ICO will need to continuously upgrade their system to stay competitive with land-lines, it will presumably be possible to
Why would you want the birds brought to earth safely? It's so much cheaper to just deorbit them and let them burn up in the atmosphere. Provided that the orbital people designed the deorbits properly, it's not difficult to get something the size of a VW bug to burn up in the atmosphere (BTW, do you have a public source for that size?)
I think you're missing the value proposition of LEOs though... they cover the low-population-density areas of the planet for zero marginal cost. Sell some bandwidth in high-population areas, make a killing off of those not covered by GSM/UMTS...
If satellites|cable|fiber|your-choice-of- technology-to-reach-the-planet is so good, why don't YOU get the funding and do it yourself? The blimp idea is viable, it's just that the FAA would hate you, the FCC would be a pain, you'd only be able to cover metro markets, and your main market is mobile users, who can't roam without a massive blimp network. The point I'm trying to make is that any kind of reach the planet solution is going to cost big bucks and have tradeoffs, and to rip one strategy without seeing whether or not it affects other strategies is uninformed at best.
BTW sparing for blimps is a heck of a lot easier than sparing for satellite constellations because you only need one spare which can be flown to the city with the failure, as opposed to one or two spares _PER PLANE_ (ICO is only two planes, but that means twice the sparing requirement) After all, fedex next-flight-out latency is a lot better than scheduling a launch latency.
Why does no one use it? cuz it's toxic and hard to store. So it's not always about pursuing the theoretical best. liquid oxygen and hydrogen is a good trade in terms of specific impulse for a shuttle-type launch.
I would question NASA's rockets costing an order of magnitude more than Russia's. Sure, Shuttle is expensive, but even with all the extra overhead in that over a proton launch, it's not an order of magnitude more...
Slashdot Mirror
You're assuming two things: perfect filters and infinite bit depth. As I argued in my previous post, higher bit depth is a Good Thing. Better filters are left as an exercise to the high-end CD player manufacturers.
More frequency range isn't going to be recorded, played, or heard by anyone.
First of all, things above 22kHz aren't picked up by ordinary mics... Even the ultra-high-end Neumann U87Ai only claims 20-20kHz frequency response (http://www.neumann.com/mics/u87ai.htm)
Secondly, most speakers won't crank out those high frequencies without a severe falloff in response: the high-end Genelec 1038A triamped monitor gets you 33-20k Hz (-3dB). (http://www.genelec.com/products/1038a/1038a.htm)
Finally, most people can't hear above 20kHz, especially those people who are incessantly blasting their ears out with loud music.
The best reason for Super CD (or DVD or whatever) is higher bit depth, NOT higher sampling rate; going from 16/44.1 (CD quality) to 24/44.1 takes just 50% more space, for nontrivially better quality, while going from 16/44.1 to 16/88.2 brings minimal benefit at a 100% space penalty.
NP does not stand for NON polynomial, it stands for nondeterministic polynomial.
Every problem in P is in NP.
Decryption brute force by definition could not be accomplished in polytime; the problem is if you can always guess right in polytime, then it's a different story (essentially how a nondeterministic turing machine works).
</rant
doesn't matter, since graph edges must be polynomial (actually O(n^2)) in the number of vertices.
You're being paranoid; there are too many amateurs in audio and video for analog to be phased out/made illegal. Quite frankly, if the semipro audio equipment was killed (and any semipro equipment has to allow copies either in digital or analog domain), the RIAA would be killing itself because it would have no way to listen to demo tapes of up and coming artists.
Besides, can you come up with a single piece of legacy equipment which was declared illegal by a US court? Even scanners that can pick up cell phone bands were grandfathered in when that ban was put in place.
Analog audio will always exist because it's easy to build. I can build a mic and speakers from raw components, and digitizing it can be done by any number of off-the-shelf parts, which will need to continue to exist in order for manufacturers to create microphones.
Frankly I don't care that much about video; I already have my APEX...
too expensive to put the amp/decoder in each speaker... there's a reason people use passive monitors (I don't think that's a good thing, but if you're dealing with sound reinforcement, you pretty much either shell out the bucks or go passive).
...
hacking into the cabinets isn't the only way to do it... my trusty MD player or WAV recorder can be hooked up to a pretty high quality mic,
legacy will always be a problem. just like you can burn the IBM CD and then rip it to mp3.
eq isn't necessarily going to get rid of your watermarks, especially if I use some kind of low-level, spread spectrum approach.
On the other hand, cassette recorders and MD recorders exist too, so it's not clear to me why watermarking mattters.
Also, a silly plug, we (the monarch group) haven't updated our latest internet draft because we've been busy writing working code =) Another interesting protocol is DSR: http://www.monarch.cs.cmu.edu/ietf.html
I'm a wireless nets guy, so I'll talk about your radio points. As for (1), you use some CSMA/CA protocol to share a channel among many users. You could even use TDMA to get some guarantees about latency and bandwidth (subject to an error-free channel)
(2): if you're CDMA, it's not easy. Also, it's kinda illegal, so if people actually used these things, the FCC people with their scopes will come knocking at your door.
(3): end to end encryption. enough said.
Even if you encrypt in the reader and decrypt in the software, someone can reverse engineer the software and release the keys, as was done to DVD's CSS. If you let software do some descrambling, and give that software to anyone, even as object code, you're already doomed; you may as well not scramble in the first place.
Bah, I disagree about English and German being close enough to consider "dialects;" as a mandarin speaker, I understand a nontrivial amount of Szechuan dialect (without having to learn any), and almost all phrases translate word-for-word. The written forms are the same as well. This is true for Taiwanese and Cantonese as well, AFAIK.
Typing Chinese turns out to be easier than writing Chinese and only marginally harder than reading Chinese if you know any of the standard phonetic alphabets.
It's not particularly complicated either; generally the only strange part is having to look at the screen to pick your character. But OTOH if you used zhuyin to type the previous sentence, you're going to type 20-25 characters, or about 40-75 keystrokes. Besides, if you've ever gone to an internet cafe in Taiwan you'll see people who type Chinese pretty fast.
So saying that you have to learn quite a lot to type chinese is like saying you need to learn quite a lot to type english... you need to learn the alphabet for english (you need to learn an alphabet for Chinese).
FreeBSD network stack is similar to other UN*Xs (much more so than Linux at the kernel level) so changes are easier to port to other OSes. One reason I use FreeBSD, at least.
Music is an expensive hobby. Some of us can pay for it because we have a day job, some of us have to depend on it for a living. Do you think people can afford to put together $100k studios for high-quality pristene sound and spend $1k/hr for recording if music is free? How about software? Software is relatively cheap to make, except for labor. Should all software be free?
10GHz? Don't you have serious propogation problems at that frequency? No one even does 5.8 GHz ISM in spite of the fact that you can put out 4W over a section of that band for two reasons: cost of electronics handling 5.8GHz, and reduced range per watt in that frequency band (vs 2.4GHz)
Realistically, you have to trade bitrate for power, or range for power, or range for bitrate. You can't have it all.
RSA equivalent to factoring (*) If I recall correctly, I took a crypto class where they proved that if you could get nontrivial advantage on the high log_2 n bits of an arbitrary message encrypted using an n-bit RSA key in polytime, you could factor in polytime. (that is, finding the high 10 bits of a message encrypted with RSA is roughly equivalent to factoring the key... so you can use 10 bits per 1024 bit block, and be pretty secure.)
Yes, so even at $6 billion, the finance costs are ridiculous. But you're covering the whole planet. I don't know about your ballpark sales/marketing/incidentals numbers, those should be mostly the responsibility of the people who resell iridium. As I mentioned in my earlier post, O&M alone is killing Iridium. But that's not because of the market; partially Iridium should price discriminate against oceanic markets (where Inmarsat is it's only serious competitor) and slash prices over land (compete with analog roaming in the US, for example)... if they had designed in a less capacity for more power trade, they could sell competitively sized handsets, and they'd have a nice niche market. Besides, DoD is a nice customer; they doubtless have their own LEOs, but don't want to use them in anything but the most dire circumstances.
But all (I think) Iridium phones take SIM-Cards and roam on GSM networks... the point is to have a system that works when you're out of range of GSM networks. Of course, you have to be outdoors and all that, but...
So GSM doesn't cover oceans and such, and Iridium could have made a great market out of that, but I haven't seen that side of it's marketing. (In fact, given the alternatives in that space, Iridium is small and more convienient).
Even if someone got rid of all of Iridium's debt and gave the constellation to you, you'd still lose money on the O&M (operations and maintenence) agreement with Motorola. MOT's been running the constellation free of charge so that IRID could attempt to make debt payments, but that didn't quite work, but even without debt payments, income < O&M costs, so.... you'd have to get a better O&M deal with MOT to even break even.
Basically what running the constellation means is making sure that the satellites don't fall out of their orbits due to perturbations caused by the sun and moon, and the fact that the Earth is not a perfect sphere. Also there's network maintence costs, and billing costs.
You could get a Motorola L7089, which works on all 3 GSM bands, and you can interface to with IR... Pretty sure you can get a global charger. You're going to be in pain when you see the cell phone bills though; if you stay in any one place long enough, you're advised to buy the local disposable SIMs. (eg Orange SIMs or Vodaphone SIMs in UK, ...)
You still have worries (or at least I hope you do) of software going bonkers and such. Especially with an unmanned vehicle. Evil Linux hackers might h@x0r your blimp and run it into Air Force One. Seriously though, I'm not even comfortable with 100% computer controlled cars, much less aircraft.
As for covering the world, if you're ever in the middle of the ocean, you'll know why you need to cover the world with high availability. Of course, there are GEOs for such situations, but...
As far as monopoly goes, cellular (another McCaw thing) was expensive, but then competition came. Since Teledesic/ICO will need to continuously upgrade their system to stay competitive with land-lines, it will presumably be possible to
Why would you want the birds brought to earth safely? It's so much cheaper to just deorbit them and let them burn up in the atmosphere. Provided that the orbital people designed the deorbits properly, it's not difficult to get something the size of a VW bug to burn up in the atmosphere (BTW, do you have a public source for that size?)
I think you're missing the value proposition of LEOs though... they cover the low-population-density areas of the planet for zero marginal cost. Sell some bandwidth in high-population areas, make a killing off of those not covered by GSM/UMTS...
If satellites|cable|fiber|your-choice-of- technology-to-reach-the-planet is so good, why don't YOU get the funding and do it yourself? The blimp idea is viable, it's just that the FAA would hate you, the FCC would be a pain, you'd only be able to cover metro markets, and your main market is mobile users, who can't roam without a massive blimp network. The point I'm trying to make is that any kind of reach the planet solution is going to cost big bucks and have tradeoffs, and to rip one strategy without seeing whether or not it affects other strategies is uninformed at best.
BTW sparing for blimps is a heck of a lot easier than sparing for satellite constellations because you only need one spare which can be flown to the city with the failure, as opposed to one or two spares _PER PLANE_ (ICO is only two planes, but that means twice the sparing requirement) After all, fedex next-flight-out latency is a lot better than scheduling a launch latency.
What's the best fuel for a rocket?
flourine and hydrogen.
Why does no one use it? cuz it's toxic and hard to store. So it's not always about pursuing the theoretical best. liquid oxygen and hydrogen is a good trade in terms of specific impulse for a shuttle-type launch.
I would question NASA's rockets costing an order of magnitude more than Russia's. Sure, Shuttle is expensive, but even with all the extra overhead in that over a proton launch, it's not an order of magnitude more...