NASA were not over-optimistic. They were outright lying. Recent evidence reveals that the development of the shuttle wasn't really over budget - it cost about as much as internal NASA estimates said it would. It did cost quite a bit more than what they promised to the congress, though. In other words - they lied so the project wouldn't get cancelled and relied on getting more money down the road.
To bring back these launch vehicles safely to earth, we have refine the science of aerodynamics and special materials," he said." ... For orbital missions, as the vehicle needs to deliver Mach number of the order of 25 (25 times the speed of sound), it becomes mandatory that rocket-based systems have to be combined with air breathing systems leading to what is termed as Rocket Based Combined Cycle System, for meeting the total orbital velocity requirement.
This attitude is exactly what makes NASA stuff so expensive. There are credible space experts who believe a cost reduction by a factor 10 to 100 is possible with proper application of nothing but existing materials, engineering and aerodynamic designs (and in particular no heavy airbreathing engines). In fact, use of mature technology and not confusing research vehicles with operational vehicles is a prerequisite for any significant cost reductions.
If this NASAThink has infected the ISRO I'm afraid the only cost reduction they will be capable of is because of somewhat lower labor costs.
12,000 seconds Isp? Sheesh. You may need to see a therapist about your specific impulse fetish.
Once you get over ~3000 seconds Isp you don't really need to keep improving it. Who cares if your propellant fraction is 15% or 20% ? As long as it's not over 60% (as is often the case with chemical propulsion) you are doing fine.
Most space probe engineers would gladly trade lower Isp for higher thrust so they don't get too old by the time the vehicle finishes accelerating. Higher *energy* efficiency would probably be welcome, too.
I hope there's a foot of lead included in that shielding somewhere. To me that would seem the most vital shielding they could provide.
There's nothing magical about lead. It's more dense so it takes up less volume than other shielding but sufficient mass of almost any material also work as radiation shielding. A few meters of concrete or frozen earth will do nicely, thank you.
This new tunneling mode requires upgrading both the client and server. It should be possible to get more-or-less the same functionality using only client-side support: capture the packets sent to the tun interface, decode individual TCP streams (similar to slirp) and convert them to port forwarding requests compatible with old servers.
If you want to use those "repressed memories" as legal evidence - absolutely not a good idea.
When used as therapy it's ok. As long as the client thinks that he's uncovered the reason for his problems and it results in improvement who cares if these memories are 100% accurate? Of course, this needs to be done responsibly. You don't want someone believing they have been sexually abused by someone who is innocent.
Instead of reporting specific incidents where a vehicle exceeds a speed limit collect statistical data and hand it over to the insurance companies.
Owners of vehicles which consistently drive faster than the traffic around them (regardless of the speed limit) or consistently keep a shorter than average distance from the vehicle ahead will have to pay more because they are more likely to be involved in accidents.
While the semantic web people are arguing forever about ontologies and schemas Google go ahead and implement a practical way of adding meaningful metadata which real people can actually use.
Instead of having strict schemas which will never be quite right you can just add whatever attribute you like and see which attributes are popular for the type of entity you are entering.
Not appropriate for all types of workloads
on
New Server Chip Niagara
·
· Score: 4, Insightful
You need at least a dozen concurrent threads or processes before you can make good use of this CPU's power. Certainly not a good idea for desktops. An excellent match for web servers. Other server-type workloads (e.g. database, file server) may need some tuning to make the best of this architecture.
VMSK was just phase modulation with very small modulation index. While the bandwidth between the -3db points is indeed very narrow (apparently breaking the Shannon limit) the actual information was carried in the sidebands. This means, of course, that the central carrier is just a waste of transmitted power.
This "xG" modulation seems to be a more subtle variation on the same idea as VMSK: the central carrier is in the 900 mHz band, where you are allowed to transmit quite a lot of power without a license. Unlike VMSK it's not totally useless since it is used as a timing signal to help the receiver synchronize. It is also used as a regulatory fig leaf (see below). The rest of the data is carried on very wide sidebands that leak hundreds of MHz outside the 902-928MHz unlicensed band. The total power of these sidebands is just tens of milliwatts and they are spread over hundreds of MHz so the power density at each frequency is below the level that devices like your PC are allowed to emit as background noise.
So xG is ultrawideband plus a pilot signal in the 900MHz band.
Ultrawideband is real. It can transmit a lot of data at very low power levels. But current FCC rules only allow UWB operation above 3.1GHz (up to 10GHz) and places additional restrictions like limiting it to indoor operation (outdoor operation is allowed for handheld devices, not fixed installations). One problem with these high frequencies is that they do not penetrate walls very well. UWB has been demonstrated below 1GHz and it works extremely well at these lower frequencies - but that's not allowed by part 15 of the FCC rules.
So how can xG legally transmit UWB in the coveted first GHz of the spectrum? By twisting the rules a bit. A device is allowed to emit low levels of RF energy all over the spectrum as noise (like your PC) but it's not allowed to transmit in these frequencies an an "intentional radiator". But there's another way in which a device may transmit small amounts of power in an otherwise restricted band: as sidebands of a carrier which is centered inside a legitimate band. This is what xG is doing - they claim that this wideband signal is the sidebands of the central carrier in the unlicensed 900MHz band. And it really is. That's what you get when you modulate a 900MHz signal with sub-cycle pulses.
So xG may meet the letter of the law but probably not the intent.
...all of the scientific community is studying data covering a half million years to try and figure out how big the CO2 effect is.
That is exactly the problem - they are trying to find "how big the CO2 effect is", in other words, they are assuming it exists. They should be treating it like any other hypothesis, not as some holy grail than needs to be "proven" at any cost. I don't know if the effect exists or not. Many respected scientists agree that the data is not conclusive. Unfortunately, the average guy on the street already "knows" because he has been brainwashed.
This won't be cryptographically safe until the data held on the card is not directly readable. So magstripe cards are insufficient for this. All it needs is for the user to have spyware installed which snoops the data from the magstripe card next time it's scanned.
How can spyware snoop data that only passses through the computer already encrypted? This device will NOT be available to the computer as a general-purpose magstripe reader.
It may not be "cryptographically safe" in the general sense but it would require someone to build a fake reader and put it in your house instead of the one you got from the bank or gain access to your card AND watch you type in your code. Not impossible, but it raises the bar significantly and all of these attacks put the attacker in significant personal risk - not something that can be done anonymously over the net to millions all over the world with little chances of being caught.
There is already system featuring two factor authentication (something you have + something you know) fully deployed and already distributed to millions of bank customers. They keep the token in their wallet and remember the password.
I'm talking about ATM cards.
How about this: a small USB device with a magstripe reader, numeric keypad and a big notice saying "always enter your PIN on this keypad, never on your computer's keyboard".
This device will not verify the PIN number itself - it will just encrypt the magstripe data and entered PIN to the bank's public key and send it to the PC which will forward it to to the bank for authorization.
You are ignoring the fact that all modern d/a converters are oversampling already. Most do at least 4x oversampling to make the reconstruction filter design easier/cheaper/better sounding.
Yes, all modern DACs are oversampling. But the the quality of the anti-aliasing filter used in this oversampling is not good enough. Ideally, the filter should have a "brick wall" reponse, going from 0db attenuation to minus infinity in 0 hz. In practice, these filters are always a compromise: the transition band always has a non-zero width, the attenuation is not infinity and the passband is not 100% flat. The compromise implemented by most of these filters aims for excellent flatness, good stopband attenuation and minimal silicon space - at the cost of a wider transition band. Usually they start the transition at around 20kHz (the -3db point) and achieve the full stopband attentuation of -80db or better at around 24 kHz. Notice that this is ABOVE the 22.05kHz nyquist frequency. At 22.05 many of these DACs get no more than 24db attenuation. They assume that these frequencies are inaudible anyway and allow the transition band to include some near ultrasonic aliases.
Doing a better job requires either a lot of silicon or compromising some other property. None of these options is attractive in a competitive market. They need specs that sound impressive. A DAC that starts attentuating at 19kHz instead of 20kHz may actually sound better, but systems using it won't be able to boast "20Hz to 20kHz" response.
That line about MP3s sounding better at 96kHz is a bunch of marketing BS.
In most marketing bullshit there is usually some hint of truth.
If you take the raw 44.1ksps PCM of the CD (not mp3) and play it through this 96ksps upsampling it may sound somewhat better. The digital to analog converters used on most CD players leave some aliased frequencies above 22.05 kHz. The full 80+db alias suppression of the interpolation low pass filter is usually not achieved until you get to 23 or 24 kHz. You can't hear these frequencies but when they hit nonlinearities in your speakers they get intermodulated with audio content and "folded" down into the first few kHz of the audible band where the ear is most sensitive. The reason this is troublesome is because content at these frequencies is not random noise - it's correlated to the audio signal and nonlinear intermodulation results in distortion.
This type of alias intermodulation distortion has been discovered relatively recentlu and many audio engineers are not yet aware of it. The "compromise" lowpass filter on DACs by itself is ok, the nonlinearities in speakers are ok when the audio source does not have these aliases, but when you combine them you get new audio artifacts.
How to fix it?
If you are mixing content to be played through a standard CD 44.1ksps DAC, make sure that right before you dither down to 16 bits you clean up the 1.5-2kHz region just below the nyquist frequency. Apply a really long and high quality filter (4096+ tap FIR) starting at around 20kHz with a transition band no more than 500Hz wide so you achieve the full suppression well before hitting 22.05. This will ensure that there is nothing to be aliased into the first 1.5-2 kHz above 22.05kHz by the DAC. Note that some dithering/noiseshaping processes that reduce the word length to 16 bits may add significant energy to this band you have just cleaned up but it's not correlated to the rest of the audio signal so it shouldn't be too much of a problem.
If you are playing 44.1ksps content through a DAC running at a higher frequency (and longer word length to avoid rounding errors introduced by the process) you can upsample like Creative does on this card. When it's not doing multichannel 3D audio. their card should have plenty of spare DSP power to do upsampling at much better quality than typical DACs.
In the context of digital signal processing MIPS refers to the number of multiply-accumulate operations per second, including incrementing buffer pointers. It is a well-defined number and comparing it is not meaningless even across different architectures.
...it is dishonest for Level-3 to blame Cogent for this exclusively. Level-3 had a peering arrangement with Cogent for a long time. If you look at Level-3's interconnection policy page:
It still looks like Cogent and Level-3 could peer under these terms. It was Level-3 that pulled the plug, not Cogent.
Quick quote from the introduction to these terms:
It is Level 3's view that IP transit is the rule and that settlement free interconnection or "peering" should be the exception when backbones interconnect.
Seems to me they were pretty up-front about their terms for peering this from the start. I expect their peering contract with Cogent reflects this and allows either party to terminate the agreement at any time.
During forward flight the vehicle generates a significant part of its lift using airfoils. These airfoils are small, but at cruising speed they are up to the task. Remember that normal aircraft use wings sized for low-speed takeoff and landing, not for cruising speed. Low-speed hovering will naturally consume much more fuel but it's only for a short time at takeoff and landing.
Ducted fan vehicles are feasible and should theoretically be capable of doing all Dr Moller claims they can do. It's a different question whether Moller's company can achieve it. People tend to become skeptical after decades of of promises...
Is the closed source code of Windows preventing us from actively defending our systems?
Apparently it isn't an obstacle to the rootkit writers. Why should it be any more of a problem for someone writing a rootkit defense system? Most Windows system programmers are pretty used to soft-icing into windows code by now.
Because of the way in which languages change, it's very unlikely that we'll ever be able to do so, at least if we are observing accepted standards of scientific rigor.
Unfortunately, when studying the past you can't make experiments, reproducible or not. As a result, fields like evolution or linguistic history will always be less scientifically rigorous than other fields.
There is one, and only one, method for determining relatedness between languages which is generally accepted by specialists in the field: namely, by identifying a core of lexical and morphological items which show systematic correspondences in their sounds between languages
There are documented cases of languages gradually adopting most of the vocabulary from another dominant regional language while retaining their original grammatic structure nearly untouched. Why is grammatic structure not valid evidence for the origins of languages? Can't you offer anything better than invoking the old argument from authority?
Slashdot Mirror
NASA were not over-optimistic. They were outright lying. Recent evidence reveals that the development of the shuttle wasn't really over budget - it cost about as much as internal NASA estimates said it would. It did cost quite a bit more than what they promised to the congress, though. In other words - they lied so the project wouldn't get cancelled and relied on getting more money down the road.
If this NASAThink has infected the ISRO I'm afraid the only cost reduction they will be capable of is because of somewhat lower labor costs.
12,000 seconds Isp? Sheesh. You may need to see a therapist about your specific impulse fetish.
Once you get over ~3000 seconds Isp you don't really need to keep improving it. Who cares if your propellant fraction is 15% or 20% ? As long as it's not over 60% (as is often the case with chemical propulsion) you are doing fine.
Most space probe engineers would gladly trade lower Isp for higher thrust so they don't get too old by the time the vehicle finishes accelerating. Higher *energy* efficiency would probably be welcome, too.
I hope there's a foot of lead included in that shielding somewhere. To me that would seem the most vital shielding they could provide.
There's nothing magical about lead. It's more dense so it takes up less volume than other shielding but sufficient mass of almost any material also work as radiation shielding. A few meters of concrete or frozen earth will do nicely, thank you.
This new tunneling mode requires upgrading both the client and server. It should be possible to get more-or-less the same functionality using only client-side support: capture the packets sent to the tun interface, decode individual TCP streams (similar to slirp) and convert them to port forwarding requests compatible with old servers.
So what has he achieved?
If you assume nothing but a huge software conglomerate is "achievement" then I guess he has not achieved much.
His job satisfaction must beat the hell out of most of us in the software industry. I'd call that success.
Ouch! I know you don't like him too much, but this is just cruel.
Here is the original article by Amit Raz et al, published in 2002.
If you want to use those "repressed memories" as legal evidence - absolutely not a good idea.
When used as therapy it's ok. As long as the client thinks that he's uncovered the reason for his problems and it results in improvement who cares if these memories are 100% accurate? Of course, this needs to be done responsibly. You don't want someone believing they have been sexually abused by someone who is innocent.
Instead of reporting specific incidents where a vehicle exceeds a speed limit collect statistical data and hand it over to the insurance companies.
Owners of vehicles which consistently drive faster than the traffic around them (regardless of the speed limit) or consistently keep a shorter than average distance from the vehicle ahead will have to pay more because they are more likely to be involved in accidents.
While the semantic web people are arguing forever about ontologies and schemas Google go ahead and implement a practical way of adding meaningful metadata which real people can actually use.
Instead of having strict schemas which will never be quite right you can just add whatever attribute you like and see which attributes are popular for the type of entity you are entering.
You need at least a dozen concurrent threads or processes before you can make good use of this CPU's power. Certainly not a good idea for desktops. An excellent match for web servers. Other server-type workloads (e.g. database, file server) may need some tuning to make the best of this architecture.
VMSK was just phase modulation with very small modulation index. While the bandwidth between the -3db points is indeed very narrow (apparently breaking the Shannon limit) the actual information was carried in the sidebands. This means, of course, that the central carrier is just a waste of transmitted power.
This "xG" modulation seems to be a more subtle variation on the same idea as VMSK: the central carrier is in the 900 mHz band, where you are allowed to transmit quite a lot of power without a license. Unlike VMSK it's not totally useless since it is used as a timing signal to help the receiver synchronize. It is also used as a regulatory fig leaf (see below). The rest of the data is carried on very wide sidebands that leak hundreds of MHz outside the 902-928MHz unlicensed band. The total power of these sidebands is just tens of milliwatts and they are spread over hundreds of MHz so the power density at each frequency is below the level that devices like your PC are allowed to emit as background noise.
So xG is ultrawideband plus a pilot signal in the 900MHz band.
Ultrawideband is real. It can transmit a lot of data at very low power levels. But current FCC rules only allow UWB operation above 3.1GHz (up to 10GHz) and places additional restrictions like limiting it to indoor operation (outdoor operation is allowed for handheld devices, not fixed installations). One problem with these high frequencies is that they do not penetrate walls very well. UWB has been demonstrated below 1GHz and it works extremely well at these lower frequencies - but that's not allowed by part 15 of the FCC rules.
So how can xG legally transmit UWB in the coveted first GHz of the spectrum? By twisting the rules a bit. A device is allowed to emit low levels of RF energy all over the spectrum as noise (like your PC) but it's not allowed to transmit in these frequencies an an "intentional radiator". But there's another way in which a device may transmit small amounts of power in an otherwise restricted band: as sidebands of a carrier which is centered inside a legitimate band. This is what xG is doing - they claim that this wideband signal is the sidebands of the central carrier in the unlicensed 900MHz band. And it really is. That's what you get when you modulate a 900MHz signal with sub-cycle pulses.
So xG may meet the letter of the law but probably not the intent.
...all of the scientific community is studying data covering a half million years to try and figure out how big the CO2 effect is.
That is exactly the problem - they are trying to find "how big the CO2 effect is", in other words, they are assuming it exists. They should be treating it like any other hypothesis, not as some holy grail than needs to be "proven" at any cost. I don't know if the effect exists or not. Many respected scientists agree that the data is not conclusive. Unfortunately, the average guy on the street already "knows" because he has been brainwashed.
This won't be cryptographically safe until the data held on the card is not directly readable. So magstripe cards are insufficient for this. All it needs is for the user to have spyware installed which snoops the data from the magstripe card next time it's scanned.
How can spyware snoop data that only passses through the computer already encrypted? This device will NOT be available to the computer as a general-purpose magstripe reader.
It may not be "cryptographically safe" in the general sense but it would require someone to build a fake reader and put it in your house instead of the one you got from the bank or gain access to your card AND watch you type in your code. Not impossible, but it raises the bar significantly and all of these attacks put the attacker in significant personal risk - not something that can be done anonymously over the net to millions all over the world with little chances of being caught.
There is already system featuring two factor authentication (something you have + something you know) fully deployed and already distributed to millions of bank customers. They keep the token in their wallet and remember the password.
I'm talking about ATM cards.
How about this: a small USB device with a magstripe reader, numeric keypad and a big notice saying "always enter your PIN on this keypad, never on your computer's keyboard".
This device will not verify the PIN number itself - it will just encrypt the magstripe data and entered PIN to the bank's public key and send it to the PC which will forward it to to the bank for authorization.
You are ignoring the fact that all modern d/a converters are oversampling already. Most do at least 4x oversampling to make the reconstruction filter design easier/cheaper/better sounding.
Yes, all modern DACs are oversampling. But the the quality of the anti-aliasing filter used in this oversampling is not good enough. Ideally, the filter should have a "brick wall" reponse, going from 0db attenuation to minus infinity in 0 hz. In practice, these filters are always a compromise: the transition band always has a non-zero width, the attenuation is not infinity and the passband is not 100% flat. The compromise implemented by most of these filters aims for excellent flatness, good stopband attenuation and minimal silicon space - at the cost of a wider transition band. Usually they start the transition at around 20kHz (the -3db point) and achieve the full stopband attentuation of -80db or better at around 24 kHz. Notice that this is ABOVE the 22.05kHz nyquist frequency. At 22.05 many of these DACs get no more than 24db attenuation. They assume that these frequencies are inaudible anyway and allow the transition band to include some near ultrasonic aliases.
Doing a better job requires either a lot of silicon or compromising some other property. None of these options is attractive in a competitive market. They need specs that sound impressive. A DAC that starts attentuating at 19kHz instead of 20kHz may actually sound better, but systems using it won't be able to boast "20Hz to 20kHz" response.
Most audio encoded by MP3 or other lossy codecs doesn't encode anything the top band at all so this process will not improve anything in that case.
That line about MP3s sounding better at 96kHz is a bunch of marketing BS.
In most marketing bullshit there is usually some hint of truth.
If you take the raw 44.1ksps PCM of the CD (not mp3) and play it through this 96ksps upsampling it may sound somewhat better. The digital to analog converters used on most CD players leave some aliased frequencies above 22.05 kHz. The full 80+db alias suppression of the interpolation low pass filter is usually not achieved until you get to 23 or 24 kHz. You can't hear these frequencies but when they hit nonlinearities in your speakers they get intermodulated with audio content and "folded" down into the first few kHz of the audible band where the ear is most sensitive. The reason this is troublesome is because content at these frequencies is not random noise - it's correlated to the audio signal and nonlinear intermodulation results in distortion.
This type of alias intermodulation distortion has been discovered relatively recentlu and many audio engineers are not yet aware of it. The "compromise" lowpass filter on DACs by itself is ok, the nonlinearities in speakers are ok when the audio source does not have these aliases, but when you combine them you get new audio artifacts.
How to fix it?
If you are mixing content to be played through a standard CD 44.1ksps DAC, make sure that right before you dither down to 16 bits you clean up the 1.5-2kHz region just below the nyquist frequency. Apply a really long and high quality filter (4096+ tap FIR) starting at around 20kHz with a transition band no more than 500Hz wide so you achieve the full suppression well before hitting 22.05. This will ensure that there is nothing to be aliased into the first 1.5-2 kHz above 22.05kHz by the DAC. Note that some dithering/noiseshaping processes that reduce the word length to 16 bits may add significant energy to this band you have just cleaned up but it's not correlated to the rest of the audio signal so it shouldn't be too much of a problem.
If you are playing 44.1ksps content through a DAC running at a higher frequency (and longer word length to avoid rounding errors introduced by the process) you can upsample like Creative does on this card. When it's not doing multichannel 3D audio. their card should have plenty of spare DSP power to do upsampling at much better quality than typical DACs.
In the context of digital signal processing MIPS refers to the number of multiply-accumulate operations per second, including incrementing buffer pointers. It is a well-defined number and comparing it is not meaningless even across different architectures.
...it is dishonest for Level-3 to blame Cogent for this exclusively. Level-3 had a peering arrangement with Cogent for a long time. If you look at Level-3's interconnection policy page:
Seems to me they were pretty up-front about their terms for peering this from the start. I expect their peering contract with Cogent reflects this and allows either party to terminate the agreement at any time.http://www.level3.com/1511.html
It still looks like Cogent and Level-3 could peer under these terms. It was Level-3 that pulled the plug, not Cogent.
Quick quote from the introduction to these terms:
During forward flight the vehicle generates a significant part of its lift using airfoils. These airfoils are small, but at cruising speed they are up to the task. Remember that normal aircraft use wings sized for low-speed takeoff and landing, not for cruising speed. Low-speed hovering will naturally consume much more fuel but it's only for a short time at takeoff and landing.
Ducted fan vehicles are feasible and should theoretically be capable of doing all Dr Moller claims they can do. It's a different question whether Moller's company can achieve it. People tend to become skeptical after decades of of promises...
Is the closed source code of Windows preventing us from actively defending our systems?
Apparently it isn't an obstacle to the rootkit writers. Why should it be any more of a problem for someone writing a rootkit defense system? Most Windows system programmers are pretty used to soft-icing into windows code by now.
Because of the way in which languages change, it's very unlikely that we'll ever be able to do so, at least if we are observing accepted standards of scientific rigor.
Unfortunately, when studying the past you can't make experiments, reproducible or not. As a result, fields like evolution or linguistic history will always be less scientifically rigorous than other fields.
There is one, and only one, method for determining relatedness between languages which is generally accepted by specialists in the field: namely, by identifying a core of lexical and morphological items which show systematic correspondences in their sounds between languages
There are documented cases of languages gradually adopting most of the vocabulary from another dominant regional language while retaining their original grammatic structure nearly untouched. Why is grammatic structure not valid evidence for the origins of languages? Can't you offer anything better than invoking the old argument from authority?