>> Patterson even suspects that finite state machines are inherently >> serial and CANNOT be parallelized.
That's because each state of a DFA already represents a (possibly very large) set of states in an NFA. Exploding it back into an NFA would be an anti-optimization. Patterson did not touch on this subtlety.
FA's derived from regex's go through this transform automatically (via grep or a parser generator for example). FA's from pretty flowcharts designed by humans are just the same level of monothink as anything else they do and probably belong in one of the other categories.
The 37W3 is about the cheapest 1080p LCD you can get, so one wonders if westinghouse (or more specifically, whatever chinese company actually built it) just cut corners left and right. You buy cheap stuff, you have to expect some problems.
>> By only using a 2.5" drive rather than 3.5 of course the average >> seek time is lower, because the read head doesn't have the >> extra 1" to cover.
it's even more trivial than you paint. The 2.5 and 3.5 numbers represent diameter, but the head only travels on one side of the disk so to it the difference is only 0.5 inch as far as it is concerned.
>> We just have to be careful that while we enslave the algae, they >> don't know it's happening so they don't start an uprising.
Obviously, one would construct a virtual reality to keep them occupied. At first, you might try to construct a virtual paradise but eventually they would get suspicious and revolt. So the second virtual reality would be more like they are used to, but maybe there would be one algae, let's call him Geo, who can feel that this virtual algae reality isn't quite right. Eventually he will face the decision of whether to take the green pill or the dark green pill. His choice will determine whether he will be able to fight against his evil human-being overlords and free his people to have some sort of weird algae sex-orgy in a cave or something.
Just to clarify, drive-by-wire is not a requirement for a/c clutch disengagement. Anything with EFI (i.e. anything sold in the US since 1986 or later) has enough data to determine full throttle. (Pre-EFI, lots of cars had pressure switches either near the gas pedal or on the carburetor.) Basically, it's a matter of cost--the extra switches, relays and/or wiring probably adds $5 to manufacturing cost and as most people wouldn't notice it anyway it got left out a lot.
I am an American so I am perhaps blinded by my proximity to the problem, but as near as I can figure...
* SUV's are the only category of vehicle that is available with 7+ passenger seating, aside from minivans. There are whole swaths of suburbia where mommies and daddies cart their kids and their kids' friends around every weekend. A 4 passenger econobox is simply not adequate for this task and I absolutely guarantee that this is part of the purchase decision. (Minivans, as an alternative, are not much more fuel efficient than SUVs.)
* There is a big safety hangover from when someone published the obvious conclusion that a heavier car will suffer less damage in a crash than a smaller car. Since then, very few people will buy anything smaller than a Honda Civic or Toyota Corolla and the rest will buy the biggest vehicle they can afford.
* The big motors are there because we have a lot of hills (in some parts of the US) and a *lot* of stoplights. The big motors supply a lot of torque and they accelerate without fuss. A small motor will of course suffice but they can be noisy and at times require work on the part of the driver to select the correct gear. (Yanks like auto transissions, don't forget)
* A lot of us Americans have owned and driven small cars, especially during high school and college years. Generally, these are cheap used cars that, frankly, suck. The first thing everyone does when they get a decent job is to buy a new car, and they always get a larger, more powerful one to erase the bad memories of the datsun or paseo or whatever it was.
Anyway, my point is that it's not entirely about the big american wasteful image. There are some practical concerns that weigh into the choices Americans make re which cars to buy and if someone wants to get us to use more fuel efficient vehicles, they need to address these concerns rather than just insulting us for our choices.
it is not unheard of for the a/c clutch to disengage when the driver commands full throttle, but there's certainly no fuel economy involved there.
other item is climate control... (a) it can switch off the compressor whenever it feels like (b) which is usually never (c) the EPA should take this into account with its tests--cars with automatic climate control should be in the default automatic mode throughout the test since that's how people drive them
>> Buy a rubbermaid container and put any liquids you're leaving >> at the house in it- shampoo, 409, drain cleaner, liquid soap, >> batteries, whatever.
Heh, you left bleach off the list:-)
Seriously, it's a bad idea. Chemicals can leak out of their container and react with each other, producing flames or toxic gases.
The idea is that one well-built power supply, possible incorporated into a normal outlet would be more efficient than multiple individual transformers, aka wall warts.
It will cost less because less idle current will be needed.
It *might* be more convenient if it is designed properly (a lot of people find the power bricks cumbersome and annoying to travel with, for example.)
It isn't flexible in and of itself, but if the voltages are standardized the user will have more interchangeability regarding cables and chargers.
Wall warts only use about a watt when idle, but keep in mind there are maybe a billion of them in the world. A billion watts 24/7 is a lot of wasted power. If we can reduce that by gradually replacing them with a more efficent solution, it would probably be a good idea overall, even if you personally don't find it convenient.
The thinking is that any black holes that are created by the LHC would be so small that they would evaporate in an instant, probably within milliseconds of devouring the earth and sun. So there's nothing to worry about really.
The idea is that a common DC circuit would eliminate 5 or 10 or more wall warts. Also, a dedicated device could be designed with more sophisticated circuitry to drop its idle current down to nearly zero.
There are a few complaints about CD's that are irrespective of the nyqyest limit.
1. the first couple years of recorded CD's did not contain proper dither (a small amount of deliberately added noise). This is not a big deal but it meant that the theoretical noise floor was off by a few decibels. Nobody heard this because:
2. The first generation of CD players had vary gradual filters that kicked in around 12khz and killed everything about 15khz.
3. The second generation of players had some extra logic that figured out if the recording contained proper dither and only turned on the heavy filter when it was missing, otherwise you got the "2X oversampling" feature which helped a lot: the analog filter didn't have to kick in until 19khz or so.
4. The first several zillion CD's released were "remastered" from tapes. I.e. they just played the analog tape into a digital mastering board and left it at that. All the hiss, snap/crackle/pop, fret noise, sibilance, guy sneezing in 3rd row, all of it made it onto the CD perfectly preserved. In my opinion it wasn't until the late nineties when remastered versions of analog recordings were actually better than vinyl.
5. CD players have to spin the damn disk and most cheap-o CD players have little 0.00001 horsepower motors that spin at a 5 thousand rpm and they have really annoying physical noise that usually makes it out into the room. I hate that.
6. The first couple generations of CD changers had a bad reputation of breaking with 50 of your CD's inside...
7. My personal gripe: the advent of fully digital recording enabled a lot of no-talent performers to use various digital processors to make them not sound like the screeching walruses that they are, and to replace a lot of good session musicians with drum machines and beat boxes. Damn shame.
Anyway, most of these problems have been fixed over the years, but I think a lot of audiophile snobs were right about the first 2 generations of equipment and recordings on CD. The bits were there and capable of a clean 20khz, but the source material and some engineering mistakes really f'd things up at first.
I've looked at various people's ipods and in every single case the contents were just their CD collection plus some stuff they bought from itunes. No copying.
Plus, ipods don't count as repositories of music--you can only download music to the ipod, you can't upload it back out so any unauthorized copy is a dead-end once it gets to the ipod. It's not in the same category as Kazaa or bittorrent.
I think that what you say indeed helped the early adopters, but DVD player prices fell so fast (in the US at least) that for the most part, people here just bought PS2's to play games. I'm sure there are exceptions, but that's my observation.
I agree with your sentiment, and additionally things like fingerprints and retinal scans cannot be re-issued if compromised. This isn't a problem yet, but as biometric tokens are more widely used and thus more widely attacked it will become a problem.
PS1's played audio CD's and PS2's played regular DVDs but hardly anyone used them to do that. If sony is counting on blue-ray to be different in that regard, they may be deluding themselves.
Point taken. It might be more reliable to have a "fuzzy" frame fingerprint that kind of works but gets lots of false positives. Then check every frame in the commercial. If a whole bunch match and there is correlation in the timecodes then you've ID'd the commercial.
Cool thing is, you could set your DVR to show you new commercials you haven't seen before. I don't think I'm alone--what I hate about commercials isn't the content, it's the repetition. In that sense, so long as the fingerprint is generated from the same analog feed each time, some of the variation goes away and the fingerprinting works better.
Might also be interesting to try to flag entry/exit frames from known commercials (with a checksum or fingerprint sort of thing) into a shared database, then just look up the frames. If your show has 2.5 minutes with matches every 30 seconds, that's a commercial break. People can do a P2P vote with their FF button to contribute. Just a thought.
Slashdot Mirror
>> Patterson even suspects that finite state machines are inherently
>> serial and CANNOT be parallelized.
That's because each state of a DFA already represents a (possibly very large) set of states in an NFA. Exploding it back into an NFA would be an anti-optimization. Patterson did not touch on this subtlety.
FA's derived from regex's go through this transform automatically (via grep or a parser generator for example). FA's from pretty flowcharts designed by humans are just the same level of monothink as anything else they do and probably belong in one of the other categories.
It used to be Theater High Altitude Area Defense.
Yeah, li-ion batteries never have safety issues.
</sarcasm>
>> ultracapacitors.. we had supercapacitors till now..
>> whats next.. ubercapacitors? ubersuperultracapacitors..
googlecapacitors!
The 37W3 is about the cheapest 1080p LCD you can get, so one wonders if westinghouse (or more specifically, whatever chinese company actually built it) just cut corners left and right. You buy cheap stuff, you have to expect some problems.
Does anyone have any theories on the point of having the panels blurred out in some of the images (/156 and /151 from the above list)?
>> By only using a 2.5" drive rather than 3.5 of course the average
>> seek time is lower, because the read head doesn't have the
>> extra 1" to cover.
it's even more trivial than you paint. The 2.5 and 3.5 numbers
represent diameter, but the head only travels on one side of
the disk so to it the difference is only 0.5 inch as far as it
is concerned.
>> We just have to be careful that while we enslave the algae, they
>> don't know it's happening so they don't start an uprising.
Obviously, one would construct a virtual reality to keep them occupied. At first, you might try to construct a virtual paradise but eventually they would get suspicious and revolt. So the second virtual reality would be more like they are used to, but maybe there would be one algae, let's call him Geo, who can feel that this virtual algae reality isn't quite right. Eventually he will face the decision of whether to take the green pill or the dark green pill. His choice will determine whether he will be able to fight against his evil human-being overlords and free his people to have some sort of weird algae sex-orgy in a cave or something.
Just to clarify, drive-by-wire is not a requirement for a/c clutch disengagement. Anything with EFI (i.e. anything sold in the US since 1986 or later) has enough data to determine full throttle. (Pre-EFI, lots of cars had pressure switches either near the gas pedal or on the carburetor.) Basically, it's a matter of cost--the extra switches, relays and/or wiring probably adds $5 to manufacturing cost and as most people wouldn't notice it anyway it got left out a lot.
I am an American so I am perhaps blinded by my proximity to the problem, but as near as I can figure...
* SUV's are the only category of vehicle that is available with 7+ passenger seating, aside from minivans. There are whole swaths of suburbia where mommies and daddies cart their kids and their kids' friends around every weekend. A 4 passenger econobox is simply not adequate for this task and I absolutely guarantee that this is part of the purchase decision. (Minivans, as an alternative, are not much more fuel efficient than SUVs.)
* There is a big safety hangover from when someone published the obvious conclusion that a heavier car will suffer less damage in a crash than a smaller car. Since then, very few people will buy anything smaller than a Honda Civic or Toyota Corolla and the rest will buy the biggest vehicle they can afford.
* The big motors are there because we have a lot of hills (in some parts of the US) and a *lot* of stoplights. The big motors supply a lot of torque and they accelerate without fuss. A small motor will of course suffice but they can be noisy and at times require work on the part of the driver to select the correct gear. (Yanks like auto transissions, don't forget)
* A lot of us Americans have owned and driven small cars, especially during high school and college years. Generally, these are cheap used cars that, frankly, suck. The first thing everyone does when they get a decent job is to buy a new car, and they always get a larger, more powerful one to erase the bad memories of the datsun or paseo or whatever it was.
Anyway, my point is that it's not entirely about the big american wasteful image. There are some practical concerns that weigh into the choices Americans make re which cars to buy and if someone wants to get us to use more fuel efficient vehicles, they need to address these concerns rather than just insulting us for our choices.
it is not unheard of for the a/c clutch to disengage when the driver commands full throttle, but there's certainly no fuel economy involved there.
other item is climate control...
(a) it can switch off the compressor whenever it feels like
(b) which is usually never
(c) the EPA should take this into account with its tests--cars with automatic climate control should be in the default automatic mode throughout the test since that's how people drive them
>> Buy a rubbermaid container and put any liquids you're leaving
:-)
>> at the house in it- shampoo, 409, drain cleaner, liquid soap,
>> batteries, whatever.
Heh, you left bleach off the list
Seriously, it's a bad idea. Chemicals can leak out of their container and react with each other, producing flames or toxic gases.
The idea is that one well-built power supply, possible incorporated into a normal outlet would be more efficient than multiple individual transformers, aka wall warts.
It will cost less because less idle current will be needed.
It *might* be more convenient if it is designed properly (a lot of people find the power bricks cumbersome and annoying to travel with, for example.)
It isn't flexible in and of itself, but if the voltages are standardized the user will have more interchangeability regarding cables and chargers.
Wall warts only use about a watt when idle, but keep in mind there are maybe a billion of them in the world. A billion watts 24/7 is a lot of wasted power. If we can reduce that by gradually replacing them with a more efficent solution, it would probably be a good idea overall, even if you personally don't find it convenient.
The thinking is that any black holes that are created by the LHC would be so small that they would evaporate in an instant, probably within milliseconds of devouring the earth and sun. So there's nothing to worry about really.
The idea is that a common DC circuit would eliminate 5 or 10 or more wall warts. Also, a dedicated device could be designed with more sophisticated circuitry to drop its idle current down to nearly zero.
No matter what you do, the lack of porn on the disk is a dead giveaway that it's been wiped :-)
There are a few complaints about CD's that are irrespective of the nyqyest limit.
1. the first couple years of recorded CD's did not contain proper dither (a small amount of deliberately added noise). This is not a big deal but it meant that the theoretical noise floor was off by a few decibels. Nobody heard this because:
2. The first generation of CD players had vary gradual filters that kicked in around 12khz and killed everything about 15khz.
3. The second generation of players had some extra logic that figured out if the recording contained proper dither and only turned on the heavy filter when it was missing, otherwise you got the "2X oversampling" feature which helped a lot: the analog filter didn't have to kick in until 19khz or so.
4. The first several zillion CD's released were "remastered" from tapes. I.e. they just played the analog tape into a digital mastering board and left it at that. All the hiss, snap/crackle/pop, fret noise, sibilance, guy sneezing in 3rd row, all of it made it onto the CD perfectly preserved. In my opinion it wasn't until the late nineties when remastered versions of analog recordings were actually better than vinyl.
5. CD players have to spin the damn disk and most cheap-o CD players have little 0.00001 horsepower motors that spin at a 5 thousand rpm and they have really annoying physical noise that usually makes it out into the room. I hate that.
6. The first couple generations of CD changers had a bad reputation of breaking with 50 of your CD's inside...
7. My personal gripe: the advent of fully digital recording enabled a lot of no-talent performers to use various digital processors to make them not sound like the screeching walruses that they are, and to replace a lot of good session musicians with drum machines and beat boxes. Damn shame.
Anyway, most of these problems have been fixed over the years, but I think a lot of audiophile snobs were right about the first 2 generations of equipment and recordings on CD. The bits were there and capable of a clean 20khz, but the source material and some engineering mistakes really f'd things up at first.
I've looked at various people's ipods and in every single case the contents were just their CD collection plus some stuff they bought from itunes. No copying.
Plus, ipods don't count as repositories of music--you can only download music to the ipod, you can't upload it back out so any unauthorized copy is a dead-end once it gets to the ipod. It's not in the same category as Kazaa or bittorrent.
Not only that, but life on earth in that time frame was all in the ocean, which cosmic rays don't penetrate very far.
I think that what you say indeed helped the early adopters, but DVD player prices fell so fast (in the US at least) that for the most part, people here just bought PS2's to play games. I'm sure there are exceptions, but that's my observation.
I agree with your sentiment, and additionally things like fingerprints and retinal scans cannot be re-issued if compromised. This isn't a problem yet, but as biometric tokens are more widely used and thus more widely attacked it will become a problem.
PS1's played audio CD's and PS2's played regular DVDs but hardly anyone used them to do that. If sony is counting on blue-ray to be different in that regard, they may be deluding themselves.
Point taken. It might be more reliable to have a "fuzzy" frame fingerprint that kind of works but gets lots of false positives. Then check every frame in the commercial. If a whole bunch match and there is correlation in the timecodes then you've ID'd the commercial.
Cool thing is, you could set your DVR to show you new commercials you haven't seen before. I don't think I'm alone--what I hate about commercials isn't the content, it's the repetition. In that sense, so long as the fingerprint is generated from the same analog feed each time, some of the variation goes away and the fingerprinting works better.
Might also be interesting to try to flag entry/exit frames from known commercials (with a checksum or fingerprint sort of thing) into a shared database, then just look up the frames. If your show has 2.5 minutes with matches every 30 seconds, that's a commercial break. People can do a P2P vote with their FF button to contribute. Just a thought.