Run it through twenty wildly different compression algorithms.
You'll eventually find a sort of "limit", towards which only incremental improvement can be made. The size of the data, after this compression, is generally referred to as the entropy level of the data. It's not always a perfect relationship between what the present algorithms find and the actual entropy of the data -- but what's key is, if a new method is found for representing the data in fewer bytes, it's de facto proof that there was less entropy in the data source than previously thought.
So that's what I mean when I refer to as entropy. It's the core measure of "what's there". To bridge the computer description with the physics description, over time, the universe gets harder and harder to describe. There's more information -- but less meaning to it. This is kind of like how striking a cymbal contains a huge amount of noise, yet creates an enormous amount of entropy (as we can see in the fact that it's the single hardest thing for any codec to encode).
Interestingly enough, compression strips data of all but its actual entropy, whereas encryption "armors" the entropy up to the size of the data. By that measure, decompression and cryptanalysis are interestingly interrelated.
I'm a crypto guy, AC. One of the things I look for are skewing issues -- for example, you can determine file transfer vs. typing over an SSH link by monitoring inter-packet timings and packet sizes. This is a source of information leakage that SSH does not protect against.
Deviation from the norm is high entropy, because there's a high amount of information content describing the shift.
The point of iTunes -- yes, iTunes, not DRM -- is to make it so convenient to buy music, that it's no longer worth using alternative methods of acquiring non-DRM'd (and increasingly illegal) music.
At that point, iTunes is iTunes Music Store. And Apple's made it pretty clear to other MP3 vendors -- they're not interested in letting anyone else play with their format. Compare to MS, which has bent over backwards, to the point of fabbing chips.
People play too many games with semantics; it's like a shell game. You don't get out of the fact that Apple is making an extremely closed music architecture by saying I'm bitching about the wrong product. It's Apple -- they're _famous_ for tying together everything so wonderfully! "It Just Works!" has a dark side.
No other DRM system actually lets you choose your player. iTunes only talks to iPod, at least w/o burning and re-ripping. You want to talk product tying -- MS doesn't even sell an MP3 player, let alone force you to use theirs.
But heh. Don't listen to me. I'm just a hardcore Linux user w/ a half terabyte RAID-5 FreeBSD box with fond memories of his old Apple IIgs days.
Not to mention I think this round of DRM won't end up any differently than it did for DAT/Minidisc/Dataplay -- eventual marginalization vs. products that actually want to work.
Yours Truly,
Dan Kaminsky
DoxPara Research
http://www.doxpara.com
The Helios Prototype was a unique electrically powered experimental lightweight flying wing developed by AeroVironment, Inc., under NASA's Environmental Research Aircraft and Sensor Technology (ERAST) program.
Metafilter pointed me towards a really interesting model for managing deception: Recognition detection. The idea is, rather than find out if someone is lying or not, simply find out if they recognize an object or scene they could only recognize if they were guilty. A certain brainwave, coined the P300 Wave, is emitted within a certain number of milliseconds of seeing an item one recognizes. One study, done by a group called Brain Wave Science, was able to reliably (and perfectly) separate FBI agents from average civilians by showing pictures of items from FBI training courses and operations. Detailed information may be found here.
I, of course, make no claims as to the veracity or accuracy of this material. But this wave is not pure pseudoscience -- the NYT has an article showing how weak P300's correspond to weak signal recognition. And BWS isn't the only group looking into P300 and deception.
There are other approaches -- blood flow and PET scans come to mind -- but this has the advantage of involving just a few electrodes.
So -- we may yet see a lie detector functional in our lifetime. Of course, it won't always be trusted, for reasons similar to the legalistic need for occasional exceptions to the rule of unique suspect DNA identifiers. But it'll be there.
Weeks certainly, months probably, years possibly. Everything degrades; what's very nice is that (noticably unlike a satellite), when you need to run maintenance on one of these things, you just tell it to land. I believe it'll even fly its way back up after you're done.
It's pretty sweet tech. Nice to see Aero getting more exposure:-)
Heh, this made Slashdot. Cool. So, yeah. I know these guys.
Basically, the airships is question are built by a company named Aerovironment (www.aerovironment.com). I've known about them for a while; one of my good friends works for the company. Really cool stuff; the basic idea is that this giant fixed wing circles around a rural area in the mid-to-upper atmosphere (where the air is thin enough to reduce drag, but thick enough to support lift) using solar power during the day and battery power at night. Then you drop some cell / wireless data relays on the bottom of the plane (UAV, to be more accurate), and poof: Regional visibility of a satellite relay, without the lag of communicating with a device being 22,500 miles away in geosynchronous orbit. That it's much cheaper to deploy the device (and possible to recover it as needed) is just gravy.
Things haven't been trivial for Aerovironment -- they lost one of their fixed wings some time ago during a test flight in Hawaii -- but as far as I know, they're the leaders in developing UAV's that simply don't need to land.
Actually, a couple people had some interesting observations on Groklaw about the request for further evidence.
For one, demanding newer versions of Dynix past 4.6.1 is apparently amusing, considering no newer versions exist. I suppose IBM could write one, but that's pushing discovery a bit far.
Secondly, failing to find misappropriated code between Linux and a version of AIX SCO has rights to is significant -- it means anything AIX-like that IBM has in Linux has to post-date the granting of code from SCO (or SCO's predecessors). Since the contract explicitly gives property rights to IBM for all of their own modifications, IBM has neatly caused SCO to show that Linux's similarities to AIX, if any, did not occur within the "protected window" that SCO purchased ownership of.
It's below freezing on the surface (no atmosphere to retain heat). Not to mention that whole thin atmosphere thing doesn't provide enough pressure to prevent liquid water from boiling away anyway.
Mud is water spatially mixed with soil, but not chemically bonded. It would freeze (as we saw in Boston, when they froze the soil for three years straight to prevent it from collapsing during the Big Dig).
I was going to get XM or Sirius setups for one or two relatives (as in, was at the store, had liked XM, but Sirius had a nicer receiver).
Too bad you couldn't obviously give people the actual subscriptions as gifts. Problem was, providing gifts as subscriptions exposed the fact that there was a subscription fee -- and if you look around the packaging and brochures, you'll find said fee is nowhere to be found (at least at the three stores I looked -- Fry's, Good Guys, and Best Buy).
Since there's no fee to be found, there's no "six month subscription" to be found either.
I wasn't about to give a gift that came with a recurring fee w/ no demo. So I bought something more interesting.
That can be your last response or not, but trust me -- you're quite a bit off.
Sessions are indeed supported by ISO-9660 circuitry -- by the readers. Where they aren't supported worth a damn are in the writing process. Drag and drop for CD's w/ ISO-9660 does not exist because it cannot exist -- the file system is too static. DirectCD (and other CD-RW solutions) use a packetized file system -- another way for referring to sector oriented. Notably, they do not work by default.
That's the fact you've utterly missed. Let me describe the process of using an MP3 player that doesn't support the MSC profile (the one that makes the device show up as a hard drive):
1) Insert CD with driver and software. 2) Install software. 3) Insert device. 4) Launch software. 5) Learn software. 6) Manipulate it to move files as needed.
Compared to:
1) Insert device. 2) Copy files.
See, I can say this, because I'm looking at (no joke) my ELEVENTH MP3 PLAYER, just bought a few hours ago. (You may mock me for this.) I've used quite a few of these players. Things that don't show up as a drive -- don't just work -- well, they suck. FAT32 is the only game in town that "just works". The grand critique of DRM is that the user needs to learn a whole new interface paradigm, compared to what they're used to (just copy the files to the player and go, no need to view the latest bizarrely skinned application of the day).
Developing a competing standard isn't hard. Developing one that works on arbitrary machines -- that's impossible, because MS controls what ships. You being able to only use your player on your computer is only bad to you. Remember, part of the DRM game is suppressing file sharing; the idea of "heh, that's a cool song, lemme pull it off your player" is anathema. Anything that suppresses this is Good.
I'm proud of your dad, but *ahem* I'm no slouch either. MS is caught between a rock and a hard place -- they're traditionally the 800lb gorilla that's enabled as much access to their users as possible. (Little realized fact is that MS was the first company to embed MP3 into their OS, through an ACM driver.) But they're doing alot to try to woo Hollywood -- Black Hat Windows last year was held w/ the SMPTE meeting (hollywood video folks), and MS had rented out an entire theatre to try to woo the guys to Windows Media for Theatres. Pushing the industry has become a story of compromise, and compromise means your system follows restrictions you didn't select (like your DVD player showing you 45 seconds of copyright warnings / movie previews whether you want to see them or not).
Migrating people away from FAT, which (as a sector level interface) is very difficult to add fine grained permissions to, is part of such compromises.
This comment is just wrong, and I think you know it: "Businesses will be eager to have the opportunity to use a supported codebase for their FAT access and thus not have to deal with possible bug problems in their own code." Businesses are never, ever eager to change something that works. Ever. EVER. Ask your dad.
Not to mention I think some of the FAT implementations are in hardware. (Note, I said 'I think'.) Switching to the MS code would be a total rebuild.
What are you saying with regards to China? This doesn't affect them, because they'll just ignore the rules? So it doesn't matter that MS is trying to set them? That means a plan will fail, not that there is no plan.
I will make one claim of ignorance...I don't know what YHBT means. Certainly I don't think you're stupid; you're pretty well spoken. But you're a bit misinformed -- you see the general rule (people can use competing standards, a $0.25 per device is cheap) and ignore the particularities of the computer market (anything that doesn't "just work" fails enough to kill profit margins, and those teeny chinese co
By file systems, I mean mechanisms for moving files onto a sector-addressable block device, a la a hard drive or flash disk.
ISO-9660 does not support additions after the fact; the way this was managed on CD's, there'd be multiple "sessions" upon which the file systems would stack. Notably, this is not supported by default. I also suspect that simply moving the file system onto a USB disk would be responded to with a blank stare; the OS does not likely know it can break out the CD-ROM driver on something that looks like a hard drive.
Bottom line, it's a question of what works by default for both the MP3 manufacturers and MS Windows. Right now, it's FAT32. By adding a 25 cent levy, they start to be able to change that.
You probably aren't aware of the huge fight China's starting to push against this. Go check out what's happening with the new Chinese WiFi security standard, or EVD. Trust me, there's some _big_ wars going on over patent power; MS is just fanning those flames.
Hello? ISO-9660 is a fixed-directory, non-fragmentable format, i.e. yes, the MP3 player could expose its contents to be read, but you're not uploading files to the drive via drag and drop.
There is no other embeddable and rewritable file system supported by Windows machines, and it's a miserable process mucking with the DDK to add drivers for alternate systems (witness the number of failed and unstable attempts).
HeySubcontinent's story linkage analyzes the automatic stegoplagarization of documents written in the language derived from Britain. Expected to displace at some point journalists, these hacks presently bash with the force of a small child. Good.
$1/unit was enough to kill firewire. You don't think $.25/unit on a mass-margin device is a huge deal, especially if you _have_ to use a MS implementation?
There are slightly broken implementations of FAT out there, but nothing too serious. I've seen Microsoft's pissed off docs (went through 'em when I build an efficient sector copier); the problems aren't that bad.
Slashdot Mirror
rangek--
Take some data.
Run it through twenty wildly different compression algorithms.
You'll eventually find a sort of "limit", towards which only incremental improvement can be made. The size of the data, after this compression, is generally referred to as the entropy level of the data. It's not always a perfect relationship between what the present algorithms find and the actual entropy of the data -- but what's key is, if a new method is found for representing the data in fewer bytes, it's de facto proof that there was less entropy in the data source than previously thought.
So that's what I mean when I refer to as entropy. It's the core measure of "what's there". To bridge the computer description with the physics description, over time, the universe gets harder and harder to describe. There's more information -- but less meaning to it. This is kind of like how striking a cymbal contains a huge amount of noise, yet creates an enormous amount of entropy (as we can see in the fact that it's the single hardest thing for any codec to encode).
Interestingly enough, compression strips data of all but its actual entropy, whereas encryption "armors" the entropy up to the size of the data. By that measure, decompression and cryptanalysis are interestingly interrelated.
--Dan
I'm a crypto guy, AC. One of the things I look for are skewing issues -- for example, you can determine file transfer vs. typing over an SSH link by monitoring inter-packet timings and packet sizes. This is a source of information leakage that SSH does not protect against.
Deviation from the norm is high entropy, because there's a high amount of information content describing the shift.
--Dan
The more we deviate from normality, the more value we place on privacy.
The more we deviate from normality, the more information value there exists within our deviation.
In other words, the greater our individual entropy, the more value we attach to it.
This is an interesting result; a first step towards quantification of something I had not really conceived of as quantifiable.
--Dan
AC--
The point of iTunes -- yes, iTunes, not DRM -- is to make it so convenient to buy music, that it's no longer worth using alternative methods of acquiring non-DRM'd (and increasingly illegal) music.
At that point, iTunes is iTunes Music Store. And Apple's made it pretty clear to other MP3 vendors -- they're not interested in letting anyone else play with their format. Compare to MS, which has bent over backwards, to the point of fabbing chips.
People play too many games with semantics; it's like a shell game. You don't get out of the fact that Apple is making an extremely closed music architecture by saying I'm bitching about the wrong product. It's Apple -- they're _famous_ for tying together everything so wonderfully! "It Just Works!" has a dark side.
--Dan
No other DRM system actually lets you choose your player. iTunes only talks to iPod, at least w/o burning and re-ripping. You want to talk product tying -- MS doesn't even sell an MP3 player, let alone force you to use theirs.
But heh. Don't listen to me. I'm just a hardcore Linux user w/ a half terabyte RAID-5 FreeBSD box with fond memories of his old Apple IIgs days.
Not to mention I think this round of DRM won't end up any differently than it did for DAT/Minidisc/Dataplay -- eventual marginalization vs. products that actually want to work.
Yours Truly,
Dan Kaminsky
DoxPara Research
http://www.doxpara.com
Wow. Who would have thought we'd see cellular automata in, um, cells.
OK, I'm being a bit harsh -- this is very cool work. But yes, life does play the game of life. It's called that for a reason.
--Dan
Gaze! Gaze upon the horrifying work conditions my roommate is forced to tolerate.
Man. What people will do for a paycheck. Poor guy, in a cube all day...
--Dan
The Helios Prototype was a unique electrically powered experimental lightweight flying wing developed by AeroVironment, Inc., under NASA's Environmental Research Aircraft and Sensor Technology (ERAST) program.
Metafilter pointed me towards a really interesting model for managing deception: Recognition detection. The idea is, rather than find out if someone is lying or not, simply find out if they recognize an object or scene they could only recognize if they were guilty. A certain brainwave, coined the P300 Wave, is emitted within a certain number of milliseconds of seeing an item one recognizes. One study, done by a group called Brain Wave Science, was able to reliably (and perfectly) separate FBI agents from average civilians by showing pictures of items from FBI training courses and operations. Detailed information may be found here.
I, of course, make no claims as to the veracity or accuracy of this material. But this wave is not pure pseudoscience -- the NYT has an article showing how weak P300's correspond to weak signal recognition. And BWS isn't the only group looking into P300 and deception.
There are other approaches -- blood flow and PET scans come to mind -- but this has the advantage of involving just a few electrodes.
So -- we may yet see a lie detector functional in our lifetime. Of course, it won't always be trusted, for reasons similar to the legalistic need for occasional exceptions to the rule of unique suspect DNA identifiers. But it'll be there.
--Dan
Weeks certainly, months probably, years possibly. Everything degrades; what's very nice is that (noticably unlike a satellite), when you need to run maintenance on one of these things, you just tell it to land. I believe it'll even fly its way back up after you're done.
:-)
It's pretty sweet tech. Nice to see Aero getting more exposure
--Dan
Heh, this made Slashdot. Cool. So, yeah. I know these guys.
Basically, the airships is question are built by a company named Aerovironment (www.aerovironment.com). I've known about them for a while; one of my good friends works for the company. Really cool stuff; the basic idea is that this giant fixed wing circles around a rural area in the mid-to-upper atmosphere (where the air is thin enough to reduce drag, but thick enough to support lift) using solar power during the day and battery power at night. Then you drop some cell / wireless data relays on the bottom of the plane (UAV, to be more accurate), and poof: Regional visibility of a satellite relay, without the lag of communicating with a device being 22,500 miles away in geosynchronous orbit. That it's much cheaper to deploy the device (and possible to recover it as needed) is just gravy.
Things haven't been trivial for Aerovironment -- they lost one of their fixed wings some time ago during a test flight in Hawaii -- but as far as I know, they're the leaders in developing UAV's that simply don't need to land.
--Dan
Some of the most intensely religious people you'll find are brilliant physicists.
It's not a GPL violation unless their code is integrated, on an object level, with the licensed code.
Mac OSX can ship with GPL'd code, for example, and not have to open up Quartz.
--Dan
Actually, a couple people had some interesting observations on Groklaw about the request for further evidence.
For one, demanding newer versions of Dynix past 4.6.1 is apparently amusing, considering no newer versions exist. I suppose IBM could write one, but that's pushing discovery a bit far.
Secondly, failing to find misappropriated code between Linux and a version of AIX SCO has rights to is significant -- it means anything AIX-like that IBM has in Linux has to post-date the granting of code from SCO (or SCO's predecessors). Since the contract explicitly gives property rights to IBM for all of their own modifications, IBM has neatly caused SCO to show that Linux's similarities to AIX, if any, did not occur within the "protected window" that SCO purchased ownership of.
Elegant.
--Dan
http://www.doxpara.com/apps/cdcygssh/
/bin/bash --login -i ...can do to your unix experience. Look, ma. Windows has a shell :-)
Go there in IE.
It's amazing what:
C:\cygwin\bin\rxvt.exe -sl 20000 -rv -fn "Lucida Console-10" -e
--Dan
It's below freezing on the surface (no atmosphere to retain heat). Not to mention that whole thin atmosphere thing doesn't provide enough pressure to prevent liquid water from boiling away anyway.
Mud is water spatially mixed with soil, but not chemically bonded. It would freeze (as we saw in Boston, when they froze the soil for three years straight to prevent it from collapsing during the Big Dig).
--Dan
Isn't there some clever way to work around these limitations?
There will be.
That's the cost of 33 months of service at $12.95 a month. That's almost 3 years before you reach break-even!
Oops.
I was going to get XM or Sirius setups for one or two relatives (as in, was at the store, had liked XM, but Sirius had a nicer receiver).
Too bad you couldn't obviously give people the actual subscriptions as gifts. Problem was, providing gifts as subscriptions exposed the fact that there was a subscription fee -- and if you look around the packaging and brochures, you'll find said fee is nowhere to be found (at least at the three stores I looked -- Fry's, Good Guys, and Best Buy).
Since there's no fee to be found, there's no "six month subscription" to be found either.
I wasn't about to give a gift that came with a recurring fee w/ no demo. So I bought something more interesting.
--Dan
Canar,
That can be your last response or not, but trust me -- you're quite a bit off.
Sessions are indeed supported by ISO-9660 circuitry -- by the readers. Where they aren't supported worth a damn are in the writing process. Drag and drop for CD's w/ ISO-9660 does not exist because it cannot exist -- the file system is too static. DirectCD (and other CD-RW solutions) use a packetized file system -- another way for referring to sector oriented. Notably, they do not work by default.
That's the fact you've utterly missed. Let me describe the process of using an MP3 player that doesn't support the MSC profile (the one that makes the device show up as a hard drive):
1) Insert CD with driver and software.
2) Install software.
3) Insert device.
4) Launch software.
5) Learn software.
6) Manipulate it to move files as needed.
Compared to:
1) Insert device.
2) Copy files.
See, I can say this, because I'm looking at (no joke) my ELEVENTH MP3 PLAYER, just bought a few hours ago. (You may mock me for this.) I've used quite a few of these players. Things that don't show up as a drive -- don't just work -- well, they suck. FAT32 is the only game in town that "just works". The grand critique of DRM is that the user needs to learn a whole new interface paradigm, compared to what they're used to (just copy the files to the player and go, no need to view the latest bizarrely skinned application of the day).
Developing a competing standard isn't hard. Developing one that works on arbitrary machines -- that's impossible, because MS controls what ships. You being able to only use your player on your computer is only bad to you. Remember, part of the DRM game is suppressing file sharing; the idea of "heh, that's a cool song, lemme pull it off your player" is anathema. Anything that suppresses this is Good.
I'm proud of your dad, but *ahem* I'm no slouch either. MS is caught between a rock and a hard place -- they're traditionally the 800lb gorilla that's enabled as much access to their users as possible. (Little realized fact is that MS was the first company to embed MP3 into their OS, through an ACM driver.) But they're doing alot to try to woo Hollywood -- Black Hat Windows last year was held w/ the SMPTE meeting (hollywood video folks), and MS had rented out an entire theatre to try to woo the guys to Windows Media for Theatres. Pushing the industry has become a story of compromise, and compromise means your system follows restrictions you didn't select (like your DVD player showing you 45 seconds of copyright warnings / movie previews whether you want to see them or not).
Migrating people away from FAT, which (as a sector level interface) is very difficult to add fine grained permissions to, is part of such compromises.
This comment is just wrong, and I think you know it: "Businesses will be eager to have the opportunity to use a supported codebase for their FAT access and thus not have to deal with possible bug problems in their own code." Businesses are never, ever eager to change something that works. Ever. EVER. Ask your dad.
Not to mention I think some of the FAT implementations are in hardware. (Note, I said 'I think'.) Switching to the MS code would be a total rebuild.
What are you saying with regards to China? This doesn't affect them, because they'll just ignore the rules? So it doesn't matter that MS is trying to set them? That means a plan will fail, not that there is no plan.
I will make one claim of ignorance...I don't know what YHBT means. Certainly I don't think you're stupid; you're pretty well spoken. But you're a bit misinformed -- you see the general rule (people can use competing standards, a $0.25 per device is cheap) and ignore the particularities of the computer market (anything that doesn't "just work" fails enough to kill profit margins, and those teeny chinese co
Canar,
Dude.
You don't understand the technology. That's fine, but stop acting like you do. Here are the file systems that Windows machines can support:
FAT32: Readable, writable, embeddable
NTFS: Readable, writable
ISO-9660:Readable, embeddable
By file systems, I mean mechanisms for moving files onto a sector-addressable block device, a la a hard drive or flash disk.
ISO-9660 does not support additions after the fact; the way this was managed on CD's, there'd be multiple "sessions" upon which the file systems would stack. Notably, this is not supported by default. I also suspect that simply moving the file system onto a USB disk would be responded to with a blank stare; the OS does not likely know it can break out the CD-ROM driver on something that looks like a hard drive.
Bottom line, it's a question of what works by default for both the MP3 manufacturers and MS Windows. Right now, it's FAT32. By adding a 25 cent levy, they start to be able to change that.
You probably aren't aware of the huge fight China's starting to push against this. Go check out what's happening with the new Chinese WiFi security standard, or EVD. Trust me, there's some _big_ wars going on over patent power; MS is just fanning those flames.
--Dan
Hello? ISO-9660 is a fixed-directory, non-fragmentable format, i.e. yes, the MP3 player could expose its contents to be read, but you're not uploading files to the drive via drag and drop.
There is no other embeddable and rewritable file system supported by Windows machines, and it's a miserable process mucking with the DDK to add drivers for alternate systems (witness the number of failed and unstable attempts).
HeySubcontinent's story linkage analyzes the automatic stegoplagarization of documents written in the language derived from Britain. Expected to displace at some point journalists, these hacks presently bash with the force of a small child. Good.
Canar,
$1/unit was enough to kill firewire. You don't think $.25/unit on a mass-margin device is a huge deal, especially if you _have_ to use a MS implementation?
There are slightly broken implementations of FAT out there, but nothing too serious. I've seen Microsoft's pissed off docs (went through 'em when I build an efficient sector copier); the problems aren't that bad.
--Dan
I tend do be rather deferential to those who have latex gloves, and believe you me, are willing to use them.
About the most that's safe to do is state a complaint will be made -- and actually make it. After.