Easy to rectify: hack the car's firmware so it doesn't allow third party's control. Or use a car with good old mechanic clutch, with no pesky electronics in the way (a brute-force option).
Suggestion: Unless there will be a challenge-response scheme, this will be susceptible to replay attack. The detector gets the response it wants, it passes you as the car you met at the detector yesterday. I don't suppose the first versions will be designed as too secure.
The basic rule with machine checkpoints is to tell the machine what it wants to hear.
If there is no one topic whose inclusion or exclusion will make or break the sale, can't we take it further, and use the same argument to state there is no need to include anything at all into the encyclopedia.
Personally, I like the idea of encyclopedia that has a piece of everything inside. Face it, are there any real space limitations in computers? In paper books, there are definitely physical limits; but with the falling cost of disk space and memory, does it make sense to omit few kilobytes just because of space? Credible experts are another issue, but I personally prefer a not-entirely-credible answer (if marked as such) than nothing at all.
However, there were way too many cases of Americans flying taxpayer-paid laser-guided bombs into foreign buildings, often cowardly (or should I say tactically?) from high altitude. No need for civilian aircrafts when the military ones do better.
Everyone uses what they have. Welcome in fourth-generation warfare world.
The parasite signals emitted by the radios are very weak. An oscillator producing signal of suitable frequency is a trivial thing to do. If the billboard would employ Doppler effect recognition, or measure passing cars, add amplitude and frequency modulation, which will move the frequency up a little, raise the output power from zero to max, shift the freq down a little, then lower the output level. Then repeat the cycle in random intervals; a microcontroller will do. The whole assembly may be powered from a solar cell, for long life, and if equipped with a directional antenna, may be quite far away from the offending billboard (or a group of them); the required signal intensity is so low that the likeliness of interfering with anything other than the intended target is low even over lower ranges and relatively high power outputs.
Maybe we should give a culture jamming hint to Adbusters crew...
The lawyers may win, but the movie will live.
on
Star Wreck Trailer
·
· Score: 1
On P2P networks, forever being part of the culture that can be forced underground but can't be killed.
Lawyers can shut down their site, but at that moment "civil disobedience" sites crop up (hope) and the movie becomes one more symbol of big media corporations going after the individual creators. Which is in effect free advertising for the authors.
If it's on the Net, it can be scraped from the screen. A central service translating the data can be shut down, but if everybody scraps and translates their own data, there is nobody to come down upon unless they want to pull the information from the Net entirely. (There is no service known to me in this country, so I had to write my own downloader/extractor of TV guide and a program that automatically switches the TV on for news and sf, and downloads by keywords in the item description.)
They then can make the decoding more complicated by morphing the page generating template dynamically. But then the extraction can be done on parsed output, as the human-readable information is in known format; a time information looks one way, the title another way, the description too, and a set of suitable regexps can take care of it.
It's bandwidth-wasting in comparison with downloads of direct XML feed, but there is plenty of bandwidth on an average cable modem or DSL; or you can get a friend to fetch and pre-parse the data for you.
I believe it should be possible even to take a look at the page structure, analyze the layout of the tables and their content, and automatically recognize the data by their characteristics, omitting the need for a page-specific extractor at all; just tell the program how the data it looks for look and let it do the rest. Maybe there are some artificial intelligence experts here who could refine this approach?
It seems to be the ATA password. As it can be enabled and disabled and set at will, there is no encryption of the data stored on the maginetic surface itself, otherwise the disk would need awful lot of time to encrypt/decrypt everything. As most disks have only the head amplifiers inside the box with the platters and heads, I suppose the password itself will be located in some chip on the circuitboard, which is exchangeable. I'd love if someone here would try to swap the boards between passworded and unpassworded disk of the same type and report if the lock moved too.
The password itself could be maybe disabled as well, but that would probably require a good amount of luck (the data would have to travel through one of the accessible traces on the board, then they can be zeroed at the right moment by pulling the bus to L, and the drive then can think it has zero-length password; this works for many motherboards where the BIOS password is located in an EEPROM chip, usually a 24C08). The concrete approach will be very vendor-specific here, and good luck if everything including the NVRAM is inside one of the big flat chips. But the data recovery still could be possible, and disks are comparatively cheap these days.
About half year ago I designed something similar, with a pair of SMT160-30 sensors, a joystick port, some wire, some custom software, and gnuplot. I chose these because of not exactly stellar success I had earlier with an attempt to calibrate a thermistor, and these little babies output temperature as pulse-width modulated TTL signal with duty cycle calibrated against temperature; not really much to screw up there. I could also use more complicated 1-wire Dallas Semiconductors sensors, but they didn't have them in my favorite electronics shop and I wouldn't be able to connect them to input-only port.
Complete documentation is at here, including measurement data and graphs for last 8+ months with exception of few days of downtime.
Software and hardware may be pretty close. Lot of modern electronics, especially in smaller series where custom-manufacturing of chips is not effective yet, is based on FPGAs, universal chips that are just arrays of building blocks where some kind of "software" selects what block is connected to what other blocks and how it is configured to perform its function. To make things even more complicated, with suitably big FPGA you can configure part of it to form a RISC microcontroller and another part of it as a ROM with the controller's software (and then another part of the chip gets configured to be eg. USB controller, then yet another part's outputs directly drive the power FETs in H-bridges of the stepper motors - I got a nice small explosion when I did a software mistake once and opened all the transistors in the bridge and shorted a fat 24V power supply to the ground through them). This all embedded in one box that looks pretty much like just a piece of hardware. See eg. www.opencores.org to see what I mean.
How long does it to assemble a computer? Hint: once it took me under 15 minutes, though that was a bet, and the other time over half-hour (but I was blindfolded then and it was another bet). No warranty on the unit? No big deal - at least you can get in anytime you want and add/replace anything you want; if you have some old spare parts or less-important machines to cannibalize them from, it's a huge advantage - things like dead videocard or dead power supply then turn into a routine shutdown-swap-restart that can be done in the middle of the night (or weekend) without an Authorized Service Center, and then you can get the faulty part replaced under warranty at your nearest convenience. No worrying when the machine will return back from the shop, no hauling of big boxes when you can carry a card or a comparatively small box in your pocket, no (or much fewer) prolonged downtimes, no exposure of potentially sensitive data on the disks to untrusted third-party techs.
Do you still prefer warranty on whole units over warranty on parts, boxes sealed with warranty stickers, and machines made from non-commodity parts?
I care about TCO and even more about lack of frustration. I build my machines from parts.
> Who would you blame for that?
The switch designer. The device may stop working for many different reasons; worm, chewed wire, power blackout, corrosion, human error, bad luck. Railway switch should be designed to "fail gracefully".
For a camera, sure. But one megapixel is enough for many many purposes. Eg, you may want a hidden cam as part of home security system, and it's only the first idea I just got. Does the CMOS chip work in infrared? Digital cameras can have WAY too many purposes.
...or just hack the firmware to skip the encryption step. Or disable (or at least compromise) the RNG necessary for the RSA stuff, which reduces the work to crack the key to somewhere close to zero. Once you own the internals, you have many possibilities.
...and the next round will be SpamAssassin-like plug-ins for the clients. It's even possible to write a proxy that sits between the server and the client, and in a MITM-like way filters the communication, lying to the server that the message was shown and the ad was seen. I remember having some program (don't ask me what it was, it's couple years ago) which insisted on showing ads and refused to work when they were blocked. A simple Internet Junkbuster based proxy that served a solid-black GIF back for specified requests (instead of blindly blocking them) took care of it; the program still thought it gets advertising, and the ad window wasn't animated and colorful and distracting anymore.
Even if the watermark would be Really Bloody Good, a workaround still exists.
Watermarks are designed to survive modifications that don't distort the content significantly. So let's introduce a reversible distortion. Insert a scrambler before the watermark detector and a descrambler after it. If the detector is in the recording stage, use analog scrambler and a math model of the descrambler. Or if the hardware will be unmodifiable, you still can have a descrambler circuit between the headphones and the player. With low-noise op-amps the quality loss won't necessarily have to be meaningful.
Slashdot Mirror
Easy to rectify: hack the car's firmware so it doesn't allow third party's control. Or use a car with good old mechanic clutch, with no pesky electronics in the way (a brute-force option).
The basic rule with machine checkpoints is to tell the machine what it wants to hear.
If there is no one topic whose inclusion or exclusion will make or break the sale, can't we take it further, and use the same argument to state there is no need to include anything at all into the encyclopedia. Personally, I like the idea of encyclopedia that has a piece of everything inside. Face it, are there any real space limitations in computers? In paper books, there are definitely physical limits; but with the falling cost of disk space and memory, does it make sense to omit few kilobytes just because of space? Credible experts are another issue, but I personally prefer a not-entirely-credible answer (if marked as such) than nothing at all.
However, there were way too many cases of Americans flying taxpayer-paid laser-guided bombs into foreign buildings, often cowardly (or should I say tactically?) from high altitude. No need for civilian aircrafts when the military ones do better. Everyone uses what they have. Welcome in fourth-generation warfare world.
If they can't legally open Word documents, they still can send them to their European branch to un-DRM them under different jurisdiction.
Poof - problem gone. Bigger companies can open their own tiny overseas offices for this purpose, smaller ones can subcontract individuals.
Who modded this "insightful" instead of "funny"?
The parasite signals emitted by the radios are very weak. An oscillator producing signal of suitable frequency is a trivial thing to do. If the billboard would employ Doppler effect recognition, or measure passing cars, add amplitude and frequency modulation, which will move the frequency up a little, raise the output power from zero to max, shift the freq down a little, then lower the output level. Then repeat the cycle in random intervals; a microcontroller will do. The whole assembly may be powered from a solar cell, for long life, and if equipped with a directional antenna, may be quite far away from the offending billboard (or a group of them); the required signal intensity is so low that the likeliness of interfering with anything other than the intended target is low even over lower ranges and relatively high power outputs.
Maybe we should give a culture jamming hint to Adbusters crew...
On P2P networks, forever being part of the culture that can be forced underground but can't be killed.
Lawyers can shut down their site, but at that moment "civil disobedience" sites crop up (hope) and the movie becomes one more symbol of big media corporations going after the individual creators. Which is in effect free advertising for the authors.
If it's on the Net, it can be scraped from the screen. A central service translating the data can be shut down, but if everybody scraps and translates their own data, there is nobody to come down upon unless they want to pull the information from the Net entirely. (There is no service known to me in this country, so I had to write my own downloader/extractor of TV guide and a program that automatically switches the TV on for news and sf, and downloads by keywords in the item description.)
They then can make the decoding more complicated by morphing the page generating template dynamically. But then the extraction can be done on parsed output, as the human-readable information is in known format; a time information looks one way, the title another way, the description too, and a set of suitable regexps can take care of it.
It's bandwidth-wasting in comparison with downloads of direct XML feed, but there is plenty of bandwidth on an average cable modem or DSL; or you can get a friend to fetch and pre-parse the data for you.
I believe it should be possible even to take a look at the page structure, analyze the layout of the tables and their content, and automatically recognize the data by their characteristics, omitting the need for a page-specific extractor at all; just tell the program how the data it looks for look and let it do the rest. Maybe there are some artificial intelligence experts here who could refine this approach?
It seems to be the ATA password. As it can be enabled and disabled and set at will, there is no encryption of the data stored on the maginetic surface itself, otherwise the disk would need awful lot of time to encrypt/decrypt everything. As most disks have only the head amplifiers inside the box with the platters and heads, I suppose the password itself will be located in some chip on the circuitboard, which is exchangeable. I'd love if someone here would try to swap the boards between passworded and unpassworded disk of the same type and report if the lock moved too. The password itself could be maybe disabled as well, but that would probably require a good amount of luck (the data would have to travel through one of the accessible traces on the board, then they can be zeroed at the right moment by pulling the bus to L, and the drive then can think it has zero-length password; this works for many motherboards where the BIOS password is located in an EEPROM chip, usually a 24C08). The concrete approach will be very vendor-specific here, and good luck if everything including the NVRAM is inside one of the big flat chips. But the data recovery still could be possible, and disks are comparatively cheap these days.
Complete documentation is at here, including measurement data and graphs for last 8+ months with exception of few days of downtime.
Then underground shops offering TEMPEST-grade shielding start popping up fast.
Every attack has its defense and this chess game won't end anytime soon.
Software and hardware may be pretty close. Lot of modern electronics, especially in smaller series where custom-manufacturing of chips is not effective yet, is based on FPGAs, universal chips that are just arrays of building blocks where some kind of "software" selects what block is connected to what other blocks and how it is configured to perform its function. To make things even more complicated, with suitably big FPGA you can configure part of it to form a RISC microcontroller and another part of it as a ROM with the controller's software (and then another part of the chip gets configured to be eg. USB controller, then yet another part's outputs directly drive the power FETs in H-bridges of the stepper motors - I got a nice small explosion when I did a software mistake once and opened all the transistors in the bridge and shorted a fat 24V power supply to the ground through them). This all embedded in one box that looks pretty much like just a piece of hardware. See eg. www.opencores.org to see what I mean.
How long does it to assemble a computer? Hint: once it took me under 15 minutes, though that was a bet, and the other time over half-hour (but I was blindfolded then and it was another bet). No warranty on the unit? No big deal - at least you can get in anytime you want and add/replace anything you want; if you have some old spare parts or less-important machines to cannibalize them from, it's a huge advantage - things like dead videocard or dead power supply then turn into a routine shutdown-swap-restart that can be done in the middle of the night (or weekend) without an Authorized Service Center, and then you can get the faulty part replaced under warranty at your nearest convenience. No worrying when the machine will return back from the shop, no hauling of big boxes when you can carry a card or a comparatively small box in your pocket, no (or much fewer) prolonged downtimes, no exposure of potentially sensitive data on the disks to untrusted third-party techs.
Do you still prefer warranty on whole units over warranty on parts, boxes sealed with warranty stickers, and machines made from non-commodity parts?
I care about TCO and even more about lack of frustration. I build my machines from parts.
> Who would you blame for that? The switch designer. The device may stop working for many different reasons; worm, chewed wire, power blackout, corrosion, human error, bad luck. Railway switch should be designed to "fail gracefully".
For a camera, sure. But one megapixel is enough for many many purposes. Eg, you may want a hidden cam as part of home security system, and it's only the first idea I just got. Does the CMOS chip work in infrared? Digital cameras can have WAY too many purposes.
...or just hack the firmware to skip the encryption step. Or disable (or at least compromise) the RNG necessary for the RSA stuff, which reduces the work to crack the key to somewhere close to zero. Once you own the internals, you have many possibilities.
...and the next round will be SpamAssassin-like plug-ins for the clients.
It's even possible to write a proxy that sits between the server and the client, and in a MITM-like way filters the communication, lying to the server that the message was shown and the ad was seen.
I remember having some program (don't ask me what it was, it's couple years ago) which insisted on showing ads and refused to work when they were blocked. A simple Internet Junkbuster based proxy that served a solid-black GIF back for specified requests (instead of blindly blocking them) took care of it; the program still thought it gets advertising, and the ad window wasn't animated and colorful and distracting anymore.
Even if the watermark would be Really Bloody Good, a workaround still exists. Watermarks are designed to survive modifications that don't distort the content significantly. So let's introduce a reversible distortion. Insert a scrambler before the watermark detector and a descrambler after it. If the detector is in the recording stage, use analog scrambler and a math model of the descrambler. Or if the hardware will be unmodifiable, you still can have a descrambler circuit between the headphones and the player. With low-noise op-amps the quality loss won't necessarily have to be meaningful.