What makes you so amazing that your off-the-cuff development model is better than the development model the Ubuntu team has used to produce the third most popular operating system distribution in the world?
Has it ever occured to you that people titled "developers" just might know how to develop things?
What does loyalty have to do with giving you free stuff?
If she can sell tickets for $250 and instead choses to sell them for $50, that's basically deciding to give a random selection of her fans a free $200.
Just because you spent $30 on two of her CDs means you should get a chance for her to give you $200?
How, exactly, do you plan on getting people to use less energy?
Interesting answer. Now, how do you plan on getting people to decrease their energy consumption at a greater rate than that at which energy use is increasing?
Is that actually worth it? Go ahead and reply with the detailed cost-benifit analysis of your scheme in comparison to increasing our exploitation of renewable resources such as wind and tidal power.
Using a one time pad is actually a pretty solid plan. I was going to recommend it myself, but a remembered cryptosystem and key has an important property that a one time pad does not: you can't find it when you search the agent's luggage.
When selecting what cryptosystem to use for a given application, there are a number of tradeoffs. It's entirely possible that a playing card based system is the best deal.
This works great to protect against an attack that rarely matters (other users on the same server reading your data). I mean... I guess that comes up in shared hosting environments, but I don't see how it would matter in any serious application.
On the other hand, it does nothing to protect against the most common security issue - a bug in a CGI or PHP script that lets web users read arbitrary files. This more common issue can be fixed by storing the password in a place readable only by root.
I'd guess that using the playing card based shift register would be slightly safer than using solitaire for exactly one reason: In order to figure out the shift register thing you'll need to do a lot more research, which will tend to help you avoid some of the stupid usage mistakes.
If you can come up with a cryptosystem that can be used conveniently by hand yet is secure in the face of of computer-assisted cryptanalysis that'd be most impressive.
Any system that can be done by hand *at all* and has that property is potentially pretty useful in a field-agent situation.
You can increase the difficulty to crack the message by increasing the complexity of the encoding and decoding process all day long. The problem with that strategy is that for the system to be useful you need to be able to encode and decode your messages in a reasonable amount of time - and the attacker will usually have more time than that to try to crack your message.
In the real world, every robot doesn't turn against and kill it's master. Frankenstien was amusing, but the moral of that story is not some kind of natural law.
Replacing a human driver is a difficult AI problem - true. That doesn't make it impossible. A computer would be able to have more and more accurate sensory inputs (radar, IR, etc) and would be able to have massively faster reaction times.
You seem to think that there would be trouble handling exceptional situations (i.e. normal human driver behaviour, a deer, weather). You'd just program the computer to react the way I do when driving - assume that every object that you've seen move is going to try to jump in front of you, and keep your speed and direction such that you can evade their attempt to get hit.
Now, I'm not saying that a robot car that can drive in normal trafic will be developed in the next year or so, but I absolutely expect to see it in the next 20 years. Driving is a very robotic activity - you mostly just obey a simple set of traffic rules while trying to follow a path. Identifying the gestures of police officers directing traffic is the hardest subproblem I can think of.
It's not your responsibility to keep things from getting messy - that's the responsibility of the package manager.
Under either Windows or Linux, when you look at the root filesystem you see a bunch of implementation-detail crap that is completely irrelevent to you as a user. Both systems promote you never looking there - Windows goes so far as to hide stuff from you in the file manager.
From an actual user perspective, the two are exactly the same: You have your home folder, your desktop folder, some external disk drives that automount or autoplay, and a bunch of system files that you shouldn't be messing with.
I wonder how effective a scam someone could come up with based on selling these people spam mailing service and then not actually ever sending the emails...
Can users easily not make a normal user account and then work in a root desktop? If so, I'm not seeing what you've gained over the Mac OS X / Ubuntu sudo-by-default model.
It sucks when my wife wants to do something basic, and I have to go download a bunch of stuff and try to make it work with my current version of Linux.
I keep thinking this, but then I try to get Windows to do similar stuff and that's harder. In the end, the stuff that matters just works under Linux unless you're doing something stupid like upgrading Slackware 5 to udev by hand without upgrading glibc.
You fail to define "information". Is a game like Halo "information"? Is a program like Photoshop "information"? Not in any colloquial sense.
They are mathematical formulas that are the output of other mathematical formulas. They are no less "information" than the answer to questions like: "What are the first 1000 prime numbers?" or "What is the equation to find the diameter of a sphere given its surface area?"
Although the concept of no privacy is an interesting discussion topic, it's not nessisary for "freedom of the bitstream" as you put it.
A video of you murdering someone is could be perfectly legal. It's also obviously excellent evidence for your murder trial. Invasion of privacy can easily be a criminal offense without creating any undistributable bit patterns - i.e. it might be illegal for you to take photos of someone in the shower without their permission - and damages could be determined based on the results of their distribution (if that model works - it might not).
It's not terribly hard to trace where spam comes from. www.spamcop.net can take any spam email and give you the appropriate abuse@ address to report it to.
The problem is that all it takes to send spam is an internet connection and there's no good way to default to anything but "Accept" when email comes from an unknown sending server.
Adding a single bit to the key length doubles the search space.
Although that's true for symetric key ciphers, it's somewhat less than that for public keys. I'm not sure about the specifics, but it's such that going from 1024 to 2048 bits for a public key is a smaller jump than going from 96 to 128 bits for a symetric key.
Slashdot Mirror
What makes you so amazing that your off-the-cuff development model is better than the development model the Ubuntu team has used to produce the third most popular operating system distribution in the world?
Has it ever occured to you that people titled "developers" just might know how to develop things?
What does loyalty have to do with giving you free stuff?
If she can sell tickets for $250 and instead choses to sell them for $50, that's basically deciding to give a random selection of her fans a free $200.
Just because you spent $30 on two of her CDs means you should get a chance for her to give you $200?
You don't taper it to increase the energy production. The taper is to prevent supersonic airspeeds at the intakes for safety reasons.
There's a problem with that plan - there's strong evolutionary pressure against it.
How, exactly, do you plan on getting people to use less energy?
Interesting answer. Now, how do you plan on getting people to decrease their energy consumption at a greater rate than that at which energy use is increasing?
Is that actually worth it? Go ahead and reply with the detailed cost-benifit analysis of your scheme in comparison to increasing our exploitation of renewable resources such as wind and tidal power.
Yea. Good job. Do you realize how small a number 6,000 is?
It's a hell of a lot less than the number of *people* who die each year because coal power is used instead of wind power.
That kind of spy doesn't use an OTP. They use directional microwave trancievers that they point at a spy satilite by hand.
When selecting what cryptosystem to use for a given application, there are a number of tradeoffs. It's entirely possible that a playing card based system is the best deal.
On the other hand, it does nothing to protect against the most common security issue - a bug in a CGI or PHP script that lets web users read arbitrary files. This more common issue can be fixed by storing the password in a place readable only by root.
I'd guess that using the playing card based shift register would be slightly safer than using solitaire for exactly one reason: In order to figure out the shift register thing you'll need to do a lot more research, which will tend to help you avoid some of the stupid usage mistakes.
If you can come up with a cryptosystem that can be used conveniently by hand yet is secure in the face of of computer-assisted cryptanalysis that'd be most impressive.
Any system that can be done by hand *at all* and has that property is potentially pretty useful in a field-agent situation.
You can increase the difficulty to crack the message by increasing the complexity of the encoding and decoding process all day long. The problem with that strategy is that for the system to be useful you need to be able to encode and decode your messages in a reasonable amount of time - and the attacker will usually have more time than that to try to crack your message.
Why are you so untrusting of computers?
In the real world, every robot doesn't turn against and kill it's master. Frankenstien was amusing, but the moral of that story is not some kind of natural law.
Replacing a human driver is a difficult AI problem - true. That doesn't make it impossible. A computer would be able to have more and more accurate sensory inputs (radar, IR, etc) and would be able to have massively faster reaction times.
You seem to think that there would be trouble handling exceptional situations (i.e. normal human driver behaviour, a deer, weather). You'd just program the computer to react the way I do when driving - assume that every object that you've seen move is going to try to jump in front of you, and keep your speed and direction such that you can evade their attempt to get hit.
Now, I'm not saying that a robot car that can drive in normal trafic will be developed in the next year or so, but I absolutely expect to see it in the next 20 years. Driving is a very robotic activity - you mostly just obey a simple set of traffic rules while trying to follow a path. Identifying the gestures of police officers directing traffic is the hardest subproblem I can think of.
It's not your responsibility to keep things from getting messy - that's the responsibility of the package manager.
Under either Windows or Linux, when you look at the root filesystem you see a bunch of implementation-detail crap that is completely irrelevent to you as a user. Both systems promote you never looking there - Windows goes so far as to hide stuff from you in the file manager.
From an actual user perspective, the two are exactly the same: You have your home folder, your desktop folder, some external disk drives that automount or autoplay, and a bunch of system files that you shouldn't be messing with.
I wonder how effective a scam someone could come up with based on selling these people spam mailing service and then not actually ever sending the emails...
Can users easily not make a normal user account and then work in a root desktop? If so, I'm not seeing what you've gained over the Mac OS X / Ubuntu sudo-by-default model.
It makes at least as much sense as the Windows structure.
"C:\Documents and Settings\Bob Jones\My Documents" vs. "/home/bob"?
"C:\Windows\System32\etc\hosts" vs "/etc/hosts"?
I keep thinking this, but then I try to get Windows to do similar stuff and that's harder. In the end, the stuff that matters just works under Linux unless you're doing something stupid like upgrading Slackware 5 to udev by hand without upgrading glibc.
They are mathematical formulas that are the output of other mathematical formulas. They are no less "information" than the answer to questions like: "What are the first 1000 prime numbers?" or "What is the equation to find the diameter of a sphere given its surface area?"
Although the concept of no privacy is an interesting discussion topic, it's not nessisary for "freedom of the bitstream" as you put it.
A video of you murdering someone is could be perfectly legal. It's also obviously excellent evidence for your murder trial. Invasion of privacy can easily be a criminal offense without creating any undistributable bit patterns - i.e. it might be illegal for you to take photos of someone in the shower without their permission - and damages could be determined based on the results of their distribution (if that model works - it might not).
The problem is that all it takes to send spam is an internet connection and there's no good way to default to anything but "Accept" when email comes from an unknown sending server.
Although that's true for symetric key ciphers, it's somewhat less than that for public keys. I'm not sure about the specifics, but it's such that going from 1024 to 2048 bits for a public key is a smaller jump than going from 96 to 128 bits for a symetric key.
Bittorrent works fine through a firewall. It just makes it so that you can only connect to unfirewalled peers.
Find me a DVD copy of the old black & white Zorro TV series. It's owned by Disney, but they've apparently buried it.