Think of it this way... design a blueprint of your building in something like sketchup, and tag all the different surfaces according to type. Then put antennas in strategic places in the structure (could do 3 in each room, or surround the entire building) and turn on the multiband antennas.

The RF interference should be able to be mapped in this way fairly easily; especially if you set state and then open/close doors, turn WiFi on/off, turn on lights/heaters, etc.

Record all this state information, and then with minimal training, the system should be able to identify all mobile objects and when they moved.

If you toss in the RFID chips here, you could serialize each major mobile item, and actively track them anywhere in the structure.

This isn't new, and doesn't need all that much knowledge of RF, just a good AI that can learn.

See: http://www.extremetech.com/ext... for one way it's been done.

I presume you read this?
http://idle.slashdot.org/story...

I've got a solution that will make everyone happy:
Have NetFlix partner with the NFB to distribute NFB content... globally. Nothing like providing global access to Canadian content. NetFlix could even provide it for free to everyone in Canada with an account but no current subscription. Under this setup, the CRTC wouldn't have a leg to stand on, as at that point, they will get their Canadian Content on NetFlix (not sure about the French/English ratio though).

HOWEVER

I'm pretty sure this really has nothing to do with NetFlix and EVERYTHING to do with the new consortium raising a Canadian NetFlix "competitor" (Shomi) whispering nasty things in the CRTC's ear. Yes, blame Rogers/Shaw for this fracas, as they're likely where the blame really lies.

Hmm... scratch that; it's not a re-branding, as the Dreamer uses more filament types, is heated, and has a few other goodies missing in the Builder. And it's only $200 more.

Oh, and it appears to be a re-branded one of these http://www.flashforge-usa.com/... which say they support Linux. I'm guessing that FlashForge Dreamer drivers for Linux would drive the Idea Builder without much modification.

If it works with OS X and Vista, that means you should be able to reverse engineer the protocol and make it work with Linux without too many issues. The drivers seem to work with AutoCAD tools, so my guess is the drivers aren't proprietary either but use a common protocol.

I'll bite. I've used a Dremel-brand dremel tool in the late 90's, and found it solid (if made of a lot of plastic), dependable, and accurate. The accessories were way too expensive, but Black & Decker accessories are of the same quality and fit in the Dremel opening.

B&D, Ryobi, Makita and similar manufacturer's dremel tools though -- I've found to be underpowered, made of cheap components, and have a shaft locking mechanism that is abysmal, not holding the shaft in a centred manner at all. DeWalt is also pretty good.

Likewise, I've had hit-and-miss experience with Dremel's other offerings -- some are good, some aren't. But their original tool still works as well as it ever did.

True... which is why talking about large-bit encryption isn't really the issue; it's the implementations that are the issue. I was mostly rebutting the part about exponential difficulty with bitsize making your encryption more secure. I'd give *properly implemented* AES-256 another decade at least before it has any security issues whatsoever. By the time AES-256 can be cracked via brute force, the entire algorithm will be out of date, so increasing bitsize won't be much of a gain.

But it doesn't matter how many bits are used or what algorithm, or even what implementation, if even one password at, or above your level on the system being protected is in the Adobe password file, people.

This actually is a good idea for an Apple Accessory(TM) -- make a line of jewelery that can store your passcode in a "hidden" compartment. Any attacker would need to get both the jewelery and the phone to gain access, which is better than nothing at all.

However, the device unlocking problem has already been solved on iPhones: TouchID. You don't need your passcode to unlock a device, you need your passcode to manage the device in cases where your thumb is missing or you're not actually on the device, but need access to its remotely-stored data.

As such, it makes much more sense to make your passcode really long, write it down, and store it in a safety deposit box. Day to day, you won't need it. But if you do, one trip to the bank and 10 min in the safe room with your phone will be enough to recover from whatever situation you've got yourself into.

I think this requires clarification: If you back up to iCloud instead of your personal computer, the backups are encrypted with keys that Apple has. And anyone who has your UUID (which they can likely pick up by sitting on the same open WiFi as you) can spoof your device for a restore of said backup, without requiring 2-factor authentication (they'll still need to figure out your Apple ID and password, or have those given to them by Apple).

If you back up locally, you control the backup key, and it never gets broadcast over a rogue WiFi AP, as any backup attempts (even if you enable WiFi backup) have to happen to a local server, and are not tied to your UUID.

One other thing to note: on iDevices, if you select a non-simple passcode that is only numbers, the device still presents the simple PIN screen instead of a full keypad. The difference is that it sticks an "OK" button in the text field that you press when you're done.

This provides a passcode of uncertain length (X choose 10, 0 x 4096 or so, realistically 16) that is still relatively easy to enter. It's not as secure as a full-on textual passcode, but it beats a 4-digit PIN even if you only use a 4-digit PIN -- as the attacker has no means to know how many digits long your PIN is -- as it *could* be "11151111" or even "1231230123123" which is pretty quick and easy to enter on a PIN pad (almost as fast as 12345), is 13 characters long, and really difficult to guess.

4096-bit encryption sounds great, but there are always ways to shave orders of magnitude off of the actual sample space, such that encryption strength really tends to grow at about the same rate as processing power.

Once you get encryption keys of that size, you've got storage and transmission issues, which increase the probability of other attack vectors working. Plus, your PRNG has to be REALLY random -- and there isn't really all that much true entropy to go around when you get right down to it.

The issue here is that as your random seed gets larger, the probability that it isn't truly random also increases, and analysis of data encrypted with this seed becomes easier through replay analysis. It won't remove the actual entropy, but forces acting on the values generated will create patterns that will still limit the amount of true entropy stored in the resulting value.

Kind of like if you flip a coin once, anyone guessing really doesn't know if it'll come up heads or tails. But if you flip a coin ten times, the method you use for flipping the coin and the environmental factors will start to have an impact on which side comes up more often, and also on the pattern of what influences a heads or tails result. If you flip the coin 2^4096 times, you'd probably be able to pretty accurately predict the result before the coin had even fully launched into the air.

Yeah; the coin experiment is often used as an example of how entropy is entropy and the probability doesn't change from toss to toss. But if you take all other factors into consideration, you limit the effect of entropy such that your guess on a given toss can actually improve over time. Try it, and you'll see I'm correct.

True as far as it goes, but "Hollywood" is a subsidized arm of the US government that mass exports modified propaganda to the rest of the world -- it's why the prevailing view of many Chinese of the US prior to the Internet was based on 1950's films. The result is not that everyone believes the movies (although if you repeat something false often enough, a large portion of the listening population will begin to take it as fact), but that everyone's perception of the subjects of the movies is strongly shaped by them (so as a result, people become wary of giant robots because they might possibly beat us to the moon).

Jingoistic whining is one thing, but oversimplifying herd dynamics isn't much better and helps nobody.

If petty jingoistic whining really bothers you that much, complain about it on slashdot -- not that this will stop the whining.

Exactly. This also means that any data exfiltration HAS to be targeted, unless Apple is forced to install a keylogger on/in every device (which is possible, what with secret courts/warrants/etc). As soon as you have real targeting, warrants come back in full force, and you have a working set of checks and balances. The issue here isn't with governments gaining access to targeted data, it's with governments gaining access to everyone's data and then rummaging around to see what they can find that is "bad" or "useful". In short, there's less room for abuse, and more transparency with regards to any actions actually taken.