I meant to say 'generate a result that could be misinterpreted as a valid message', thus maintaining the so-thought integrity of the fake message.
Ah, I think I see where my confusion arose. When you said "one of the escrow holders manages to create a fake key" you didn't mean that the key itself was fake - it would still be right key, according to the key escrow process - but that the original encryption could have been done in such a way as to cause the correct key to return a misleading result?
produces an entirely different result than the key holder's genuine key should generate.
Not if by "genuine key" you mean the key used by the proprietary device, and for which step it also generates the secret split key to allow decryption by agencies. Any decryption by either of those keys will result in the correct decryption.
If by "genuine key" you mean the key used to encrypt the data before it went through device encryption, well, then the escrow decryption process will still produce the correct intermediate (once-encrypted) text.
Almost nobody writes texts on the bus because most people take the bus during peak periods, when there's no ROOM to text.
That's not been my anecdotal experience.
And during off-peak hours, it's still too bumpy even on good roads.
Again, not my experience. Maybe we have better roads around here, or people are more tolerant than you think. But fine, let's shift the example - which was only supposed to be about people in close proximity - to the bus station where everyone's waiting for a bus. All the points I made still stand. Ambient noise, other people speaking, and privacy. Which people do want, even if they don't need it. Try asking a real person sometime. They're funny like that.
People in offices use headsets
Not in my office. Surely what you mean is that people in offices will have to use headsets, should your keyboard-less future come about. Which it won't. Keyboards are far and away a more practical input device than speech recognition, certainly for many years to come, if not indefinitely.
Being able to offer solutions to the problems that SR causes is not an argument in their favour when those problems don't exist with keyboards.
Call centers are an example.
In call centers headsets are a necessity. Given the choice I'm sure most people who work in one would prefer not to have to put up with the constant hub-bub. Call centers are also a great example of when speech recognition would be utterly impractical, since call center staff often need to both operate a computer and be free to speak to the person on the opposite end of the call at the same time.
People dealing with sensitive stuff should already have an office with a door that closes, so again, no need for a keyboard.
Again you're dictating how things must be done in your keyboard-less utopia, instead of making the argument for why people should prefer SR over keyboards. "Sensitivity" is also a sliding scale, like privacy, and there are plenty of things on it that'll fall into the "doesn't need a separate office, but doesn't need to be blurting it out loud" category for which a keyboard is more than adequate.
And for your private stuff, do it on your own time at home, same as everyone else.
There you go again, dictating when and how I should do something. Are there only two places people ever go, work and home? And why must they be expected to happily forego any notion of privacy when they're not in the latter?
Not that it makes much difference to the substance of your point, but I don't think anyone's proposing literally a single key. It could (hypothetically, naively) be one split key per company, or per product, or batch of a product, or maybe even one split key per "real" key.
I might be missing something which rules out any or all of those possibilities, though.
one of the escrow holders manages to create a fake key
Not quite sure what you mean. Do you mean one of the escrow holders providing a fake "part" of the key, to be joined with the other real parts, thus producing a full, but false, key? Or producing an entire fake key by themselves?
when used to decrypt some given message, produces an entirely different result than the key holder's genuine key should generate
Isn't that what all incorrect keys do? Generate a result different from what the genuine key would produce?
You are the one who brought up the bogus bus example, so don't blame me that it's weak as wet toilet paper.
How is it weak? And in what did I blame you for said imagined weakness?
You can point out how it might be slightly more awkward to text on a bus than when you're sitting at home until you're blue in the face, but that's not going to make up for the fact that speech recognition is going to create more problems that it would solve in that situation for most people. It'd have to work over any ambient noise, including other people trying to send text messages by babbling at their phones or - ugh! - having conversations. It would mean your texts becoming known to anyone within earshot, which plenty of people wouldn't much care for - or care to be subjected to - even if it is just about getting more milk.
If that's fine with you, that's your choice. But do you at least see that plenty of other people won't feel that way? Can you really not imagine any circumstance in which you might not be happy with using speech recognition to send a text for privacy reasons?
And if you believe that ANY of your messages are private, you're incredibly naive.
I never said I believed that. Leaping from there to "well, might as well just speak my every communication out loud, it makes no difference" doesn't make a whole lot of sense though.
Some database somewhere might well have a list of all the websites I visited last night, but that doesn't mean I'd be fine with the idea of having the list automatically posted to a website for anyone to browse.
The masses will always choose convenience over security.
Really? Because I'd find it really convenient not to need a key to open my front door or start my car.
But if you're that paranoid about your text messages being public
What's paranoid about just not wanting strangers on a bus knowing your business, no matter how mundane?
Again, have you ever tried to type on a moving bus?
Again, my only point is that a keyboard is generally a far more useful and usable input method than speech recognition for such a situation and many, many others.
Also, most of your messages don't need to be private
Most people don't work with computers with keyboards (think point-of-sale touchscreens, etc)
Pens aren't obsolete even though they're used far less in business since the invention of the typewriter.
So what's more natural for them than to just say your message and hit send?
It may be "natural" but there are a multitude of times and places where it would be entirely impractical or inappropriate to do so. Imagine the fun of trying to send a text via speech recognition on the bus while even just one person sitting near you is doing the same.
It is safe to say that no one should be running this software in its current form.
I'd say it's safe to say that the software shouldn't have done this without informing the user, but if someone wants to run it while knowing it is less secure than might otherwise reasonably be expected, who are you to tell them they shouldn't?
I disable selinux and in some cases I always log in as root, because I've decided that's the way I want to do things - I'd rather have the extra convenience than the extra security.
And on the basis of that you think that keyboards are becoming obsolete? I can get everywhere I need to go on foot, but that doesn't mean my car is going to become obsolete.
Keyboards will be around until something that's actually better comes along. Speech recognition certainly isn't it, no matter how good it gets.
Wait - when this 96% extinction happened, where the oceans acidic as they are now, or were they more acidic? As far as I can tell the substance of the article only talks rate of change of acidity, not the actual pH.
So, okay, the ocean pH is going down at a high rate. But that doesn't mean we're looking at the same kind of circumstances as occured 252m years ago.
if what you are saying is true, then we would be much taller than we are.
I think your counter-argument is intrinsically nonsensical, or at least circular. "Taller than we are"? We can only be as tall as we are right now. If we were taller, we wouldn't know it, we'd just think it was average, and we could all still be having the same conversation. Only you'd be saying "we would be much taller than we are (6'6 average)" instead of "we would be much taller than we are (5'6" average)".
It feels like there's a connection here to the similarly ridiculous notion of the Doomsday Argument.
giraffes got really tall, we could have too.
And maybe we will. We're still growing. So at some point we'd have to reach an average between early man and giraffe height. That's now.
Of course, giraffe height is pretty unlikely in reality. At some point the disadvantages of being tall will outweight the advantages (or new disadvantages will arise due to a change in the environment, lack of food for example) and we'll stop growing - maybe even start shrinking.
Slashdot Mirror
I really hate summaries that conspicuously omit the only thing that's really important in the whole article.
I hate comments that say that something is "really important" but don't explain what it is.
So, which "really important" thing in the article is missing from the summary?
I meant to say 'generate a result that could be misinterpreted as a valid message', thus maintaining the so-thought integrity of the fake message.
Ah, I think I see where my confusion arose. When you said "one of the escrow holders manages to create a fake key" you didn't mean that the key itself was fake - it would still be right key, according to the key escrow process - but that the original encryption could have been done in such a way as to cause the correct key to return a misleading result?
produces an entirely different result than the key holder's genuine key should generate.
Not if by "genuine key" you mean the key used by the proprietary device, and for which step it also generates the secret split key to allow decryption by agencies. Any decryption by either of those keys will result in the correct decryption.
If by "genuine key" you mean the key used to encrypt the data before it went through device encryption, well, then the escrow decryption process will still produce the correct intermediate (once-encrypted) text.
Almost nobody writes texts on the bus because most people take the bus during peak periods, when there's no ROOM to text.
That's not been my anecdotal experience.
And during off-peak hours, it's still too bumpy even on good roads.
Again, not my experience. Maybe we have better roads around here, or people are more tolerant than you think. But fine, let's shift the example - which was only supposed to be about people in close proximity - to the bus station where everyone's waiting for a bus. All the points I made still stand. Ambient noise, other people speaking, and privacy. Which people do want, even if they don't need it. Try asking a real person sometime. They're funny like that.
People in offices use headsets
Not in my office. Surely what you mean is that people in offices will have to use headsets, should your keyboard-less future come about. Which it won't. Keyboards are far and away a more practical input device than speech recognition, certainly for many years to come, if not indefinitely.
Being able to offer solutions to the problems that SR causes is not an argument in their favour when those problems don't exist with keyboards.
Call centers are an example.
In call centers headsets are a necessity. Given the choice I'm sure most people who work in one would prefer not to have to put up with the constant hub-bub. Call centers are also a great example of when speech recognition would be utterly impractical, since call center staff often need to both operate a computer and be free to speak to the person on the opposite end of the call at the same time.
People dealing with sensitive stuff should already have an office with a door that closes, so again, no need for a keyboard.
Again you're dictating how things must be done in your keyboard-less utopia, instead of making the argument for why people should prefer SR over keyboards. "Sensitivity" is also a sliding scale, like privacy, and there are plenty of things on it that'll fall into the "doesn't need a separate office, but doesn't need to be blurting it out loud" category for which a keyboard is more than adequate.
And for your private stuff, do it on your own time at home, same as everyone else.
There you go again, dictating when and how I should do something. Are there only two places people ever go, work and home? And why must they be expected to happily forego any notion of privacy when they're not in the latter?
If a single key can open all the doors
Not that it makes much difference to the substance of your point, but I don't think anyone's proposing literally a single key. It could (hypothetically, naively) be one split key per company, or per product, or batch of a product, or maybe even one split key per "real" key.
I might be missing something which rules out any or all of those possibilities, though.
Yes, but unless you have all the parts you can't get the whole story.
assuming encryption is stacked
What does "stacked" mean in this context?
one of the escrow holders manages to create a fake key
Not quite sure what you mean. Do you mean one of the escrow holders providing a fake "part" of the key, to be joined with the other real parts, thus producing a full, but false, key? Or producing an entire fake key by themselves?
when used to decrypt some given message, produces an entirely different result than the key holder's genuine key should generate
Isn't that what all incorrect keys do? Generate a result different from what the genuine key would produce?
ranging from black and brown to baby blue
If your baby's blue, you should probably get it to a doctor. It's a-sposed-ta be pink.
You are the one who brought up the bogus bus example, so don't blame me that it's weak as wet toilet paper.
How is it weak? And in what did I blame you for said imagined weakness?
You can point out how it might be slightly more awkward to text on a bus than when you're sitting at home until you're blue in the face, but that's not going to make up for the fact that speech recognition is going to create more problems that it would solve in that situation for most people. It'd have to work over any ambient noise, including other people trying to send text messages by babbling at their phones or - ugh! - having conversations. It would mean your texts becoming known to anyone within earshot, which plenty of people wouldn't much care for - or care to be subjected to - even if it is just about getting more milk.
If that's fine with you, that's your choice. But do you at least see that plenty of other people won't feel that way? Can you really not imagine any circumstance in which you might not be happy with using speech recognition to send a text for privacy reasons?
And if you believe that ANY of your messages are private, you're incredibly naive.
I never said I believed that. Leaping from there to "well, might as well just speak my every communication out loud, it makes no difference" doesn't make a whole lot of sense though.
Some database somewhere might well have a list of all the websites I visited last night, but that doesn't mean I'd be fine with the idea of having the list automatically posted to a website for anyone to browse.
The masses will always choose convenience over security.
Really? Because I'd find it really convenient not to need a key to open my front door or start my car.
But if you're that paranoid about your text messages being public
What's paranoid about just not wanting strangers on a bus knowing your business, no matter how mundane?
To me it seems to be used to explain the unexplainable, much like the aether of former times.
The aether was a pretty reasonable postulation given the observations of the time.
It got disproved. Dark energy might. Or might not.
Again, have you ever tried to type on a moving bus?
Again, my only point is that a keyboard is generally a far more useful and usable input method than speech recognition for such a situation and many, many others.
Also, most of your messages don't need to be private
Who are you to decide that?
It'd still be easier and more private than trying to use speech recognition.
Most people don't work with computers with keyboards (think point-of-sale touchscreens, etc)
Pens aren't obsolete even though they're used far less in business since the invention of the typewriter.
So what's more natural for them than to just say your message and hit send?
It may be "natural" but there are a multitude of times and places where it would be entirely impractical or inappropriate to do so. Imagine the fun of trying to send a text via speech recognition on the bus while even just one person sitting near you is doing the same.
It is safe to say that no one should be running this software in its current form.
I'd say it's safe to say that the software shouldn't have done this without informing the user, but if someone wants to run it while knowing it is less secure than might otherwise reasonably be expected, who are you to tell them they shouldn't?
I disable selinux and in some cases I always log in as root, because I've decided that's the way I want to do things - I'd rather have the extra convenience than the extra security.
You can text without one on a smartphone.
And on the basis of that you think that keyboards are becoming obsolete? I can get everywhere I need to go on foot, but that doesn't mean my car is going to become obsolete.
Keyboards will be around until something that's actually better comes along. Speech recognition certainly isn't it, no matter how good it gets.
reduces the rate of swipe interpretation mistakes by 50.1% compared to the QWERY keyboard.
I think you accidentally a letter.
Never consumer ready
Never really meant to be.
Wake me when tape is reliable AND costs 10% of the $/GB of hard drive storage.
Yeah? Well, wake me up when they cost five precent of the $/GB of HDDs!
Arbitrary criteria FTW!
they claim can already deliver data speeds of 10 Gigabits per second using millimeter Wave (mmW) frequency bands of 73GHz.
Over what distance? And with what power requirements? How many concurrent users at what bandwidth?
The Last Time Oceans Got This Acidic This Fast
Wait - when this 96% extinction happened, where the oceans acidic as they are now, or were they more acidic? As far as I can tell the substance of the article only talks rate of change of acidity, not the actual pH.
So, okay, the ocean pH is going down at a high rate. But that doesn't mean we're looking at the same kind of circumstances as occured 252m years ago.
The site also targets unidentified "John Does"
It's the suit which is also targetting unidentified John Does, not the site.
Although there aren't any high-profile cases of 3D printed weapons being used in crimes in the country
"High-profile" might be a bit redundant. I doubt any crime involving a 3D printed weapon will get treated as "low profile" for a while yet.
Or it could just be the author's fancy way of saying "I haven't heard of any such cases."
Malware is something computer users are now more aware of than ever.
You might say we're... *sunglasses* mal-aware of it.
YEEEAAAAH!
"No decent person," write Geoffrey Manne and Ben Sperry in a special issue of Reason, "should be *for* net neutrality."
No decent person? You might want to try not insulting people if you're trying to win them round to your point of view.
if what you are saying is true, then we would be much taller than we are.
I think your counter-argument is intrinsically nonsensical, or at least circular. "Taller than we are"? We can only be as tall as we are right now. If we were taller, we wouldn't know it, we'd just think it was average, and we could all still be having the same conversation. Only you'd be saying "we would be much taller than we are (6'6 average)" instead of "we would be much taller than we are (5'6" average)".
It feels like there's a connection here to the similarly ridiculous notion of the Doomsday Argument.
giraffes got really tall, we could have too.
And maybe we will. We're still growing. So at some point we'd have to reach an average between early man and giraffe height. That's now.
Of course, giraffe height is pretty unlikely in reality. At some point the disadvantages of being tall will outweight the advantages (or new disadvantages will arise due to a change in the environment, lack of food for example) and we'll stop growing - maybe even start shrinking.
Being tall requires more food, for instance
which can be raided from the higher cupboards with impugnity.
I've nearly had perfect scores on multiple occasions
Exactly. Nearly. Gah!