The linux kernel's recent UBIFS flash support is I believe separated into 2 distinct layers. There's a layer for logical to physical address translations with wear-leveling and free space tracking (UBI). And a separate layer for organising the storage of the filesystem within those used blocks while keeping stored data in block sizes that match the underlying physical media and re-writing the whole block at once.
I think that kind of abstraction is useful enough for the OS, potentially with the UBI layer provided by a hardware device.
When you delete data, you are deleting little bits spread all over the physical drive.
The biggest problem is that a delete in most filesystems simply marks the space in the index on the device as free. However most filesystems leave the deleted data in place without writing anything over the top until that space is re-allocated. Hard disks don't typically need to know which sectors of the physical storage are actually in use. If you tell an SSD that this block is no longer required it can start erasing the physical chips and add them to the internal free list ready for the next data to be written.
Ideally filesystems will need to be modified so they are aware of the different characteristics of SSD's.
It is possible to add features to a product that everyone purchasing it gets. What I suppose I meant was that manufacturers could be selling fully capable phones to Sprint and Verizon that also have the changes they need to access their network. And they could be selling the exact same phones to everyone else at the same time.
In an imaginary world where all manufacturers respected the spirit of the standards this kind of chicanery wouldn't be happening.
One of the problems with Sprint and Verizon is that the radio has to be specifically designed for them which mean you can only use a phone that they sell directly
And the manufacturers are enabling this behaviour. If Sprint and Verizon didn't have your help we wouldn't be in this mess. Sometimes you have to say "NO".
I dunno. Sure it's a big change, and it's going to be annoying to relearn the interface. But I think if you had to learn the old 2003 interface and the new 2007 interface from scratch you wouldn't be as annoyed.
What they really needed was a quick tutorial to point out where all the old features have moved to. And they need to make that big circle look more like a button than a window decoration or add some annoying first use animation to point out that you need to click on it to do anything.
He has tried prosthetics a couple of times, but they were always more hassle than they were worth. He'd lose tactile feedback and dexterity. Plus these days you need to swap from mouse to keyboard fairly quickly, having a pointing device attached to his arm would probably be more annoying to deal with.
Though talking about my dad in this topic seems a bit unfair. I don't think anyone who's met him would call him disabled. The only things he's incapable of doing by himself are fiddly things he can't reach, like tying a necktie.
My father is an amputee, he lost both his hands when he was about 6. His left arm has about 4 inches after the elbow, and his right ends at the elbow.
In the early 8-bit hobby computing era he gave up his teaching job and started working as an analyst / developer. He types on a normal keyboard by holding a pen between his arms. Sometimes using his left elbow on Shift / Control keys.
However he is far more productive than most of the able bodied developers he works with because he's written so many macros in vim to automate just about everything.
I think the best question to answer when reviewing code is "Can I maintain this code if it needs to be modified?". Depending on the style of the developer you are reviewing, this could take seconds or hours to establish.
Yep. Once you've got DNSSEC you can publish a self signed cert in your DNS records (or public key or whatever standard people can agree on). Then you just need client support to fetch the details from DNS when connecting to the host over SSL.
I dunno, dogs have been trained to smell cancer. They might be able to smell the difference in the plastic of the disc or the printing if you gave them enough training. My guess is they gave the dogs plenty of examples of the legit discs and pirate discs they were trying to distinguish between. Who knows which smell they used to distinguish between them.
I dunno. In Australia they have recently begun random driver testing for a couple of "recreational" drugs. There's been all kinds of media advertising to warn you of the risk of being caught.
Yeah, but we can publish public keys in DNS for use in end to end encryption, or authentication. If the chain of trust starts at the root of DNS and flows down all the way to the connection to the web server, how are you planning to spoof it?
I found a way to beat the AI I wrote by forcing the AI to block then placing a piece on top of it for a double win. But my AI just weighted positions and didn't look ahead at all.
While I think of it... I think the most effective way to avoid a similar crisis in future is to limit the security value of an asset to say 9 years of its earning potential. Eg a bank should only be able to loan an amount equal to 9 years of rent on a property, putting a ceiling on any housing asset bubble.
What they assumed, is that each security had X% chance of defaulting, and when you bundle them together those percentages can be multiplied together. Giving an infinitesimal percentage that the whole bundle will go bad at once.
They basically assumed that all the fundamentals of the economy (eg GDP growth, unemployment rates) would remain at current levels. That current rates of default would always apply.
Well guess what, all those mountains of debt we've been encouraging people to bury themselves under has finally dragged down the economy. Without all that debt fueled spending, GDP is dropping, unemployment is rising, and so are defaults across the board.
Yeah he gets pretty worked up. And the maths he tries to use is too complex for most economists to follow. But that's the point. Economists make far too many assumptions about the world that are just plain wrong. And when the world doesn't behave like their model they assume the world is wrong. They assume everything would have turned out ok if only people understood their models better.
Slashdot Mirror
Yes and No.
The linux kernel's recent UBIFS flash support is I believe separated into 2 distinct layers. There's a layer for logical to physical address translations with wear-leveling and free space tracking (UBI). And a separate layer for organising the storage of the filesystem within those used blocks while keeping stored data in block sizes that match the underlying physical media and re-writing the whole block at once.
I think that kind of abstraction is useful enough for the OS, potentially with the UBI layer provided by a hardware device.
Yeah, but I've been seeing that in Firefox 3.5 as well.
When you delete data, you are deleting little bits spread all over the physical drive.
The biggest problem is that a delete in most filesystems simply marks the space in the index on the device as free. However most filesystems leave the deleted data in place without writing anything over the top until that space is re-allocated. Hard disks don't typically need to know which sectors of the physical storage are actually in use. If you tell an SSD that this block is no longer required it can start erasing the physical chips and add them to the internal free list ready for the next data to be written.
Ideally filesystems will need to be modified so they are aware of the different characteristics of SSD's.
It is possible to add features to a product that everyone purchasing it gets. What I suppose I meant was that manufacturers could be selling fully capable phones to Sprint and Verizon that also have the changes they need to access their network. And they could be selling the exact same phones to everyone else at the same time.
In an imaginary world where all manufacturers respected the spirit of the standards this kind of chicanery wouldn't be happening.
One of the problems with Sprint and Verizon is that the radio has to be specifically designed for them which mean you can only use a phone that they sell directly
And the manufacturers are enabling this behaviour. If Sprint and Verizon didn't have your help we wouldn't be in this mess. Sometimes you have to say "NO".
I dunno. Sure it's a big change, and it's going to be annoying to relearn the interface. But I think if you had to learn the old 2003 interface and the new 2007 interface from scratch you wouldn't be as annoyed.
What they really needed was a quick tutorial to point out where all the old features have moved to. And they need to make that big circle look more like a button than a window decoration or add some annoying first use animation to point out that you need to click on it to do anything.
He has tried prosthetics a couple of times, but they were always more hassle than they were worth. He'd lose tactile feedback and dexterity. Plus these days you need to swap from mouse to keyboard fairly quickly, having a pointing device attached to his arm would probably be more annoying to deal with.
Though talking about my dad in this topic seems a bit unfair. I don't think anyone who's met him would call him disabled. The only things he's incapable of doing by himself are fiddly things he can't reach, like tying a necktie.
My father is an amputee, he lost both his hands when he was about 6. His left arm has about 4 inches after the elbow, and his right ends at the elbow.
In the early 8-bit hobby computing era he gave up his teaching job and started working as an analyst / developer. He types on a normal keyboard by holding a pen between his arms. Sometimes using his left elbow on Shift / Control keys.
However he is far more productive than most of the able bodied developers he works with because he's written so many macros in vim to automate just about everything.
I think the best question to answer when reviewing code is "Can I maintain this code if it needs to be modified?". Depending on the style of the developer you are reviewing, this could take seconds or hours to establish.
Or they could resolve that plot in about 5 seconds and return to earth. It's not that big a deal.
And you'll damage it after the warranty expires, and the 3GS with be worthless.
Which is one of the reasons I'm hoping clang can live up to its own hype.
You got a better idea?
Yep. Once you've got DNSSEC you can publish a self signed cert in your DNS records (or public key or whatever standard people can agree on). Then you just need client support to fetch the details from DNS when connecting to the host over SSL.
I dunno, dogs have been trained to smell cancer. They might be able to smell the difference in the plastic of the disc or the printing if you gave them enough training. My guess is they gave the dogs plenty of examples of the legit discs and pirate discs they were trying to distinguish between. Who knows which smell they used to distinguish between them.
Pffft. I taught myself the alphabet backwards in a couple of days just so I could screw with my kids. ZYX WVUT SRQP ONML KJIH G FED CBA.
I dunno. In Australia they have recently begun random driver testing for a couple of "recreational" drugs. There's been all kinds of media advertising to warn you of the risk of being caught.
Yeah, but we can publish public keys in DNS for use in end to end encryption, or authentication. If the chain of trust starts at the root of DNS and flows down all the way to the connection to the web server, how are you planning to spoof it?
I found a way to beat the AI I wrote by forcing the AI to block then placing a piece on top of it for a double win. But my AI just weighted positions and didn't look ahead at all.
Heck you can do green screen style processing with Javascript in real time.
While I think of it... I think the most effective way to avoid a similar crisis in future is to limit the security value of an asset to say 9 years of its earning potential. Eg a bank should only be able to loan an amount equal to 9 years of rent on a property, putting a ceiling on any housing asset bubble.
It's only a "black swan" if you ignore debt levels. Anyone who's been watching the growth of debt in western economies could see this crisis coming.
What they assumed, is that each security had X% chance of defaulting, and when you bundle them together those percentages can be multiplied together. Giving an infinitesimal percentage that the whole bundle will go bad at once.
They basically assumed that all the fundamentals of the economy (eg GDP growth, unemployment rates) would remain at current levels. That current rates of default would always apply.
Well guess what, all those mountains of debt we've been encouraging people to bury themselves under has finally dragged down the economy. Without all that debt fueled spending, GDP is dropping, unemployment is rising, and so are defaults across the board.
He has.
Yeah he gets pretty worked up. And the maths he tries to use is too complex for most economists to follow. But that's the point. Economists make far too many assumptions about the world that are just plain wrong. And when the world doesn't behave like their model they assume the world is wrong. They assume everything would have turned out ok if only people understood their models better.
Another great evangelist for change, an economist who foresaw this crisis; Steve Keen.