I couldn't remember, nor find it via google. But I found a few other 3.9 release stories that pointed there so I went with that. I was just surprised that TFS didn't include any actual details.
I live on the end of a fairly long piece of copper, for an urban area at least. The best ADSL speed I can get is around 4.5 Mbit. To get more than that speed what are my options? And if you're going to replace my copper with something, why not replace it with the best, most future proof technology?
$8,500 is a good investment for a technology that should last for the foreseeable future. Just don't ask each person to pay that full price right now.
"The ABS is constantly looking at ways it can simplify the website and enhance the user experience,"
While at the same time telling their actual developers to make it more difficult;
... generate a random number, which we append to the URL, to make it appear as if a complex
key is required. This is a pathetic attempt to discourage someone from downloading the ZIPs
directly (ie. without having to login), if they deduce the URL pattern.
Lets put a 20MB obsolete version of the facebook app on the system partition. Then as soon as I buy the phone I'll need to download another 20MB version from the play store...
I've been wondering for a while if it's worth building a UI focused, sandboxed virtual machine. Something that could be used to draw basic widget animations in response to mouse and keyboard events, while messages are passed back to the server your application is really running on.
So don't use your cell phone as a cell phone. Buy a pre-paid with no ID (if you can), use the data connection to open a VPN link, use whatever voice and IM protocols you want over the VPN link.
Assisted GPS in cell phones is typically just a way to download accurate satellite position information from the internet. Otherwise your receiver will have to download that data from the satellites themselves, which takes a significant amount of time. It doesn't have anything to do with remembering where the phone is, and I highly doubt it helps the cell phone company to determine your position.
The entire block chain contains all of the history of every piece of every coin. But to mine coins, and validate new blocks you don't really need to keep the entire history locally to validate new transactions, you only need to keep the details of transactions that haven't been consumed in another transaction. That's the sole purpose of the log anyway.
Sure the whole log should remain available online somewhere, but even that could be stored in a distributed way. The log itself is essentially tamper proof, since you can verify every single hash and signature if you want to.
While the total number of coins has a finite limit, each coin can be split into a huge number of transactions. Even when the active set of transactions grows too large, there will be solutions to improve the scalability of the system.
These are all solvable problems. There just hasn't been a pressing need to solve them yet.
And yet this massive increase in m0 has had no impact on any other measure of the money supply. Nor has it had much impact on the general health of the economy.
My primary point was that they cannot be manufactured *in response to* a growing market. They are created based on a well known and agreed upon formula, at a relatively predictable rate. And I believe, throwing more CPU power at the problem only increases your odds of getting a coin, it doesn't increase the overall rate that coins are produced.
And in response to some of the other replies, of course you could fork the bitcoin application and protocol, change the block validation rules to allow more coins to be created. But they wouldn't be "bitcoins" since the existing clients wont accept them.
Dollars *are* backed by debt. And that debt seemed to be ever increasing, at least up until 2007-ish when the housing market finally imploded.
The Fed doesn't really control or constrain the supply of money, though many economists still believe that they do. It's the double entry book-keeping rules of the banking industry that predominantly create and control the supply of money. A new loan creates both a future obligation and current spending power that didn't exist before. Sure the bank has to find a small amount of money to meet their deposit insurance, liquidity and capital requirements, but that's tiny in comparison to the value of new loans.
Since the level of debt is now such a huge factor in the economy, small accelerations and decelerations in the growth of debt have an enormous impact on the economy. And when everyone recently slammed on the debt brakes the economy practically died.
easy enough to make that the quantity can expand to support a growing market
Not even close. They are designed to be hard to make, to only be made at a pre-determined rate, and for new supplies to eventually run out. Bit-coins are designed to be limited in supply.
First problem, they have 1500 repo's and they want their mirrors to be able to delete them. But they built a script that deletes repo's if they are not found in a master list. Deleting because something is missing is always dangerous, something could disappear for any number of reasons. They should instead require an explicit delete command to be present.
Second problem, a mirror doesn't fsck and their master copies don't either. They should fsck each repo periodically, and verify all new patches being pushed before they appear on the master server.
They could also keep reflogs if they wish to guard against branches being deleted, but that wasn't a problem they actually faced.
The PS3 is not one version of one piece of software. It is a continually moving target that includes online services running as a back end. Every version must be cracked, in a very short time frame, to remain usable.
An almost frozen application wont move until the app's main thread is able to process the move message. XP and above allow the OS to take over after a few seconds, but that's still a very long delay that affects usability.
What Wayland is aiming for, is following the unix philosophy; "Do one thing and do it well". Wayland is a local desktop compositor, X11 is a legacy network transparency protocol, gtk+ Qt etc are application level UI frameworks. Sure it would be useful to lock some Wayland, Qt, gtk+ etc developers in a room until they build a new standard network protocol they can all agree on.
The Wayland developers are *not* trying to remove X11 support, heck most of the Wayland developers have worked on X11, they probably understand its strengths and weaknesses better than most people. X11 is a huge beast that used to do everything related to graphics, from the application layer to the display driver and the code is unmaintainable. They are replacing X11's crufty display rendering with Wayland so that the X11 layer can focus solely on the network transparency protocol. If we can improve the display rendering and the application frameworks, maybe we can even foster some better alternative to X11.
Aaron Schartz was caught in a computer closet with his laptop hooked into a network that had specifically denied him permission to connect to their system.
If he was caught in a TLA government agency in similar circumstances I'd agree with you. But this is a school, where you have to expect students to misbehave. The appropriate response should have been somewhere between a severe talking to, expulsion, or a lengthy discussion on how he should have randomised the requests to cover his tracks better;)...
It's obvious that the graphics driver is assuming that your panels are arranged on a flat plane, and that your eyes are some distance directly in-front of the center display.
Each screen really needs to be rendered from the point of view of a separate camera based on their actual orientation. This could probably be achieved using head tracking relative to each screen, then each perspective can be rendered correctly and doesn't need to be manually calibrated. Like this 5 year old demo.
Slashdot Mirror
I couldn't remember, nor find it via google. But I found a few other 3.9 release stories that pointed there so I went with that. I was just surprised that TFS didn't include any actual details.
http://www.h-online.com/open/features/What-s-new-in-Linux-3-9-1845705.html
I live on the end of a fairly long piece of copper, for an urban area at least. The best ADSL speed I can get is around 4.5 Mbit. To get more than that speed what are my options? And if you're going to replace my copper with something, why not replace it with the best, most future proof technology?
$8,500 is a good investment for a technology that should last for the foreseeable future. Just don't ask each person to pay that full price right now.
"The ABS is constantly looking at ways it can simplify the website and enhance the user experience,"
While at the same time telling their actual developers to make it more difficult;
... generate a random number, which we append to the URL, to make it appear as if a complex key is required. This is a pathetic attempt to discourage someone from downloading the ZIPs directly (ie. without having to login), if they deduce the URL pattern.
The ironing is delicious.
Lets put a 20MB obsolete version of the facebook app on the system partition. Then as soon as I buy the phone I'll need to download another 20MB version from the play store...
I've been wondering for a while if it's worth building a UI focused, sandboxed virtual machine. Something that could be used to draw basic widget animations in response to mouse and keyboard events, while messages are passed back to the server your application is really running on.
A better answer; False positive medical tests.
So don't use your cell phone as a cell phone. Buy a pre-paid with no ID (if you can), use the data connection to open a VPN link, use whatever voice and IM protocols you want over the VPN link.
Assisted GPS in cell phones is typically just a way to download accurate satellite position information from the internet. Otherwise your receiver will have to download that data from the satellites themselves, which takes a significant amount of time. It doesn't have anything to do with remembering where the phone is, and I highly doubt it helps the cell phone company to determine your position.
The entire block chain contains all of the history of every piece of every coin. But to mine coins, and validate new blocks you don't really need to keep the entire history locally to validate new transactions, you only need to keep the details of transactions that haven't been consumed in another transaction. That's the sole purpose of the log anyway.
Sure the whole log should remain available online somewhere, but even that could be stored in a distributed way. The log itself is essentially tamper proof, since you can verify every single hash and signature if you want to.
While the total number of coins has a finite limit, each coin can be split into a huge number of transactions. Even when the active set of transactions grows too large, there will be solutions to improve the scalability of the system.
These are all solvable problems. There just hasn't been a pressing need to solve them yet.
If you would believe in a god when given evidence, doesn't that make you agnostic?
And yet this massive increase in m0 has had no impact on any other measure of the money supply. Nor has it had much impact on the general health of the economy.
"Hey all you bitcoin miners, update to this software version so we can devalue all of your coins". Yeah, that's totally going to work...
he claimed that students at Galileo Academy, who had completed a web page authoring subject, had difficulty with the HTML image tag
My primary point was that they cannot be manufactured *in response to* a growing market. They are created based on a well known and agreed upon formula, at a relatively predictable rate. And I believe, throwing more CPU power at the problem only increases your odds of getting a coin, it doesn't increase the overall rate that coins are produced.
And in response to some of the other replies, of course you could fork the bitcoin application and protocol, change the block validation rules to allow more coins to be created. But they wouldn't be "bitcoins" since the existing clients wont accept them.
Dollars *are* backed by debt. And that debt seemed to be ever increasing, at least up until 2007-ish when the housing market finally imploded.
The Fed doesn't really control or constrain the supply of money, though many economists still believe that they do. It's the double entry book-keeping rules of the banking industry that predominantly create and control the supply of money. A new loan creates both a future obligation and current spending power that didn't exist before. Sure the bank has to find a small amount of money to meet their deposit insurance, liquidity and capital requirements, but that's tiny in comparison to the value of new loans.
Since the level of debt is now such a huge factor in the economy, small accelerations and decelerations in the growth of debt have an enormous impact on the economy. And when everyone recently slammed on the debt brakes the economy practically died.
easy enough to make that the quantity can expand to support a growing market
Not even close. They are designed to be hard to make, to only be made at a pre-determined rate, and for new supplies to eventually run out. Bit-coins are designed to be limited in supply.
First problem, they have 1500 repo's and they want their mirrors to be able to delete them. But they built a script that deletes repo's if they are not found in a master list. Deleting because something is missing is always dangerous, something could disappear for any number of reasons. They should instead require an explicit delete command to be present.
Second problem, a mirror doesn't fsck and their master copies don't either. They should fsck each repo periodically, and verify all new patches being pushed before they appear on the master server.
They could also keep reflogs if they wish to guard against branches being deleted, but that wasn't a problem they actually faced.
The PS3 is not one version of one piece of software. It is a continually moving target that includes online services running as a back end. Every version must be cracked, in a very short time frame, to remain usable.
And if 1,000,000 people join this group it will have 1,000,000 people in it.
An almost frozen application wont move until the app's main thread is able to process the move message. XP and above allow the OS to take over after a few seconds, but that's still a very long delay that affects usability.
What Wayland is aiming for, is following the unix philosophy; "Do one thing and do it well". Wayland is a local desktop compositor, X11 is a legacy network transparency protocol, gtk+ Qt etc are application level UI frameworks. Sure it would be useful to lock some Wayland, Qt, gtk+ etc developers in a room until they build a new standard network protocol they can all agree on.
The Wayland developers are *not* trying to remove X11 support, heck most of the Wayland developers have worked on X11, they probably understand its strengths and weaknesses better than most people. X11 is a huge beast that used to do everything related to graphics, from the application layer to the display driver and the code is unmaintainable. They are replacing X11's crufty display rendering with Wayland so that the X11 layer can focus solely on the network transparency protocol. If we can improve the display rendering and the application frameworks, maybe we can even foster some better alternative to X11.
I believe that stability is exactly what this team will be investigating next.
Aaron Schartz was caught in a computer closet with his laptop hooked into a network that had specifically denied him permission to connect to their system.
If he was caught in a TLA government agency in similar circumstances I'd agree with you. But this is a school, where you have to expect students to misbehave. The appropriate response should have been somewhere between a severe talking to, expulsion, or a lengthy discussion on how he should have randomised the requests to cover his tracks better ;)...
It's obvious that the graphics driver is assuming that your panels are arranged on a flat plane, and that your eyes are some distance directly in-front of the center display.
Each screen really needs to be rendered from the point of view of a separate camera based on their actual orientation. This could probably be achieved using head tracking relative to each screen, then each perspective can be rendered correctly and doesn't need to be manually calibrated. Like this 5 year old demo.