Take away the video card so Jr can't see the hot action? Or sit there with the computer so Jr can be monitored at all times. Cancel internet access. Encrypt the hard drive so that Jr can't use the computer at all. Put a picture of Jesus over the monitor.
But with wired ethernet there's only one physical lan you can be attached to per interface. With wireless there may be many many possible lans. How does your card know which one? There is no automagical answer to this.
Also whenever radio comes into play things get a lot harder to figure out. It's black magic to programmers. With ethernet there's the ability to easily capture what's on the wire so you can debug. With wireless with encryption with interference from your neighbors' wireless ethernet... It gets very difficult.
Many of the ethernet drivers for Linux were from the days when DOS and Windows people didn't have networks, but Unix and Windows NT people did.
How do you configure/control the parameters that aren't common with wired ethernet? Wireless is a superset of wired, so while I don't doubt that it can work, I do wonder how some of the parameters are getting set.
I'm tired of people dogging multi-core cpus because they don't improve single threaded application performance. They do improve their performance if you have multiple applications (even ones that aren't that busy) running at the same time, or at the very very least other cores can handle interrupts so the one running your one application will.
There are a lot of improvements that can be made to many applications to better take advantage of multiple processors and multi-core processors, but many of those would help the performance of those applications on single core systems. In particular I'm thinking about the Mozilla family of programs and their poor responsiveness when rendering pages and how one bug triggered by one page brings the whole app down.
This is almost a pretty cool magic trick. The magician asks the audience member which of the two she wants to be true. Only thing is that the audience can't see the 2 billiards before it starts because billiards have static color and pattern.
He's been saying that there probably wouldn't be one for a long while now, but I think there should be a 3.0, and that it should break stuff. Anything that's old and a better way has been implemented should be removed. It wouldn't be a huge architectural change, but it would be userland visible so should have a new major number.
Nothing in this really sounds like a new idea except that using this method would have some benefit on the quantum level. It's just a balanced binary decision tree implemented as a (quantum, in this case) circuit such that leaf nodes are stored data and addresses are qbit streams.
Am I missing something?
If only the minimum bid were placed that would be about $15.33 per citizen. You'd never ever ever get every citizen to go in together to place a bid, but I would definitely bid $1000 to have a non-F-ed up mobile phone bandwidth. I don't know what exactly to do with it after bidding on it (and owning some minuscule fraction of it) but surely a great cell phone co-op wouldn't be unachievable.
Linked lists are just about the most common data structure in an OS kernel.
I know they are common in Linux.
Scheduling of processes is really just moving process nodes around within various queues which are typically made of linked lists.
File systems use linked lists for unused disk blocks.
Memory management uses them for free memory lists and other stuff I know I'm forgetting.
IO uses them for chunks of data that needs to be written to a devise or that is waiting to be read by a process or processed by a driver.
Just because they are the first data structure you learned does not mean that they aren't the be best for many circumstances. Both the code and the data overhead for them is very small, and they really do work well for stacks, queues, and dequeues -- all of which are very very common in OSes.
Slashdot Mirror
Jesus will forgive you, then He will adapt and evolve to deal with the perils of living in the trash can. Or being undead in the trash can.
Take away the video card so Jr can't see the hot action? Or sit there with the computer so Jr can be monitored at all times. Cancel internet access. Encrypt the hard drive so that Jr can't use the computer at all. Put a picture of Jesus over the monitor.
But with wired ethernet there's only one physical lan you can be attached to per interface. With wireless there may be many many possible lans. How does your card know which one? There is no automagical answer to this.
Also whenever radio comes into play things get a lot harder to figure out. It's black magic to programmers. With ethernet there's the ability to easily capture what's on the wire so you can debug. With wireless with encryption with interference from your neighbors' wireless ethernet... It gets very difficult. Many of the ethernet drivers for Linux were from the days when DOS and Windows people didn't have networks, but Unix and Windows NT people did.
How do you configure/control the parameters that aren't common with wired ethernet? Wireless is a superset of wired, so while I don't doubt that it can work, I do wonder how some of the parameters are getting set.
I blame the US's policy against outed homosexual in the military.
I'm tired of people dogging multi-core cpus because they don't improve single threaded application performance. They do improve their performance if you have multiple applications (even ones that aren't that busy) running at the same time, or at the very very least other cores can handle interrupts so the one running your one application will. There are a lot of improvements that can be made to many applications to better take advantage of multiple processors and multi-core processors, but many of those would help the performance of those applications on single core systems. In particular I'm thinking about the Mozilla family of programs and their poor responsiveness when rendering pages and how one bug triggered by one page brings the whole app down.
This is almost a pretty cool magic trick. The magician asks the audience member which of the two she wants to be true. Only thing is that the audience can't see the 2 billiards before it starts because billiards have static color and pattern.
And since most states collect extra taxes when there's a rebate because they tax the price it's sold for they don't care much about outlawing them.
And I feel like my ATi fanboyness has finally paid off!
He's been saying that there probably wouldn't be one for a long while now, but I think there should be a 3.0, and that it should break stuff. Anything that's old and a better way has been implemented should be removed. It wouldn't be a huge architectural change, but it would be userland visible so should have a new major number.
I'll bet quantum computers are listed on http://snopes.com/!
Nothing in this really sounds like a new idea except that using this method would have some benefit on the quantum level. It's just a balanced binary decision tree implemented as a (quantum, in this case) circuit such that leaf nodes are stored data and addresses are qbit streams. Am I missing something?
If only the minimum bid were placed that would be about $15.33 per citizen. You'd never ever ever get every citizen to go in together to place a bid, but I would definitely bid $1000 to have a non-F-ed up mobile phone bandwidth. I don't know what exactly to do with it after bidding on it (and owning some minuscule fraction of it) but surely a great cell phone co-op wouldn't be unachievable.
Linked lists are just about the most common data structure in an OS kernel. I know they are common in Linux. Scheduling of processes is really just moving process nodes around within various queues which are typically made of linked lists. File systems use linked lists for unused disk blocks. Memory management uses them for free memory lists and other stuff I know I'm forgetting. IO uses them for chunks of data that needs to be written to a devise or that is waiting to be read by a process or processed by a driver. Just because they are the first data structure you learned does not mean that they aren't the be best for many circumstances. Both the code and the data overhead for them is very small, and they really do work well for stacks, queues, and dequeues -- all of which are very very common in OSes.
There are just too many things going wrong for them right now. I'm just not sure if they can make it!