No, sorry for my fuzziness. I meant run-time pluggability. I didn't mean it to be a general scheduler, just a generic way to call a run-time pluggable scheduler
I meant for something like a pointer to the scheduling function from the code in the interrupt vector. I used this technique in a kernel I wrote at the university. The main difference between what I described here and my implementation is that the scheduler itself was responsible for the context switch. Not a great idea in terms of maintainability, but it was quite fast.
A "mere mortal" coding in C++ would probably write a scheduler class instead. Having pointers to functions is an excellent idea, since it allows for an instantaneous jump to the requested scheduler function.
You don't even have to change the interrupt vector. You can use it to jump to a general rescheduling routine, which switches process by having the scheduler return the PID or thread ID of what it thinks should be running next. This way the scheduler is very easy to replace, and no modification of the interrupt vector is needed.
except maybe for Debian, it is the ONLY Linux distribution that won't install an X11 server by default.
Tried Arch Linux? It only installs the kernel and base packages (bash, init scripts, glibc) at it's bare minimum. Default is to also install some compilers (gcc, g++ etc.). X11 is not installed by default. You'd like it. I, as a former Slackware user, did.
...and I'm up to 3.5h of battery life (battery at 90% health) by using the powertop utility (try it!). Another great tip is that if you have an ATI card and are using the open source radeon driver with it, inserting the line
Option "DynamicClocks" "true"
into your "device" section in your xorg.conf makes it draw up to 2 watts less power (a lot in this context).
At my university (Uppsala, Sweden) we have a PC lab which still runs XP in the CS department. I don't see an upgrade to 7 any time soon because they just bought a bunch of Sun servers running Solaris 10. Most people here use thin clients (SunRay 2) to connect to the servers, and very few actually use the XP machines. At the math department, people have their own GNU/Linux installations on their machines.
We're probably safe security wise, since we've got a firewall protecting the XP machines, and they're not visible from the outside. The Solaris machines, however, are accessible via SSH.
The problem now is the lack of applications that enable end users to make benefit from having a powerful GPU. This will be the case until there's a standard API which works across multiple GPU architectures. Having both CUDA and OpenCL is one too many
...are just a way of sharing information. I think the <video> tag is a great idea, since it moves towards a standardised way to view videos on a page. Even though it's video, it's still information.
This, however, is not a clever way of "enhancing" web pages. We have the information we need, and we're satisfied. No need to put bells and whistles on it. If it were up to me (which it isn't), there would be no such thing as "web design". Web pages are not a fashion show, they're just means of sharing, displaying and publishing information. Let's keep it that way.
You're absolutely correct, but it should also be noted that IP addresses change and get recycled. Keeping logs of all IP leases is a difficult task, and failure is bound to happen at some point (DHCP server losing time/date, etc.). My point is that it's always hard to identify a person by a number that can change at any given time.
Let's wait one year from when they purchase their Windows-disabled Eees, when their machines are clogged with spyware and malware and the registry's broken and the little Atom CPU desperately tries to actually do some work whilst constantly swapping things in and out of RAM. Let's wait until then and tell them "It's Better With Windows!(tm)".
I say f*ck this! I'm not buying another Asus product in my life!
From this perspective, the Asus Eee is an interesting machine. It comes bundled with a system (modified Xandros)that you can very much treat as an appliance.
In fact, "appliance computing", as we may as well call it, is the intended target market for the Eee. You use the Eee as an appliance for surfing, word processing etc. Some people may call this limited, but lets' face it, it's all a Joe Sixpack needs in a small, convenient package.
You have a valid point. It's really hard to get any VPN connection to work under Linux at all, unless you use a GUI frontent.
Recent versions of the Network Manager applet (The network selection thing in Ubuntu, for those of you who don't know its name) claims to support VPN, albeit I've been unable to make it work even after nine years of using Linux and almost two years of CS studies.
If Linux is ever going to be widely used in corporate environments, easily configurable VPN support is a must.
When I was twelve I decided to try GNU/Linux. I was running Windows 98 at that time, and I didn't like it when it corrupted my partitions. Quite naturally I chose a distro which could be installed on top of a FAT file system, and I opted for Dragon Linux, which was slackware based. It was quite a mess. I never actually got X running with it, but I got doom running, though (albeit without sound).
The article states that these so called OAE:s can be recognized using hyper-sensitive microphones. This is a bit of a problem since phones tend to have microphones of rather poor quality compared to those required.
Furthermore, since the method requires sensitive microphones, it can't be expected to work at all, since there are a lot of noises around us which can affect the authentication process. Not to mention the signal quality required. I don't see this working over a telephone in a foreseeable future.
I'm not an American resident, so I'm wondering if what you pay for your connection is considered to be relatively good pricing? In my family, we're paying 359 SEK/month, wich is about $43/month, for 24 Mbit/s DSL with unlimited monthly traffic. We live in Sweden.
Slashdot Mirror
This makes me wish I had mod points :)
You can change it afterwards with tune2fs (for extfs family filesystems).
No, sorry for my fuzziness. I meant run-time pluggability. I didn't mean it to be a general scheduler, just a generic way to call a run-time pluggable scheduler
I meant for something like a pointer to the scheduling function from the code in the interrupt vector. I used this technique in a kernel I wrote at the university. The main difference between what I described here and my implementation is that the scheduler itself was responsible for the context switch. Not a great idea in terms of maintainability, but it was quite fast.
A "mere mortal" coding in C++ would probably write a scheduler class instead. Having pointers to functions is an excellent idea, since it allows for an instantaneous jump to the requested scheduler function.
You don't even have to change the interrupt vector. You can use it to jump to a general rescheduling routine, which switches process by having the scheduler return the PID or thread ID of what it thinks should be running next. This way the scheduler is very easy to replace, and no modification of the interrupt vector is needed.
a giant mirror on the Moon
I especially like this one. It has that "evil genius" feel about it.
Well, RHEL is a commercial product, so unless your organization feels like being criminal, it's not going to work in a legal context.
It is, however, a good idea in general.
except maybe for Debian, it is the ONLY Linux distribution that won't install an X11 server by default.
Tried Arch Linux? It only installs the kernel and base packages (bash, init scripts, glibc) at it's bare minimum. Default is to also install some compilers (gcc, g++ etc.). X11 is not installed by default. You'd like it. I, as a former Slackware user, did.
...and I'm up to 3.5h of battery life (battery at 90% health) by using the powertop utility (try it!). Another great tip is that if you have an ATI card and are using the open source radeon driver with it, inserting the line
Option "DynamicClocks" "true"
into your "device" section in your xorg.conf makes it draw up to 2 watts less power (a lot in this context).
Good luck!
whoooooosh!
At my university (Uppsala, Sweden) we have a PC lab which still runs XP in the CS department. I don't see an upgrade to 7 any time soon because they just bought a bunch of Sun servers running Solaris 10. Most people here use thin clients (SunRay 2) to connect to the servers, and very few actually use the XP machines. At the math department, people have their own GNU/Linux installations on their machines.
We're probably safe security wise, since we've got a firewall protecting the XP machines, and they're not visible from the outside. The Solaris machines, however, are accessible via SSH.
Damn, you beat me to it!
The problem now is the lack of applications that enable end users to make benefit from having a powerful GPU. This will be the case until there's a standard API which works across multiple GPU architectures. Having both CUDA and OpenCL is one too many
...the marketspeak made my ears bleed. The word "kernel" was mentioned *once* in the article linked to by the second link.
...are just a way of sharing information. I think the <video> tag is a great idea, since it moves towards a standardised way to view videos on a page. Even though it's video, it's still information.
This, however, is not a clever way of "enhancing" web pages. We have the information we need, and we're satisfied. No need to put bells and whistles on it. If it were up to me (which it isn't), there would be no such thing as "web design". Web pages are not a fashion show, they're just means of sharing, displaying and publishing information. Let's keep it that way.
You're absolutely correct, but it should also be noted that IP addresses change and get recycled. Keeping logs of all IP leases is a difficult task, and failure is bound to happen at some point (DHCP server losing time/date, etc.). My point is that it's always hard to identify a person by a number that can change at any given time.
OpenTTD > Transport Tycoon Deluxe ;)
Hello, friend! I am also quite pissed.
Let's wait one year from when they purchase their Windows-disabled Eees, when their machines are clogged with spyware and malware and the registry's broken and the little Atom CPU desperately tries to actually do some work whilst constantly swapping things in and out of RAM. Let's wait until then and tell them "It's Better With Windows!(tm)".
I say f*ck this! I'm not buying another Asus product in my life!
Damn, it feels good to vent!
From this perspective, the Asus Eee is an interesting machine. It comes bundled with a system (modified Xandros)that you can very much treat as an appliance.
In fact, "appliance computing", as we may as well call it, is the intended target market for the Eee. You use the Eee as an appliance for surfing, word processing etc. Some people may call this limited, but lets' face it, it's all a Joe Sixpack needs in a small, convenient package.
I wish I hadn't already used up my mod points, since this is one of the funnier things I've read today.
You have a valid point. It's really hard to get any VPN connection to work under Linux at all, unless you use a GUI frontent.
Recent versions of the Network Manager applet (The network selection thing in Ubuntu, for those of you who don't know its name) claims to support VPN, albeit I've been unable to make it work even after nine years of using Linux and almost two years of CS studies.
If Linux is ever going to be widely used in corporate environments, easily configurable VPN support is a must.
When I was twelve I decided to try GNU/Linux. I was running Windows 98 at that time, and I didn't like it when it corrupted my partitions. Quite naturally I chose a distro which could be installed on top of a FAT file system, and I opted for Dragon Linux, which was slackware based. It was quite a mess. I never actually got X running with it, but I got doom running, though (albeit without sound).
Oh noes! Ballmer's private army!
What sweet irony if they were programmed to attack Red Hat HQ.
If this video is removed due to copyright reasons, I'll then officially have lost all hope in mankind.
The article states that these so called OAE:s can be recognized using hyper-sensitive microphones. This is a bit of a problem since phones tend to have microphones of rather poor quality compared to those required.
Furthermore, since the method requires sensitive microphones, it can't be expected to work at all, since there are a lot of noises around us which can affect the authentication process. Not to mention the signal quality required. I don't see this working over a telephone in a foreseeable future.
I'm not an American resident, so I'm wondering if what you pay for your connection is considered to be relatively good pricing? In my family, we're paying 359 SEK/month, wich is about $43/month, for 24 Mbit/s DSL with unlimited monthly traffic. We live in Sweden.