Uhm.... i386+ chips all have two priority bits, for a total of four priority levels. I don't know of a single OS that uses more than ring0 and ring3, though. Even NT/2K/XP, which is supposed to be a microkernel, runs at either ring0 or ring3.
One obvious advantage to a microkernel is that portions of the kernel can then be (carefully) swapped out to virtual memory on disk, whereas monolithic kernels generally cannot (just don't page out the pager, or disk I/O system!). This is good, because those microkernels tend to run significantly larger than a comparable monolithic kernel. Microkernels tend to be handle to handle real-time scheduling constraints better than a monolithic kernel can because of the nature of the thing.
Now, aside from SCO UNIX products, I don't know of any OS that is completely monolithic these days. Linux has kernel threads and run-time installable modules which are rather un-monolithic from a purist stand point. And from the microkernel camp, we have NT and OSX, which runs their entire kernel in the highest priority level for performance reasons, and in NT's case, have a rather heavy microkernel. Hybrids, the whole lot of 'em.
Proxy servers can provide different kinds of security than simple filtering can. In particular, it can provide authentication. A packet filter has no way to tell if a particular packet came from an authenticated source or not. A web-proxy, an gather information about the currently logged on user, and which sites this user visits, or can deny access to certain users to certain sites, or at certain times of the day. An SMTP proxy can provide additional security to the network by doing anti-virus scanning and spam filtering.
At work, we use a web proxy server to log access to the internet. It prevents most applications that aren't specificly designed to handle the proxy from accessing the internet without authorization.
Wouldn't multi-sided practically be part of the definition of a die? A single sided die would be pretty silly -- you might as well flip a coin and play your game in binary.
Uh, right. Buying macintoshes with OSX with be MUCH cheaper than paying that $379 for RHEL... It's not cheaper if the up-front investment far outweighs the licensing of the software.
Unsanitary grinder, or butcher. If the carcass wasn't clean before butchering, it doesn't matter how clean the grinder was. And if the carcass wasn't clean, every cut the butcher makes with his knives spreads the E.coli bacteria across the surface of the meat. Then there's poultry. Even raw eggs can contain salmonella, which is to say nothing about the bird itself. Undercooked poultry has been the cause of many unpleasant post-thanksgiving dinners for many people.
Modern butchering practices are largely responsible for this danger, but are you going to go into a field, kill and animal, and then attempt to eat the meat from it's bones yourself? Nearly every part of an animal gets used when it is butchered, which is something you likely wouldn't do if you had killed the animal yourself for your own use.
Oh, it may do more damage, especially if new.net decides to change their program in order to sabotage the 'unauthorized' uninstallers. It wouldn't be the first time adware/spyware/bundleware intentionally tried to break anti-adware software. I remember a story about one software package intentionally damaging Ad-Aware's reference file in order to prevent detection, and another where installation of one software package uninstalled Ad-aware (read: deleted it). In both cases, the authors believed they were completely in the right, since Ad-aware had the capability of 'modifying' their programs as they were distributed. The guy who deleted ad-aware as part of his install process eventually caved, and changed it so his software refuses to install if ad-aware is installed.
.NET 1.0 could've just sucked because Microsoft can't release anything that works with version 1.0... It's usually version 3.0 or so that works well enough to use. Even when Microsoft tried to use that trend against itself with Windows NT 3.1, it failed, and wasn't really until Windows 2000 (NT 5.0) that it was really usable.
I doubt someone would sue them for real damage. More likely someone would sue to destroy them, and remove them from the picture. Now, I'm sure Microsoft would never do this (something about anti-trust), but if some other entity was sinking (e.g. SCO?), they might consider something like this.
I'd like to know how much first and second year calculus has changed in the last ten years. Third and fourth year, I could see changing... maybe. But introductory courses? I'm guessing the life on these books should be more than 1 or 2 years.
Perhaps you should open your eyes. Supplying weapons to religious extremists in Afghanistan. The Arabic nations (and South American, for that matter) were treated poorly by both the Soviets and Americans. Both sides used them as pawns against the other, neither really caring what happens to them. As soon as the soviets pulled out of Afghanistan, American support subsided as well. The entire cold war was nothing but a cruel game played out by the world leaders using otherwise independent countries as the pawns of the game. These are no "Soviet lies", but the difficult truth that both sides ought to face up to. Few of the countries that were unlucky enough to be involved in these games managed to obtain democratic government after the players were done with them.
The world, and indeed, your government is not so black and white as you paint it. The US government has done some good things, and they have done some bad things. Perhaps we shall be lucky enough to count the involvement in Iraq and Afghanistan as one of the good things, but only time will tell. Will the US be able to get these countries back on their own two feet again, with democratic independence? I certainly can't say.
It really depends on how quirky and problematic the application that would otherwise need to reside on the client system is. Companies with fragile fat clients may find it considerably cheaper to maintain it on a server, and use thin clients instead. Cost savings for this doesn't come down to licensing (which they're probably paying regardless), but support costs. Needing just one less support person would save the cost of that entire person's wages, as well as any the costs of downtime that may have been associated with fixing those client workstations constantly.
He also noted that it was a live server. Did he reboot the server to make sure the cache was flushed? Did both systems use the same filesystem and options (no dir hashes for ext3, for example). I truly hope he didn't just copy the files to the server, then immediately open it with his mail client, or the results are utterly useless. The results just seem fishy. A SCSI drive with 30% less throughput, and 40% better seek times can't really outperform an IDE drive by 600%. If the SCSI drive had out performed the IDE one by 1.4 minutes (5.6 minutes total), things would still be making sense. There's already benchmarks that prove IDE is slower than SCSI, but none of them ever claim that IDE is 1/6th the speed of SCSI.
This 'benchmark' doesn't really qualify as a benchmark. It's more of an observation.
It really doesn't matter if Linux was more difficult or easier to use than Windows. These days, people don't buy Windows because it's easy to use. They buy Windows because everyone else uses Windows, and all the programs they might want to use are made for Windows. If ease of use was the only factor, you'd think Mac's would be all the rage. The fact of the matter is, most people would have trouble using their computer, be it Macintosh, Windows or Linux. Computers are complex devices, and not all complexity can be hidden from the novice user.
You might think that something as simple as installing a new program from CD would be something that the average user could accomplish. The reality is, most users are scared of that, and won't do it. Even the "Compact, Normal, Full, or Custom" install choice is enough to stop many people. After all, how do they know which type of install they're going to need, and how do they change it if they choose wrong now? One might even say that software installation on Linux is MORE user friendly. There's just a blackbox RPM (or.deb) that does its thing when you choose to install it, but instead we have people trying to bring the Windows installation method to Linux, all the while claiming it's more user friendly (when it's only user familiar).
You missed a point, too. Because of things like this, companies have an extra excuse for not providing any documentation so people can make Linux drivers. In specific, people have been trying to get 802.11g drivers released for Linux, and an application like this may undermine their efforts.
Linux programs running on FreeBSD aren't really virtualized. They are run their instructions natively with alternate libraries, and a slightly different program loader.
Such a driver could still wedge the bus, or initiate some DMA operation that overwrites some critical memory. Look at XFree86, for example. I can't directly cause a kernel panic, but it can cause portions of the system to get into a state that can't be corrected without resetting the system to bring it back to a known state.
Okay, now repeat after me: "Anything running in kernel mode (ring 0) has potential to completely crash the machine". That means ANY driver that runs in kernel mode can crash the machine, video, network, sound, anything. Printer drivers are usually immune to crashing the machine except via interactions with bugs in other things that run in kernel mode (such as win32k.sys). For user-mode applications, that is very true. Bugs in user-mode applications and drivers (such as printer drivers) should never be able to crash Windows, much the same way they can't crash Linux, or any other operating system that runs fully in protected mode.
Why would driver developers use remote debugging consoles on NT, if nothing can crash the kernel except for hardware? Certainly there'd be no point, right?
We'd prefer ICANN to become more of a trade association that promotes the growth of the network rather than a regulatory body
Lets see. How many corporations who were granted a monopoly on a public resource do you see operating without a regulatory body governing them? Phone companies? Electric companies? Cable companies? Television and radio broadcasters? So long as the government is going to give out monopolies over resources that cannot be shared, regulation is a way of life for these companies who wish to compete in this sector. Maybe next time the contract comes up for renewal, ICANN could help Verisign by giving control over.com and.net to someone else. Verisign would be free to 'innovate' all they want, then.
I don't think I've seen any vendor that will guarantee 99.9999% uptime (52.5 minutes downtime per year). Most will only guarantee 99.999% uptime, but only if you use one of their clustered systems that have automatic failover (8.7 hours downtime per year). I could believe vendors offering 99.99% or 99.9% uptime for non clustered systems (87.6 or 876.5 hours of downtime per year). Anything beyond 99.99% uptime, you get into the realm of hardware failures causing downtime, and the guarantees usually expire the moment you install any non-approved software (including non-approved MS patches) on the system.
Sending out unsoliticed invoices is illegal in many places (mail fraud). If you want to get paid for misappropriation of patents, then you either negotiate with them directly, or you sue. Of course, IANAL, and I'm not from the US, either. Maybe in your corner of the world, they let people send out all kinds of unsolicited invoices for products not purchased, but I'd doubt it.
The RIAA had a chance to stop this ages ago. Napster should've been a wake-up call. my.MP3.com should've been a wake-up call. KaZaA should've been a wake-up call. Still they are sleeping. They could've at least tried online music distribution back when Napster was getting popular, but they didn't. Today, they're dabbling with online music distribution, but are unwilling to co-operate with each other enough so that real online music stores could become a reality. So, we have a half dozen online music stores, each with one or maybe two major labels each, all demanding a subscription to access material with heavy DRM locks. The public has already decided what format they want the music in, and are being ignored.
Personally, I don't care if Kazaa and services like it live or die, but I believe the recording industry is fully responsible for the situation they now find themselves in. They're going to make all the same errors they would've if they had started years ago, but now, the public's standards have been raised, so those mistakes will hurt more. Stores with tiny selections won't work, files that are so DRM encumbered as to be useless except for listening on a PC won't work, treating your customers like criminals won't work!
Slashdot Mirror
One obvious advantage to a microkernel is that portions of the kernel can then be (carefully) swapped out to virtual memory on disk, whereas monolithic kernels generally cannot (just don't page out the pager, or disk I/O system!). This is good, because those microkernels tend to run significantly larger than a comparable monolithic kernel. Microkernels tend to be handle to handle real-time scheduling constraints better than a monolithic kernel can because of the nature of the thing.
Now, aside from SCO UNIX products, I don't know of any OS that is completely monolithic these days. Linux has kernel threads and run-time installable modules which are rather un-monolithic from a purist stand point. And from the microkernel camp, we have NT and OSX, which runs their entire kernel in the highest priority level for performance reasons, and in NT's case, have a rather heavy microkernel. Hybrids, the whole lot of 'em.
Given that it's the release build for Windows XP, I'd say it's pretty official.
At work, we use a web proxy server to log access to the internet. It prevents most applications that aren't specificly designed to handle the proxy from accessing the internet without authorization.
Uh, right. Buying macintoshes with OSX with be MUCH cheaper than paying that $379 for RHEL... It's not cheaper if the up-front investment far outweighs the licensing of the software.
Modern butchering practices are largely responsible for this danger, but are you going to go into a field, kill and animal, and then attempt to eat the meat from it's bones yourself? Nearly every part of an animal gets used when it is butchered, which is something you likely wouldn't do if you had killed the animal yourself for your own use.
Fine. I want to see you eat a 1/4 cup of ground beef, and see how you feel two or three days later.
Oh, it may do more damage, especially if new.net decides to change their program in order to sabotage the 'unauthorized' uninstallers. It wouldn't be the first time adware/spyware/bundleware intentionally tried to break anti-adware software. I remember a story about one software package intentionally damaging Ad-Aware's reference file in order to prevent detection, and another where installation of one software package uninstalled Ad-aware (read: deleted it). In both cases, the authors believed they were completely in the right, since Ad-aware had the capability of 'modifying' their programs as they were distributed. The guy who deleted ad-aware as part of his install process eventually caved, and changed it so his software refuses to install if ad-aware is installed.
.NET 1.0 could've just sucked because Microsoft can't release anything that works with version 1.0... It's usually version 3.0 or so that works well enough to use. Even when Microsoft tried to use that trend against itself with Windows NT 3.1, it failed, and wasn't really until Windows 2000 (NT 5.0) that it was really usable.
I doubt someone would sue them for real damage. More likely someone would sue to destroy them, and remove them from the picture. Now, I'm sure Microsoft would never do this (something about anti-trust), but if some other entity was sinking (e.g. SCO?), they might consider something like this.
I'd like to know how much first and second year calculus has changed in the last ten years. Third and fourth year, I could see changing... maybe. But introductory courses? I'm guessing the life on these books should be more than 1 or 2 years.
The world, and indeed, your government is not so black and white as you paint it. The US government has done some good things, and they have done some bad things. Perhaps we shall be lucky enough to count the involvement in Iraq and Afghanistan as one of the good things, but only time will tell. Will the US be able to get these countries back on their own two feet again, with democratic independence? I certainly can't say.
It really depends on how quirky and problematic the application that would otherwise need to reside on the client system is. Companies with fragile fat clients may find it considerably cheaper to maintain it on a server, and use thin clients instead. Cost savings for this doesn't come down to licensing (which they're probably paying regardless), but support costs. Needing just one less support person would save the cost of that entire person's wages, as well as any the costs of downtime that may have been associated with fixing those client workstations constantly.
You poor baby! Everyone's picking on you! Don't forget that the US got second place in the Weaseliest Country category.
This 'benchmark' doesn't really qualify as a benchmark. It's more of an observation.
You might think that something as simple as installing a new program from CD would be something that the average user could accomplish. The reality is, most users are scared of that, and won't do it. Even the "Compact, Normal, Full, or Custom" install choice is enough to stop many people. After all, how do they know which type of install they're going to need, and how do they change it if they choose wrong now? One might even say that software installation on Linux is MORE user friendly. There's just a blackbox RPM (or .deb) that does its thing when you choose to install it, but instead we have people trying to bring the Windows installation method to Linux, all the while claiming it's more user friendly (when it's only user familiar).
A little OT, but the URL in your sig has expired.
You missed a point, too. Because of things like this, companies have an extra excuse for not providing any documentation so people can make Linux drivers. In specific, people have been trying to get 802.11g drivers released for Linux, and an application like this may undermine their efforts.
Linux programs running on FreeBSD aren't really virtualized. They are run their instructions natively with alternate libraries, and a slightly different program loader.
Such a driver could still wedge the bus, or initiate some DMA operation that overwrites some critical memory. Look at XFree86, for example. I can't directly cause a kernel panic, but it can cause portions of the system to get into a state that can't be corrected without resetting the system to bring it back to a known state.
Okay, now repeat after me: "Anything running in kernel mode (ring 0) has potential to completely crash the machine". That means ANY driver that runs in kernel mode can crash the machine, video, network, sound, anything. Printer drivers are usually immune to crashing the machine except via interactions with bugs in other things that run in kernel mode (such as win32k.sys). For user-mode applications, that is very true. Bugs in user-mode applications and drivers (such as printer drivers) should never be able to crash Windows, much the same way they can't crash Linux, or any other operating system that runs fully in protected mode.
Why would driver developers use remote debugging consoles on NT, if nothing can crash the kernel except for hardware? Certainly there'd be no point, right?
Lets see. How many corporations who were granted a monopoly on a public resource do you see operating without a regulatory body governing them? Phone companies? Electric companies? Cable companies? Television and radio broadcasters? So long as the government is going to give out monopolies over resources that cannot be shared, regulation is a way of life for these companies who wish to compete in this sector. Maybe next time the contract comes up for renewal, ICANN could help Verisign by giving control over .com and .net to someone else. Verisign would be free to 'innovate' all they want, then.
I don't think I've seen any vendor that will guarantee 99.9999% uptime (52.5 minutes downtime per year). Most will only guarantee 99.999% uptime, but only if you use one of their clustered systems that have automatic failover (8.7 hours downtime per year). I could believe vendors offering 99.99% or 99.9% uptime for non clustered systems (87.6 or 876.5 hours of downtime per year). Anything beyond 99.99% uptime, you get into the realm of hardware failures causing downtime, and the guarantees usually expire the moment you install any non-approved software (including non-approved MS patches) on the system.
Sending out unsoliticed invoices is illegal in many places (mail fraud). If you want to get paid for misappropriation of patents, then you either negotiate with them directly, or you sue. Of course, IANAL, and I'm not from the US, either. Maybe in your corner of the world, they let people send out all kinds of unsolicited invoices for products not purchased, but I'd doubt it.
Personally, I don't care if Kazaa and services like it live or die, but I believe the recording industry is fully responsible for the situation they now find themselves in. They're going to make all the same errors they would've if they had started years ago, but now, the public's standards have been raised, so those mistakes will hurt more. Stores with tiny selections won't work, files that are so DRM encumbered as to be useless except for listening on a PC won't work, treating your customers like criminals won't work!