And a large corporation such as Dell would be able manage all the necessary registry hacks.
The only issue is if the user fills C:\ anyway--for instance, a heavy gamer, or someone who uses applications that save files under the application data folder by default. You could include a safe partition editor as part of the default application set, but that would be too much work for a major OEM to consider. It doesn't make them any money.
I know what you're talking about, just not the name.
However, that GUI doesn't replace the default Windows GUI; it's drawn on top of it. Windows would take a performance hit if you used any GUI besides its default, even if it were as minimal as twm.
Actually, the article referred to WinXP and Server2k3. It didn't mention anything about performance, though, and was strictly referring to out-of-the-box configurations. So if you only use your computer to play Freecell, choose Windows XP with your legacy hardware.
A Linux system on comparable hardware, though, could do just about anything you want, besides running OpenOffice or Eclipse. (Those are painfully slow even on my 512MB/Athlon 2800 system.)
It's a high school website. If it got twenty hits in an average day, that's more than I expect; and if any essential school functions were handled by that server, the IT people in charge should be shot.
So, there's the law. How does it apply? The kid used speech to ask others to use a computer to disrupt non-essential functions undertaken by an educational system, but not for an educational purpose. It doesn't.
Or you could use LILO or GRUB, have a backup Linux on a separate hard drive, modify a conf file, and reboot. No special hardware or software necessary. That's still a downtime of as much as ten minutes, but it's on an unreliable server anyway. Or else it's due to routine maintenance, so you can plan for the downtime.
Still, ten minutes just to fsck the damn thing....
At some point, there's a minimum bandwidth limit. There's one between a local routing station and your cable modem and goes to 5-20Mb/s in most cases. There's another at the server, usually no more than 0.5KB/s but sometimes as much as twice that.
So if you're pulling information from multiple servers (streaming BitTorrent sort of arrangement, let's say), you can potentially get up to 20Mb/s in. If you're only working with a single server, it's much less. Duh.
Now. A particular Western Digital SATA drive can handle 681Mb/s to disk and 150MB/s to host. If the operating system takes half that, you still have gobs of bandwidth within your computer to handle the stream, if you want to save it to your hard drive. If you put it all in a page file, that's not a bottleneck.
So what's the issue? RAM? It's getting cheaper, faster, and larger these days. You can't get more than eight gigs without going multiprocessor, perhaps, but even four should be plenty for any one task.
CPU speed? This would better be served by multiple cores than a faster CPU, and that's happening. (If you have two cores, one can handle decoding the video and the other can handle everything else, if video's that important.)
Operating systems? While the major desktop OSs don't take much advantage of multiple cores to my knowledge, they still perform adequately for streaming video.
Server bandwidth, then? Most streaming video I've seen has a transfer rate well under 100KB/s. Around 50KB/s is common for commercial sites. This is quite obviously the issue.
So, once we're all living in a world of bliss where everyone has truly unlimited bandwidth, we can worry about restructuring our OSs to handle octet streams. But how will we do it?
There are two options: allow the server to directly write to your screen, or handle it just like C++ i/o streams. And the latter is quite similar to C++ sockets, while the former is a script kiddie's wet dream.
True, but it's odd that every sequence appears twice. Insignificant, I'm sure.
Also, almost every sequence contains a repeated digit or pair of digits. Though there are 10**5 possible sequences, few of them would be used.
This one-time pad might well be based on something other than addition. Alternately, it might simply use very large constants. With a larger text, expectation would play a factor. If I were implementing the one-time pad, I'd probably produce each digit individually using a large modulus and then taking the most significant digit.
Anyway, this all reminds me of the Conet Project (Irdial). Check it out; it's Creative Commons.
Why search for your files every time? If you know where your files are, there's no issue, and no need to wait for the shell to search through all your files. (And I thought it was bad when GNOME file-chooser wanted to load a listing of my/usr/bin directory every time I manually entered a path inside/usr/bin. I'm almost tempted to gut Firefox so it uses a different system.)
Google's current search engine relies on hyperlinks as much as text searches. It would do poorly with any system without hyperlinks. So you'd have a crappy search system to find files rather than just using a bit of organization. Or you'd have a new system that worked better for monolithic files, but that would take more effort.
Sons in college who visit once a year and find virus definition files that are 363 days old, and some sort of msbb process that's sucking up fully half the available RAM. And sons who have a very strong urge to replace the OS with their favorite Linux distribution.
And then those parents wonder why their computer was getting slow in the interim.
Microsoft's income based on IE is negligible if not negative. It costs them absolutely nothing for Dell to preinstall Firefox; the only cost is altering the Windows Update site to allow for both browsers--which they're doing anyway.
Eventually, MS might reach a deal with the Mozilla Foundation to bundle Firefox with Windows, saving them the expense of maintaining IE. Then everyone wins--the users get a better browser and a more secure OS; Microsoft's reputation improves all around; Firefox spreads further; and website designers get redesign their sites to support standards and nix ActiveX controls.
Google's products look and behave better than Microsoft's for a number of reasons, but the most important: 1. They're generally a lot simpler. An email service and a search engine and a number of media hosting services are much simpler than an operating system and can be implemented easily in high-level languages, without paying much attention to hardware issues.
2. They don't have issues with client configuration and client hardware. Or third-party applications. Especially ones that run as system.
3. Someone else handles much of the work--rendering and such. Google just has to make sure it does so properly for each project.
If Google tried making an operating system, though, they'd have to restructure the company to accept larger, more hierarchical teams. And they'd have more bugs than Microsoft in the first few releases--they have little experience with client operating systems.
I do hope you're only referring to a configuration script, not something that runs every time I start X. That way, I have final control over the settings, in case of odd hardware setups.
I've been using Windows for years. First they started with numbers after the name, then they put "Me!" instead, then something about experience points. Now that's not enough, and they want prefixes as well.
There are algorithms for relaying information without letting anyone who can intercept less than n-1 portions out of n find the content of the message. One that I'm familiar with involves sending a series of random numbers that all sum to a multiple of an arbitrary, very large number, and adding your data in numerical form to one of the numbers; this assumes that a number of separate connections are available that are substantially different.
There are other methods available; search for 'zero-knowledge proofs'.
There may be more unique arrangements of bits to collect, but as for content, most of the Internet's rather slim. Mostly it's personal websites, livejournal and its analogues, sales, and aggregators. On the side, you get some regular news sites.
I've found some informative and interesting personal websites, including nuwen.net and a fair bit about amateur linguistics. Other than that, if you want more content, the best place to look is universities.
It's not more content, generally, just more media.
And have it flash the BIOS with 0's as its first action, then force reboot after spreading. That's data loss and hardware loss. Unless we start hot-swapping motherboards.
Slashdot Mirror
If you wait long enough, someone might bring it back to life.
Or you could try it, if you wanted.
And a large corporation such as Dell would be able manage all the necessary registry hacks.
The only issue is if the user fills C:\ anyway--for instance, a heavy gamer, or someone who uses applications that save files under the application data folder by default. You could include a safe partition editor as part of the default application set, but that would be too much work for a major OEM to consider. It doesn't make them any money.
I know what you're talking about, just not the name.
However, that GUI doesn't replace the default Windows GUI; it's drawn on top of it. Windows would take a performance hit if you used any GUI besides its default, even if it were as minimal as twm.
Actually, the article referred to WinXP and Server2k3. It didn't mention anything about performance, though, and was strictly referring to out-of-the-box configurations. So if you only use your computer to play Freecell, choose Windows XP with your legacy hardware.
A Linux system on comparable hardware, though, could do just about anything you want, besides running OpenOffice or Eclipse. (Those are painfully slow even on my 512MB/Athlon 2800 system.)
Off topic, I admit, but is there a decent PDF reader for *nix? [X|G|K]PDF doesn't handle some PDFs, and Adobe for Linux is both ugly and slow.
It's a high school website. If it got twenty hits in an average day, that's more than I expect; and if any essential school functions were handled by that server, the IT people in charge should be shot.
So, there's the law. How does it apply? The kid used speech to ask others to use a computer to disrupt non-essential functions undertaken by an educational system, but not for an educational purpose. It doesn't.
Or you could use LILO or GRUB, have a backup Linux on a separate hard drive, modify a conf file, and reboot. No special hardware or software necessary. That's still a downtime of as much as ten minutes, but it's on an unreliable server anyway. Or else it's due to routine maintenance, so you can plan for the downtime.
Still, ten minutes just to fsck the damn thing....
Forget that--how long is it before the Mozilla foundation has to disable all extension functionality because it's a vehicle for murder?
Especially since Linux is barely fifteen years old!
At some point, there's a minimum bandwidth limit. There's one between a local routing station and your cable modem and goes to 5-20Mb/s in most cases. There's another at the server, usually no more than 0.5KB/s but sometimes as much as twice that.
So if you're pulling information from multiple servers (streaming BitTorrent sort of arrangement, let's say), you can potentially get up to 20Mb/s in. If you're only working with a single server, it's much less. Duh.
Now. A particular Western Digital SATA drive can handle 681Mb/s to disk and 150MB/s to host. If the operating system takes half that, you still have gobs of bandwidth within your computer to handle the stream, if you want to save it to your hard drive. If you put it all in a page file, that's not a bottleneck.
So what's the issue? RAM? It's getting cheaper, faster, and larger these days. You can't get more than eight gigs without going multiprocessor, perhaps, but even four should be plenty for any one task.
CPU speed? This would better be served by multiple cores than a faster CPU, and that's happening. (If you have two cores, one can handle decoding the video and the other can handle everything else, if video's that important.)
Operating systems? While the major desktop OSs don't take much advantage of multiple cores to my knowledge, they still perform adequately for streaming video.
Server bandwidth, then? Most streaming video I've seen has a transfer rate well under 100KB/s. Around 50KB/s is common for commercial sites. This is quite obviously the issue.
So, once we're all living in a world of bliss where everyone has truly unlimited bandwidth, we can worry about restructuring our OSs to handle octet streams. But how will we do it?
There are two options: allow the server to directly write to your screen, or handle it just like C++ i/o streams. And the latter is quite similar to C++ sockets, while the former is a script kiddie's wet dream.
True, but it's odd that every sequence appears twice. Insignificant, I'm sure.
Also, almost every sequence contains a repeated digit or pair of digits. Though there are 10**5 possible sequences, few of them would be used.
This one-time pad might well be based on something other than addition. Alternately, it might simply use very large constants. With a larger text, expectation would play a factor. If I were implementing the one-time pad, I'd probably produce each digit individually using a large modulus and then taking the most significant digit.
Anyway, this all reminds me of the Conet Project (Irdial). Check it out; it's Creative Commons.
Why search for your files every time? If you know where your files are, there's no issue, and no need to wait for the shell to search through all your files. (And I thought it was bad when GNOME file-chooser wanted to load a listing of my /usr/bin directory every time I manually entered a path inside /usr/bin. I'm almost tempted to gut Firefox so it uses a different system.)
Google's current search engine relies on hyperlinks as much as text searches. It would do poorly with any system without hyperlinks. So you'd have a crappy search system to find files rather than just using a bit of organization. Or you'd have a new system that worked better for monolithic files, but that would take more effort.
Money is less important than lives, though.
Unless you're a particularly callous insurance mogul.
Sons in college who visit once a year and find virus definition files that are 363 days old, and some sort of msbb process that's sucking up fully half the available RAM. And sons who have a very strong urge to replace the OS with their favorite Linux distribution.
And then those parents wonder why their computer was getting slow in the interim.
Microsoft's income based on IE is negligible if not negative. It costs them absolutely nothing for Dell to preinstall Firefox; the only cost is altering the Windows Update site to allow for both browsers--which they're doing anyway.
Eventually, MS might reach a deal with the Mozilla Foundation to bundle Firefox with Windows, saving them the expense of maintaining IE. Then everyone wins--the users get a better browser and a more secure OS; Microsoft's reputation improves all around; Firefox spreads further; and website designers get redesign their sites to support standards and nix ActiveX controls.
Where's the problem?
Google's products look and behave better than Microsoft's for a number of reasons, but the most important:
1. They're generally a lot simpler. An email service and a search engine and a number of media hosting services are much simpler than an operating system and can be implemented easily in high-level languages, without paying much attention to hardware issues.
2. They don't have issues with client configuration and client hardware. Or third-party applications. Especially ones that run as system.
3. Someone else handles much of the work--rendering and such. Google just has to make sure it does so properly for each project.
If Google tried making an operating system, though, they'd have to restructure the company to accept larger, more hierarchical teams. And they'd have more bugs than Microsoft in the first few releases--they have little experience with client operating systems.
I do hope you're only referring to a configuration script, not something that runs every time I start X. That way, I have final control over the settings, in case of odd hardware setups.
I've been using Windows for years. First they started with numbers after the name, then they put "Me!" instead, then something about experience points. Now that's not enough, and they want prefixes as well.
Screw the bastards. I'm going with Linux.
It's probably not that bad. Most likely, there wouldn't have been anything in its place, just a longer delay.
Not to scale, but recall that the signal's stronger--NASA wants to hear what the Mars orbiter has to say, after all.
And the backup battery.
Hell, just get a Faraday cage and shove it in there when you don't want to be tracked.
There are algorithms for relaying information without letting anyone who can intercept less than n-1 portions out of n find the content of the message. One that I'm familiar with involves sending a series of random numbers that all sum to a multiple of an arbitrary, very large number, and adding your data in numerical form to one of the numbers; this assumes that a number of separate connections are available that are substantially different.
There are other methods available; search for 'zero-knowledge proofs'.
There may be more unique arrangements of bits to collect, but as for content, most of the Internet's rather slim. Mostly it's personal websites, livejournal and its analogues, sales, and aggregators. On the side, you get some regular news sites.
I've found some informative and interesting personal websites, including nuwen.net and a fair bit about amateur linguistics. Other than that, if you want more content, the best place to look is universities.
It's not more content, generally, just more media.
Viruses talk to you!
And have it flash the BIOS with 0's as its first action, then force reboot after spreading. That's data loss and hardware loss. Unless we start hot-swapping motherboards.