So you take your computer and never change the software....
There exist operating systems that don't require a reboot after installing software. I upgrade my software all the time, and I don't need to reboot.
Neither Macs nor *NIX boxes have the reboot fetish Windows seems to have.
The fact is that windows users f%&k around with their pc's adding any of 100,000's of application, uninstalling, doing whatever they want without worrying about backing up and what not. You do that on a linux box and see how stable it is after 6 months abuse.
Funny, I've installed a fair bit of software on each of my Linux boxes, and gee... still seems pretty darn stable! Can't say the same for my WinXP laptop from work. They've got so many SMS scripts running in the background that that thing goes to 80% load for an hour at a time. Never mind Norton AV, ConnectedTLM (backup software), and CyberShield.
I think a big factor here is that installing an app under Linux doesn't dump a bunch of crap in the registry, nor does it register a bunch of crap to run at startup, nor does it crap all over your desktop and launcher bar with short-cuts to the software. (Man, what ego these companies must have!) If an app just parked itself in its own little self-contained directory and left it at that, it wouldn't matter if you had 1 or 100,000 apps installed, except maybe to your diskspace.
In theory, if it's a large enough violation, it becomes a criminal act at the Federal level, as opposed to a civil liability. In that case, I'd imagine the Department of Justice can bring suit.
Under UNIX and Linux, if you delete a file that's in use, the directory entry goes away, but the inode remains in place until all the users of the file close their file handles. So, if you delete a running executable, it can still bring pages of itself into memory as needed, even though you can't see the file any more. Once the program exits, the inode gets freed and the disk space reclaimed.
So, you could do an rm -rf / and keep running for quite awhile, until, like you said, an existing task needs to open a file that's no longer there, or you try to fork some new task.
I should try that some time. I have a UML disk image set up somewhere... I could start up UML in "COW" mode (meaning any changes to the disk image get written to a separate "delta" file that I can throw away afterwards).
As for bitchchecker, it seems entirely plausible that remote disk-erasing software would work from higher-lettered disks to lower-lettered disks, simply because system files and programs tend to be on lower disk letters (C: usually), and huge MP3/movie/warez stores tend to be on the extra volumes. You'll notice Windows crashing pretty quick. You won't notice half your pr0n stash disappeared right away (unless you're really that big of a wanker). Starting from higher-lettered drives could also be even more devastating if the PC under attack has network drives mounted to those letters.
Back when I was in school on the campus UNIX system, I used to tell people about the default "slow mode" that most accounts start under. You see, since it was a time-shared machine, the default setting for a login session was a resource-conserving "slow mode." This ensured fairness on the machine, you see.
The trick, then, to speeding things up is to exit this mode. The command for doing so? exit $SLOW_MODE
Enough people do that, and sure enough, the machine speeds up... for everyone else!
(BTW, I'm aware now that said trick doesn't work on csh/tcsh users. This was an AT&T StarServer E running SVR4, though, so nearly all of us were running the original Bourne Shell.)
127/8 does not mean "127 or 128" in this context. It means "127.x.x.x with a netmask of 255.0.0.0" (8 bits of subnetwork identifier, 24 bits of host identifier). For instance, 10/8 is reserved as a non-routable address range, as is 192.168/16. That means you might use 10.x.x.x or 192.168.x.x addresses on your private networks. In the pre-CIDR days, a/8 was called a "Class A" network, a/16 was called a "Class B" network, and a/24 was called a "Class C" network.
I'm sure those of you who are Net.Gods will skewer me on slight terminology misuse, but I think I conveyed the basic idea.
That'd be 10,000 bloated drivers. Most drivers are FAR from being 100KB. A 1GB kernel source tree is highly unlikely to persist as "the way things are."
If the kernel source tree gets to be 100MB compressed, I think you will see the tarball fork into subtrees. If it doesn't, it's only because bandwidth to each of the developers has seen at least a 10x if not 100x growth from today's current status quo.
Keep in mind that the entire kernel history from the end of 2.4 through the current 2.6.x only takes 3 GB in the new "git" management system, and it doesn't do 'diff' style compression. (Though it does gzip individual files.) That 3GB is due to amplifying all the one-liners and other trivial commits that happened in the many hundred kernel intermediate versions between the two endpoints.
Anyway, veering noticeably ontopic, ndiswrapper hell only applies to network cards which lack a native Linux driver. So, in a way, you're getting secondary Windows hell...
The algorithm I apply is to select all that either
Don't get greyed out by deselecting a hardware item I lack, or
Obviously don't apply to my hardware, such as "VIA workarounds" when I have an Intel chipset... etc.
I look at it this way: If the workaround is halfway sane, it shouldn't kick in if I have sane hardware, and should kick in if I have an affected card/mobo/CPU/whatever.
There are few cases when selecting a workaround will actually cause problems, as opposed to simply adding a little bloat. At least, that's been my experience.
I still recompile since I don't usually like the vendor's kernel. Plus, it's kinduva habit, having run Linux as long as I have. That said, at least a couple of my machines are running the vendor kernel.
Re: your sig... 7 years, eh? I've been running Linux since Nov 1993... Do I get a cookie? I guess you beat me to/. though...
Well, I suppose if you could identify some psychoactive chemical that you could trick your immune system into generating as an antibody, and then construct an artificial pathogen to stimulate production of said antibody, it might almost be a vaccine.... if you squint, tilt your head sideways and glance at it just so.:-)
The question though just sounds like something out of Clockwork Orange. Bah. Yay 01-Apr-2005.
Well, something can be flamebait and insightful at the same time.
Just like language, there are certain things that are "signal," and certain things that are "carrier." When we get overwhelmed by accents, it's because we haven't learned how to separate the signal--the information content--from the carrier. When we say a given ethnic group 'looks alike,' we're seeing the 'carrier' more than we're seeing the 'signal.'
You don't need the loopback mount. You can do mkswap/tmp/moreswap and swapon/tmp/moreswap.
Note, that assumes/tmp is a filesystem like ext2 or ext3, not tmpfs.
At least, not liquid water at anything near 1 atmosphere. Theoretically, you could do it with water vapor at a low enough pressure, but I doubt it's remotely practical.
You know, nuclear reactors move liquid, non-boiling water around at well over 100 degrees Celsius, but it's at much more than 1 atmosphere.
I vote we start controversy pointing out that many of these "science centers" (such as the one here in Ft Worth) are taking the science out. A sort-of fight-fire-with-fire approach.
Then again, it'd be like any argument with a fool. They'll drag you down to their level and then beat you with experience.... In this case, it's the science-heads vs. the fundies, and somehow I don't think the science-heads will win by throwing a temper tantrum like the fundies did.
Maybe we should study some Chick tracts to get an idea of what their playbook is... (Ok, so I'm feeling a little snarky tonight.)
It isn't really sad, I suppose.:-) I meant it in a self-deprecating-humorous way.
As for why the family of proteins and genes is named "hedgehog," it has nothing to do with the shape of the protein. Rather, according to this page, "The original hedgehog gene was found in Drosophila and was named for the appearance of the mutant phenotype which causes an embryo to be covered with pointy denticles resembling a hedgehog."
I don't know what's sadder: That you're not making that up (about Sonic hedgehog), or that I know that you're not making that up (and didn't need to look it up).
Slashdot Mirror
There exist operating systems that don't require a reboot after installing software. I upgrade my software all the time, and I don't need to reboot.
Neither Macs nor *NIX boxes have the reboot fetish Windows seems to have.
Funny, I've installed a fair bit of software on each of my Linux boxes, and gee... still seems pretty darn stable! Can't say the same for my WinXP laptop from work. They've got so many SMS scripts running in the background that that thing goes to 80% load for an hour at a time. Never mind Norton AV, ConnectedTLM (backup software), and CyberShield.
I think a big factor here is that installing an app under Linux doesn't dump a bunch of crap in the registry, nor does it register a bunch of crap to run at startup, nor does it crap all over your desktop and launcher bar with short-cuts to the software. (Man, what ego these companies must have!) If an app just parked itself in its own little self-contained directory and left it at that, it wouldn't matter if you had 1 or 100,000 apps installed, except maybe to your diskspace.
--JoeIn theory, if it's a large enough violation, it becomes a criminal act at the Federal level, as opposed to a civil liability. In that case, I'd imagine the Department of Justice can bring suit.
--JoeUhm, because both of them are odd composites that seem like improbable oxymorons?
And don't forget Teslacles Deviant to Fudd's Law: "If it goes in, it must come out."
ObRant: It's subpoena, not subpeona.
/have seen this misspelling too many times in one form.
Also, your big storage/backup drives and network shares tend to be on the higher drive letters. It really is a good strategy.
--JoeUnder UNIX and Linux, if you delete a file that's in use, the directory entry goes away, but the inode remains in place until all the users of the file close their file handles. So, if you delete a running executable, it can still bring pages of itself into memory as needed, even though you can't see the file any more. Once the program exits, the inode gets freed and the disk space reclaimed.
So, you could do an rm -rf / and keep running for quite awhile, until, like you said, an existing task needs to open a file that's no longer there, or you try to fork some new task.
I should try that some time. I have a UML disk image set up somewhere... I could start up UML in "COW" mode (meaning any changes to the disk image get written to a separate "delta" file that I can throw away afterwards).
As for bitchchecker, it seems entirely plausible that remote disk-erasing software would work from higher-lettered disks to lower-lettered disks, simply because system files and programs tend to be on lower disk letters (C: usually), and huge MP3/movie/warez stores tend to be on the extra volumes. You'll notice Windows crashing pretty quick. You won't notice half your pr0n stash disappeared right away (unless you're really that big of a wanker). Starting from higher-lettered drives could also be even more devastating if the PC under attack has network drives mounted to those letters.
--JoeMine would probably laugh briefly with the teacher at me, and then rekindle and redirect their previous ire directly at me the whole way home...
--JoeYou do know what a steer is, right?
Back when I was in school on the campus UNIX system, I used to tell people about the default "slow mode" that most accounts start under. You see, since it was a time-shared machine, the default setting for a login session was a resource-conserving "slow mode." This ensured fairness on the machine, you see.
The trick, then, to speeding things up is to exit this mode. The command for doing so? exit $SLOW_MODE
Enough people do that, and sure enough, the machine speeds up... for everyone else!
(BTW, I'm aware now that said trick doesn't work on csh/tcsh users. This was an AT&T StarServer E running SVR4, though, so nearly all of us were running the original Bourne Shell.)
--Joe127/8 does not mean "127 or 128" in this context. It means "127.x.x.x with a netmask of 255.0.0.0" (8 bits of subnetwork identifier, 24 bits of host identifier). For instance, 10/8 is reserved as a non-routable address range, as is 192.168/16. That means you might use 10.x.x.x or 192.168.x.x addresses on your private networks. In the pre-CIDR days, a /8 was called a "Class A" network, a /16 was called a "Class B" network, and a /24 was called a "Class C" network.
I'm sure those of you who are Net.Gods will skewer me on slight terminology misuse, but I think I conveyed the basic idea.
--JoeMore like, "When?" That style seems straight out of the 1700s.
That'd be 10,000 bloated drivers. Most drivers are FAR from being 100KB. A 1GB kernel source tree is highly unlikely to persist as "the way things are."
If the kernel source tree gets to be 100MB compressed, I think you will see the tarball fork into subtrees. If it doesn't, it's only because bandwidth to each of the developers has seen at least a 10x if not 100x growth from today's current status quo.
Keep in mind that the entire kernel history from the end of 2.4 through the current 2.6.x only takes 3 GB in the new "git" management system, and it doesn't do 'diff' style compression. (Though it does gzip individual files.) That 3GB is due to amplifying all the one-liners and other trivial commits that happened in the many hundred kernel intermediate versions between the two endpoints.
--Joe"I would not join a club that would have me as a member." --Groucho Marx... or how ever it goes... ;-)
Anyway, veering noticeably ontopic, ndiswrapper hell only applies to network cards which lack a native Linux driver. So, in a way, you're getting secondary Windows hell...
--JoeThe algorithm I apply is to select all that either
I look at it this way: If the workaround is halfway sane, it shouldn't kick in if I have sane hardware, and should kick in if I have an affected card/mobo/CPU/whatever.
There are few cases when selecting a workaround will actually cause problems, as opposed to simply adding a little bloat. At least, that's been my experience.
--JoeI still recompile since I don't usually like the vendor's kernel. Plus, it's kinduva habit, having run Linux as long as I have. That said, at least a couple of my machines are running the vendor kernel.
Re: your sig... 7 years, eh? I've been running Linux since Nov 1993... Do I get a cookie? I guess you beat me to /. though...
--JoeOf course, these days we have ground-based observatories that rival Hubble. Those can hold us awhile while the ESA or JAXA work out the next big space observatory... ;-)
Well, I suppose if you could identify some psychoactive chemical that you could trick your immune system into generating as an antibody, and then construct an artificial pathogen to stimulate production of said antibody, it might almost be a vaccine.... if you squint, tilt your head sideways and glance at it just so. :-)
The question though just sounds like something out of Clockwork Orange. Bah. Yay 01-Apr-2005.
--JoeWell, something can be flamebait and insightful at the same time.
Just like language, there are certain things that are "signal," and certain things that are "carrier." When we get overwhelmed by accents, it's because we haven't learned how to separate the signal--the information content--from the carrier. When we say a given ethnic group 'looks alike,' we're seeing the 'carrier' more than we're seeing the 'signal.'
--JoeYou don't need the loopback mount. You can do mkswap /tmp/moreswap and swapon /tmp/moreswap.
Note, that assumes /tmp is a filesystem like ext2 or ext3, not tmpfs.
At least, not liquid water at anything near 1 atmosphere. Theoretically, you could do it with water vapor at a low enough pressure, but I doubt it's remotely practical.
You know, nuclear reactors move liquid, non-boiling water around at well over 100 degrees Celsius, but it's at much more than 1 atmosphere.
--JoeI vote we start controversy pointing out that many of these "science centers" (such as the one here in Ft Worth) are taking the science out. A sort-of fight-fire-with-fire approach.
Then again, it'd be like any argument with a fool. They'll drag you down to their level and then beat you with experience.... In this case, it's the science-heads vs. the fundies, and somehow I don't think the science-heads will win by throwing a temper tantrum like the fundies did.
Maybe we should study some Chick tracts to get an idea of what their playbook is... (Ok, so I'm feeling a little snarky tonight.)
--JoeIt isn't really sad, I suppose. :-) I meant it in a self-deprecating-humorous way.
As for why the family of proteins and genes is named "hedgehog," it has nothing to do with the shape of the protein. Rather, according to this page, "The original hedgehog gene was found in Drosophila and was named for the appearance of the mutant phenotype which causes an embryo to be covered with pointy denticles resembling a hedgehog."
--JoeI don't know what's sadder: That you're not making that up (about Sonic hedgehog), or that I know that you're not making that up (and didn't need to look it up).