Right. Especially for laptops, which tend to have slower hard drives in the first place.
I installed TrueCrypt on my moderately old laptop, an Intel 1.6Ghz, and the only speed different I notice is that, for some reason, hibernation and unhibernation is twice as slow. I suspect this is some sort of bug. Other than that, I forget it's there except when I boot up.
TrueCrypt, by default, uses AES, which was designed for speed on modern processors. (Or, rather, was designed to use exactly the mathematical operations that CPU manufacturers optimize for in order to make games run faster, so as CPUs keep speeding those operations up AES gets faster.)
Ha, I just checked to see if that hibernation thing is a bug, and it turns out that not only is it, but it's been fixed in 6.0 and I should just upgrade instead of whining about it.
I work for a fairly small company, and while we don't have any person data off our server, and in fact don't really have any personal data beyond names, addresses and email accounts...
...we have logins to our CC processor and whatnot that could trivially be used to steal quite a lot of CC numbers. In addition to probably breaking into our bank account and draining. In addition to getting into our servers and installing backdoors.
Which is why, of course, we have Truecrypt with boot-time encryption on all laptops, so that if they get stolen we don't have to run around like chickens with our heads cut off trying to figure out every single login that needs to be changed.
For those people worried about forgetting password: Burn three or four TrueCrypt 'recovery CD' and write the password on them. In fact, write the password everywhere...just don't carry it around in the laptop bag.
Seriously, half these 'data thefts' are random laptop thieves stealing random laptop that just happen to include absurdly dangerous amounts of data on them. They aren't targeted attacks, and the thief is probably wiping them before boot. But companies have to act like they have all your data because said companies are morons who can't spend a tiny amount of time setting up free software that would stop that from happening.
People often worry about computer security in entirely the wrong direction, worrying about changing internal company-only passwords every month, and then completely ignoring actual outside risks like someone snatching a laptop bag off someone's arm.
The two people who responded about Linux hibernation seem to have it confused with suspension.
Suspend is notoriously flaky in Linux. This is because power management states are very...crappy...in a lot of hardware. Manufacturers write drivers that work for Windows, Linux people have to guess until things work. (Unlike other driver issues, this isn't a matter of manufacturers claiming it's proprietary, it's a matter of them being incredibly lazy and bad engineers and not actually following the well-defined standard and fixing it in their Windows drivers.)
But there is no conceivable way that hibernation doesn't work every time. Hibernation is just shutdown and restart, it requires no hardware support. There is a power management state designed for 'hibernation', but Linux and Windows don't use it, because half the manufacturers didn't bother to even put it in. (Possible the Intel Apples use it.) They just shut the computer down like normal, but instead of a normal shutdown process they write everything to a file, and instead of a normal boot they read everything from a file.
The only actual problems that I've ever seen is people attempting to do it as a normal user and not having permissions. For a while, Gnome would idiotically try to run the hibernation command as the user you'd logged in as, even though it'd go and run the shutdown command as the superuser. And half the distros didn't bother hooking it any acpi buttons or lid-close events, so it would never run in cases where you though it should, but that's the fault of the distro.
There are a very few pieces of hardware that need custom programs to run on startup to initialize them, and can end up with those not having run when resuming from hibernation, but that's bad driver design and actually more likely to exist on Windows than Linux, because Linux drivers that need that get fixed. (And got fixed, years ago.)
I bought a cellphone repeater to stick half on my roof and half in my house because of bad reception.
It would probably be trivially easy to open that thing up and disable the thing that makes it only broadcast it's available for handoffs if it can lock onto a tower. And then operate it without an antenna and, tada, it's a cell phone 'jammer', in that cell phones will switch to it and not actually be able to do anything. And then hook it up to a signal amplifier and blanket an area.
If cell phone networks were as fragile as some people think, there's no way in hell they'd let us have repeaters, which is literally a tiny tower.
All that is silly. You don't need to wipe the hard drive, all you have to do is make sure it is unrecoverable from items in the house plus your memory.
First, build a machine that, on bootup, reads the encrypted header (With the actual key) into memory and then overwrites it on disk. With 0s, so it's obviously invalid. On proper shutdown, it writes it back.
Then put a screensaver password on the computer. Disable any sort of proper shutdown from that screen.
Then put a UPS toggle tripwire in places where, if the door is broken down, it turns off the UPS.
So if you're there and they break down the door, it cuts off without you doing anything, so no tampering with evidence charge. If they don't break down the door, you can quickly press a screensaver hotkey, which is technically illegal but unprovable, and (openly) disable the tripwire, so you can't be charged with tampering with evidence.
While in theory if they show up and the screensaver is on, they could demand you type in the password, in actuality they will 'intelligent' immediately shut the computer off to make a disk image. (And, of course, you could do exactly what they're trying to avoid you having done, and make a duress password, just in case they aren't 'intelligent'.)
Then find someone trusted in another country to hold a copy of the encrypted header for you. Tell them to destroy it if you are unreachable or arrested. Tell the police exactly what you did, (To avoid a tampering with evidence charge) they won't have time to stop it...international demands for witnesses and evidence are very convoluted.
Sure, officer, you can have my key. It's 'I buried the bodies at the coords -32...'...hey, wait a minute, you're trying to get me to incriminate myself, I want a lawyer!
The problem with British law is that you don't have two different levels of law, with one harder to alter than the over, like the US Constitution vs. US law.
Well, you might have that in various ways, possibly, it's just that 'rights' aren't on the higher tier.
Actually, it's possible you don't have two tiers at all, as I read about things like forming the UK and in incorporating Northern Ireland via law, whereas, legally, if the US wanted to form some meta-nation with Canada we couldn't do that without a constitutional amendment. (Actually, not even then...amendments cannot reduce a state's representation without their consent.)
You guys just appear to have a big blob of laws and traditions, which do provide plenty of rights, but those rights are much easier removed than in the US.
Really? Did they have access to the US counts if they wished to prove they were, in fact, citizens?
The minute there exists a subclass of people who do not have the constitutional right of access to a court, all the government has to do is assert that you are a member of that subclass and you have no way to prove you're not, because, hey, no access to the courts.
Are people really so stupid to let people be imprisoned without any sort of judicial oversight, just because of a government-claimed attribute?
Of course, there's actually nothing in the bill of rights about it apply to citizens only. The bill of rights actually applies to the government and restricts it from doing things.
I don't know what country you live in (I honestly don't.), but in the US, no, the judge cannot demand you tell them where the key is, or demand you tell them anything at all whatsoever.
To repeat: In the US, no suspect can be jailed for failing to comply with the police or the courts demand for any information at all. You are under no obligation to tell them anything if you are a suspect in a crime except, possibly, your name. (And even that seems to be under debate.)
Except encryption keys.
Note this only applies to your own criminal activity. You can, in fact, be forced to incriminate others (Except your spouse.) and testify to their criminal activity, unless it would also incriminate you.
The people who think people can be forced to provide any information to help the police are missing the concept of 'right to remain silent'.
And people who think that encrypted data provides some sort of unique reality that demands we surrender our rights, simply because modern encryption is uncrackable, need to realize there are a dozen other 'uncrackable' things out there that law enforcement deals with. For example, the sheer amount of physical space means that it is very possible to hide 'undiscoverable' bodies or weapons, yet judges cannot imprison people until they tell the courts where the body is.
Of course, if Windows hibernation operated anything like Linux hibernation, it would work a lot better.
For reference, Linux hibernation doesn't bother writing non-writable memory pages to the hibernation file. So the hibernation file is much smaller compared to Windows. (Which is why Linux can hibernate to a swap file.)
But this is because Linux can 'swap' from the original executable file into memory. So when it unhibernates, it 'unswaps' most of the programs from their original location, only loading the data segments from the swap file.
Of course, a good portion of the program is already in swap, so what actually happens is that all data segments not in the swap file are written to it, with as much executable segments overwritten as needed to fit those in. It is very very fast.
As opposed to Windows, which sits down and writes out all of physical memory to another file, and then has to load it all back in.(It might even write out 'clean' memory pages that are already in the swap file and unchanged since they were loaded back in memory, but I bet MS is smarter than that.)
Granted, Linux still has to, eventually, load all the programs into memory too, but it can load them in via 'swap', which is fairly invisible to the end user.
The real joke, of course, is that shutdown really should be almost instant.
But MS, and Linux, and everyone else, insists on terminating each and every process for no apparent reason.
Just close the damn open files and turn the power off, you stupid computer.
Re:Building your own kernel these days ain't easy
on
Linux 2.6.27 Out
·
· Score: 1
Oh, sorry, I thought you were being sarcastic or something.
Yeah, the old low level configuration stuff is going away, finally. I remember when I had to do that on a DOS machine, too, locating drivers and putting them in the config.sys and whatnot.
Of course, the huge difference is that Microsoft has the manufacturers make the drivers, and, frankly, they suck ass at software, and often fail to keep it up to date.
I have to laugh at people who complain that Linux doesn't support 'newer hardware'...apparently they are unaware of how little older hardware Windows supports. I used to have a whole box of hardware that wouldn't work in Windows XP anymore.
Linux rarely, if ever, removes hardware support. I remember when there was an outcry over them removing support for MFM hard drives..so they didn't. (Although they at least split it out of the modern IDE driver.)
Meanwhile, I have USB devices that don't work out of the box on XP. For example, I have a USB 1.1 link cable that needs XP drivers, making it utterly useless. USB 1.1 was released in 1998, Windows XP was released 3 years later, and the manufacturer or MS never bothered to release a driver. (And, honestly, why should they? The manufacturer had moved on to 2.0 cables, and MS didn't give a damn.) I could manually install 2000 drivers, but as half the point of a link cable is link to other people's computer, and would require me carrying around the drivers, it seemed sort of stupid, even assuming I had admin access. (Granted, it'd be stupid regardless now, but that was back when USB flash drives were very expensive and very small.)
Honestly, Linux's policy of having all the drivers in one place, maintained by one team, has resulted in infinitely better hardware support. In fact, it's the few places that they aren't maintained by the kernel team that causes the problems. (Like wifi devices used to be, don't know if they are anymore, and nvidia's insistence on binary drivers, and so one.)
OTOH, the annoying habit of compiling a kernel is just that...a habit. No one has any reason to do that anymore unless they want experimental drivers or something odd like that.
Re:Building your own kernel these days ain't easy
on
Linux 2.6.27 Out
·
· Score: 1
My useful skills are running web and database servers and programming, not spending an hour answering hardware questions that actually don't make my systems work any better.
I don't understand why we just don't have each country sign each TLD, or whatever.
Frankly, no one's going to be checking that the.com key is signed by the . key. They'll have the.com key cached, check the signature on each.com DNS looked up, and only worry about the . key if the.com key changes.
Granted, I'm not entirely certain how DNSSEC is supposed to work in the first place.
Re:Building your own kernel these days ain't easy
on
Linux 2.6.27 Out
·
· Score: 1
No kidding.
I used to worry about that crap, stripping it to exactly the right thing, everything built into the kernel, with maybe a few extra modules like USB devices I didn't have but might conceivable have in my possession at some point. I'd spent ten seconds on each option, carefully considering it.
And then I realized it was damn stupid, and just went with the defaults, just setting the CPU to the exact one, and turning off things I didn't have at all, like SCSI hardware or ISDN stuff. And sometimes flipping things from 'module' to compiled in if I noticed them and knew they'd always be loaded.
Nowdays, I have no local Linux boxes, and on my servers, I just use the default kernel. I have better things to do with my time. Although I do wish there was some trivially simply way of, at least, compiling it for the exact CPU architecture, which would take almost none of my time. I guess I could track down the source RPM and recompile that.
Witnesses can refuse to testify if their testimony would incriminate them in a crime, and only if it would incriminate them. (Whereas defendants can refuse to testify regardless.)
However, adultery is not a crime, and thus you cannot refuse to testify simply because your testimony includes that. (1) If you refuse, the court will hold you in contempt.
Likewise, the court can simply grant you immunity and compel you to testify even if it will incriminate you. (Although they then can't use it against you.)
1) Although, technically, in many states laws against it, or against sex outside of marriage, are still on the books, and that would make an interesting legal question.
I am attempting to reduce lying on the stand without reducing information there.
The ability to privately request, perhaps via attorney, the right to testify in private if they can give some good explanation for it would help.
Actually, a better place it would help is the police. If suspects could give, for example, an alibi to the police with the assurance that, if it checks out and the case against them is dismissed, the alibi will never be made public, we might have a good deal less lying.
And, incidentally, your idea only works for defendants, not witnesses. A witness who is having an affair might lie about some relevant information to cover up an affair, and witnesses, obviously, can't refuse to testify.
I agree with you, although I think perjury is sometimes understandable, especially when it isn't to avoid criminal prosecution but to avoid embarrassment.
I wish that witnesses on the stand could ask to 'approach the bench' if they wished to. Or, at the very least, ask for a closed courtroom.
But, yes, anyone who lies to get out of punishment, anyone who is actually found guilty, should be prosecuted for perjury. Same with anyone who attempts to incriminate someone else.
Your claim is that average wages have gone DOWN over the past 2 decades (your timeframe?)?
Jesus Christ. If you aren't aware that we've either had rising unemployment or, at the most, steady wages (Which means dropping because of inflation.) pretty much all through the 90s and 00s, I don't know why I'm talking about the economy with you.
Look, it's really simple. When you make it so expensive to do business in the US, you make manufactured goods from overseas all the cheaper. If an American car manufacturer has to pay for highly taxed steel, highly taxed computer parts, highly taxed engines, etc, the final result is going to be even MORE expensive than a car manufactured entirely outside of the country.
Um, only if you're really stupid about how you tax things. At no point did I suggest having any sort of flat tax based on price like you appear to imagine. (Especially since, when companies import thinks for resell here, we can damn well ask to see their books and what they paid workers.)
What we're actually taxing is the work that went into the each product. Or, rather, the amount of work that didn't happen here, or the amount that they would have paid had the work happened vs. where it did happen. Which is, of course, impossible to tax directly, but is certainly possible to estimate.
Of course, we need to realize that sometimes there are legitimate reasons that price differ, For example, the example you're using isn't a very good one. Japan is making cheaper cars, but certainly doesn't have a lower standard of living. It's making cheaper cars because American car companies suck ass, and have for quite some time. I don't, in any way, propose we attempt to hinder Japanese car sales with tax. (And, ironically, a lot of 'Japanese' cars are manufactured here, and are, amazingly, just as cheap.)
What will kill all manufacturing here is the fact that is cheaper to do it elsewhere regardless. It will never be as cheap to employ Americans as it will people in third-world countries until our standard of living drops to theirs, or until every single person in the world is raised to our standard of living.
This isn't some debatable point, it's not some obscure economic theory, it's something that is instantly and blatantly obvious. It's something that has happened for the last two decades. All jobs that can go elsewhere will, leaving us with only in-person services and construction and other jobs that can't be done remotely.
The only solution is to tax companies who sell in the US but do not manufacture here. Whether they're 'US companies' or foreign, although we need different methods for each.
You can sit and whine about cheap steel all you want. I suspect you wouldn't be whining the same way if you were in the American steel industry. Oh, wait. Cheap Chinese steel already destroyed that industry. I guess having slightly lower prices for your car was more important there, although, of course, less jobs on average means lower wages on average, so in the end most people came out the same.
As has been pointed out, under the free market, it is quite possible to pay people into destroying themself, and that's exactly what's been going on...America, as a whole, has been bribed with low prices into destroying all its industries, so no one has jobs, so it needs low prices, and so on and so on.
Slashdot Mirror
Right. Especially for laptops, which tend to have slower hard drives in the first place.
I installed TrueCrypt on my moderately old laptop, an Intel 1.6Ghz, and the only speed different I notice is that, for some reason, hibernation and unhibernation is twice as slow. I suspect this is some sort of bug. Other than that, I forget it's there except when I boot up.
TrueCrypt, by default, uses AES, which was designed for speed on modern processors. (Or, rather, was designed to use exactly the mathematical operations that CPU manufacturers optimize for in order to make games run faster, so as CPUs keep speeding those operations up AES gets faster.)
Ha, I just checked to see if that hibernation thing is a bug, and it turns out that not only is it, but it's been fixed in 6.0 and I should just upgrade instead of whining about it.
It's not just personal data on the laptop.
I work for a fairly small company, and while we don't have any person data off our server, and in fact don't really have any personal data beyond names, addresses and email accounts...
Which is why, of course, we have Truecrypt with boot-time encryption on all laptops, so that if they get stolen we don't have to run around like chickens with our heads cut off trying to figure out every single login that needs to be changed.
For those people worried about forgetting password: Burn three or four TrueCrypt 'recovery CD' and write the password on them. In fact, write the password everywhere...just don't carry it around in the laptop bag.
Seriously, half these 'data thefts' are random laptop thieves stealing random laptop that just happen to include absurdly dangerous amounts of data on them. They aren't targeted attacks, and the thief is probably wiping them before boot. But companies have to act like they have all your data because said companies are morons who can't spend a tiny amount of time setting up free software that would stop that from happening.
People often worry about computer security in entirely the wrong direction, worrying about changing internal company-only passwords every month, and then completely ignoring actual outside risks like someone snatching a laptop bag off someone's arm.
The two people who responded about Linux hibernation seem to have it confused with suspension.
Suspend is notoriously flaky in Linux. This is because power management states are very...crappy...in a lot of hardware. Manufacturers write drivers that work for Windows, Linux people have to guess until things work. (Unlike other driver issues, this isn't a matter of manufacturers claiming it's proprietary, it's a matter of them being incredibly lazy and bad engineers and not actually following the well-defined standard and fixing it in their Windows drivers.)
But there is no conceivable way that hibernation doesn't work every time. Hibernation is just shutdown and restart, it requires no hardware support. There is a power management state designed for 'hibernation', but Linux and Windows don't use it, because half the manufacturers didn't bother to even put it in. (Possible the Intel Apples use it.) They just shut the computer down like normal, but instead of a normal shutdown process they write everything to a file, and instead of a normal boot they read everything from a file.
The only actual problems that I've ever seen is people attempting to do it as a normal user and not having permissions. For a while, Gnome would idiotically try to run the hibernation command as the user you'd logged in as, even though it'd go and run the shutdown command as the superuser. And half the distros didn't bother hooking it any acpi buttons or lid-close events, so it would never run in cases where you though it should, but that's the fault of the distro.
There are a very few pieces of hardware that need custom programs to run on startup to initialize them, and can end up with those not having run when resuming from hibernation, but that's bad driver design and actually more likely to exist on Windows than Linux, because Linux drivers that need that get fixed. (And got fixed, years ago.)
I bought a cellphone repeater to stick half on my roof and half in my house because of bad reception.
It would probably be trivially easy to open that thing up and disable the thing that makes it only broadcast it's available for handoffs if it can lock onto a tower. And then operate it without an antenna and, tada, it's a cell phone 'jammer', in that cell phones will switch to it and not actually be able to do anything. And then hook it up to a signal amplifier and blanket an area.
If cell phone networks were as fragile as some people think, there's no way in hell they'd let us have repeaters, which is literally a tiny tower.
All that is silly. You don't need to wipe the hard drive, all you have to do is make sure it is unrecoverable from items in the house plus your memory.
First, build a machine that, on bootup, reads the encrypted header (With the actual key) into memory and then overwrites it on disk. With 0s, so it's obviously invalid. On proper shutdown, it writes it back.
Then put a screensaver password on the computer. Disable any sort of proper shutdown from that screen.
Then put a UPS toggle tripwire in places where, if the door is broken down, it turns off the UPS.
So if you're there and they break down the door, it cuts off without you doing anything, so no tampering with evidence charge. If they don't break down the door, you can quickly press a screensaver hotkey, which is technically illegal but unprovable, and (openly) disable the tripwire, so you can't be charged with tampering with evidence.
While in theory if they show up and the screensaver is on, they could demand you type in the password, in actuality they will 'intelligent' immediately shut the computer off to make a disk image. (And, of course, you could do exactly what they're trying to avoid you having done, and make a duress password, just in case they aren't 'intelligent'.)
Then find someone trusted in another country to hold a copy of the encrypted header for you. Tell them to destroy it if you are unreachable or arrested. Tell the police exactly what you did, (To avoid a tampering with evidence charge) they won't have time to stop it...international demands for witnesses and evidence are very convoluted.
You know what's interesting? TrueCrypt lets you use a file as a passphrase.
This is 'truly' existing independently of your will...but could you be forced to identify which file it was?
Sure, officer, you can have my key. It's 'I buried the bodies at the coords -32...'...hey, wait a minute, you're trying to get me to incriminate myself, I want a lawyer!
No, you really aren't. Nothing requires you to make it easy for the police to get in.
(Of course, I remind people that booby traps are illegal, regardless of the victim having a legal right to be there or not.)
The problem with British law is that you don't have two different levels of law, with one harder to alter than the over, like the US Constitution vs. US law.
Well, you might have that in various ways, possibly, it's just that 'rights' aren't on the higher tier.
Actually, it's possible you don't have two tiers at all, as I read about things like forming the UK and in incorporating Northern Ireland via law, whereas, legally, if the US wanted to form some meta-nation with Canada we couldn't do that without a constitutional amendment. (Actually, not even then...amendments cannot reduce a state's representation without their consent.)
You guys just appear to have a big blob of laws and traditions, which do provide plenty of rights, but those rights are much easier removed than in the US.
Really? Did they have access to the US counts if they wished to prove they were, in fact, citizens?
The minute there exists a subclass of people who do not have the constitutional right of access to a court, all the government has to do is assert that you are a member of that subclass and you have no way to prove you're not, because, hey, no access to the courts.
Are people really so stupid to let people be imprisoned without any sort of judicial oversight, just because of a government-claimed attribute?
Of course, there's actually nothing in the bill of rights about it apply to citizens only. The bill of rights actually applies to the government and restricts it from doing things.
I don't know what country you live in (I honestly don't.), but in the US, no, the judge cannot demand you tell them where the key is, or demand you tell them anything at all whatsoever.
To repeat: In the US, no suspect can be jailed for failing to comply with the police or the courts demand for any information at all. You are under no obligation to tell them anything if you are a suspect in a crime except, possibly, your name. (And even that seems to be under debate.)
Except encryption keys.
Note this only applies to your own criminal activity. You can, in fact, be forced to incriminate others (Except your spouse.) and testify to their criminal activity, unless it would also incriminate you.
The people who think people can be forced to provide any information to help the police are missing the concept of 'right to remain silent'.
And people who think that encrypted data provides some sort of unique reality that demands we surrender our rights, simply because modern encryption is uncrackable, need to realize there are a dozen other 'uncrackable' things out there that law enforcement deals with. For example, the sheer amount of physical space means that it is very possible to hide 'undiscoverable' bodies or weapons, yet judges cannot imprison people until they tell the courts where the body is.
You do not have to unlock your meth lab for the police, despite what you may think.
They, however, have the right to break down the door or even go in through the walls if you do not.
Of course, if Windows hibernation operated anything like Linux hibernation, it would work a lot better.
For reference, Linux hibernation doesn't bother writing non-writable memory pages to the hibernation file. So the hibernation file is much smaller compared to Windows. (Which is why Linux can hibernate to a swap file.)
But this is because Linux can 'swap' from the original executable file into memory. So when it unhibernates, it 'unswaps' most of the programs from their original location, only loading the data segments from the swap file.
Of course, a good portion of the program is already in swap, so what actually happens is that all data segments not in the swap file are written to it, with as much executable segments overwritten as needed to fit those in. It is very very fast.
As opposed to Windows, which sits down and writes out all of physical memory to another file, and then has to load it all back in.(It might even write out 'clean' memory pages that are already in the swap file and unchanged since they were loaded back in memory, but I bet MS is smarter than that.)
Granted, Linux still has to, eventually, load all the programs into memory too, but it can load them in via 'swap', which is fairly invisible to the end user.
The real joke, of course, is that shutdown really should be almost instant.
But MS, and Linux, and everyone else, insists on terminating each and every process for no apparent reason.
Just close the damn open files and turn the power off, you stupid computer.
Oh, sorry, I thought you were being sarcastic or something.
Yeah, the old low level configuration stuff is going away, finally. I remember when I had to do that on a DOS machine, too, locating drivers and putting them in the config.sys and whatnot.
Of course, the huge difference is that Microsoft has the manufacturers make the drivers, and, frankly, they suck ass at software, and often fail to keep it up to date.
I have to laugh at people who complain that Linux doesn't support 'newer hardware'...apparently they are unaware of how little older hardware Windows supports. I used to have a whole box of hardware that wouldn't work in Windows XP anymore.
Linux rarely, if ever, removes hardware support. I remember when there was an outcry over them removing support for MFM hard drives..so they didn't. (Although they at least split it out of the modern IDE driver.)
Meanwhile, I have USB devices that don't work out of the box on XP. For example, I have a USB 1.1 link cable that needs XP drivers, making it utterly useless. USB 1.1 was released in 1998, Windows XP was released 3 years later, and the manufacturer or MS never bothered to release a driver. (And, honestly, why should they? The manufacturer had moved on to 2.0 cables, and MS didn't give a damn.) I could manually install 2000 drivers, but as half the point of a link cable is link to other people's computer, and would require me carrying around the drivers, it seemed sort of stupid, even assuming I had admin access. (Granted, it'd be stupid regardless now, but that was back when USB flash drives were very expensive and very small.)
Honestly, Linux's policy of having all the drivers in one place, maintained by one team, has resulted in infinitely better hardware support. In fact, it's the few places that they aren't maintained by the kernel team that causes the problems. (Like wifi devices used to be, don't know if they are anymore, and nvidia's insistence on binary drivers, and so one.)
OTOH, the annoying habit of compiling a kernel is just that...a habit. No one has any reason to do that anymore unless they want experimental drivers or something odd like that.
My useful skills are running web and database servers and programming, not spending an hour answering hardware questions that actually don't make my systems work any better.
I wonder if we could sell the DNS root zone to the Chinese to cover our bills until next month.
I don't understand why we just don't have each country sign each TLD, or whatever.
Frankly, no one's going to be checking that the .com key is signed by the . key. They'll have the .com key cached, check the signature on each .com DNS looked up, and only worry about the . key if the .com key changes.
Granted, I'm not entirely certain how DNSSEC is supposed to work in the first place.
No kidding.
I used to worry about that crap, stripping it to exactly the right thing, everything built into the kernel, with maybe a few extra modules like USB devices I didn't have but might conceivable have in my possession at some point. I'd spent ten seconds on each option, carefully considering it.
And then I realized it was damn stupid, and just went with the defaults, just setting the CPU to the exact one, and turning off things I didn't have at all, like SCSI hardware or ISDN stuff. And sometimes flipping things from 'module' to compiled in if I noticed them and knew they'd always be loaded.
Nowdays, I have no local Linux boxes, and on my servers, I just use the default kernel. I have better things to do with my time. Although I do wish there was some trivially simply way of, at least, compiling it for the exact CPU architecture, which would take almost none of my time. I guess I could track down the source RPM and recompile that.
That's why I prefer Linux : it maybe sometimes more trouble to make it work, but once it is setup, I never have to touch it again.
No shit, especially as the solution to Windows problems often are 'reinstall it'. Oh, how fun.
Witnesses can refuse to testify if their testimony would incriminate them in a crime, and only if it would incriminate them. (Whereas defendants can refuse to testify regardless.)
However, adultery is not a crime, and thus you cannot refuse to testify simply because your testimony includes that. (1) If you refuse, the court will hold you in contempt.
Likewise, the court can simply grant you immunity and compel you to testify even if it will incriminate you. (Although they then can't use it against you.)
1) Although, technically, in many states laws against it, or against sex outside of marriage, are still on the books, and that would make an interesting legal question.
None of that is the point.
I am attempting to reduce lying on the stand without reducing information there.
The ability to privately request, perhaps via attorney, the right to testify in private if they can give some good explanation for it would help.
Actually, a better place it would help is the police. If suspects could give, for example, an alibi to the police with the assurance that, if it checks out and the case against them is dismissed, the alibi will never be made public, we might have a good deal less lying.
And, incidentally, your idea only works for defendants, not witnesses. A witness who is having an affair might lie about some relevant information to cover up an affair, and witnesses, obviously, can't refuse to testify.
I agree with you, although I think perjury is sometimes understandable, especially when it isn't to avoid criminal prosecution but to avoid embarrassment.
I wish that witnesses on the stand could ask to 'approach the bench' if they wished to. Or, at the very least, ask for a closed courtroom.
But, yes, anyone who lies to get out of punishment, anyone who is actually found guilty, should be prosecuted for perjury. Same with anyone who attempts to incriminate someone else.
Your claim is that average wages have gone DOWN over the past 2 decades (your timeframe?)?
Jesus Christ. If you aren't aware that we've either had rising unemployment or, at the most, steady wages (Which means dropping because of inflation.) pretty much all through the 90s and 00s, I don't know why I'm talking about the economy with you.
Look, it's really simple. When you make it so expensive to do business in the US, you make manufactured goods from overseas all the cheaper. If an American car manufacturer has to pay for highly taxed steel, highly taxed computer parts, highly taxed engines, etc, the final result is going to be even MORE expensive than a car manufactured entirely outside of the country.
Um, only if you're really stupid about how you tax things. At no point did I suggest having any sort of flat tax based on price like you appear to imagine. (Especially since, when companies import thinks for resell here, we can damn well ask to see their books and what they paid workers.)
What we're actually taxing is the work that went into the each product. Or, rather, the amount of work that didn't happen here, or the amount that they would have paid had the work happened vs. where it did happen. Which is, of course, impossible to tax directly, but is certainly possible to estimate.
Of course, we need to realize that sometimes there are legitimate reasons that price differ, For example, the example you're using isn't a very good one. Japan is making cheaper cars, but certainly doesn't have a lower standard of living. It's making cheaper cars because American car companies suck ass, and have for quite some time. I don't, in any way, propose we attempt to hinder Japanese car sales with tax. (And, ironically, a lot of 'Japanese' cars are manufactured here, and are, amazingly, just as cheap.)
God, you're dumb, aren't you?
What will kill all manufacturing here is the fact that is cheaper to do it elsewhere regardless. It will never be as cheap to employ Americans as it will people in third-world countries until our standard of living drops to theirs, or until every single person in the world is raised to our standard of living.
This isn't some debatable point, it's not some obscure economic theory, it's something that is instantly and blatantly obvious. It's something that has happened for the last two decades. All jobs that can go elsewhere will, leaving us with only in-person services and construction and other jobs that can't be done remotely.
The only solution is to tax companies who sell in the US but do not manufacture here. Whether they're 'US companies' or foreign, although we need different methods for each.
You can sit and whine about cheap steel all you want. I suspect you wouldn't be whining the same way if you were in the American steel industry. Oh, wait. Cheap Chinese steel already destroyed that industry. I guess having slightly lower prices for your car was more important there, although, of course, less jobs on average means lower wages on average, so in the end most people came out the same.
As has been pointed out, under the free market, it is quite possible to pay people into destroying themself, and that's exactly what's been going on...America, as a whole, has been bribed with low prices into destroying all its industries, so no one has jobs, so it needs low prices, and so on and so on.