In theory a SCSI disk should not be much better than ATA but the reality is the best made, more reliable parts go to the high end more profitable products.
There are, typically, huge differences between your average SCSI and ATA disk beyond just manufacturing quality. USENIX paper here. Differences range from disk interface command richness, to reliability under wider operating conditions, to materials and assembly. They do the same thing, but they aren't even close to being the same thing.
You sort of disproved your own point in the second paragraph! While assigning IRQs may be irrelevant now, knowing what an interrupt is and why it is important is not. God, I can't believe I'm putting myself in the old-timer group, but my first computer was a 3MHz, 16-bit pre-IBM PC machine, so I guess I am to some degree. I've come to the conclusion over the years that any knowledge about computers, however obscure or seemingly irrelevant, will definitely be useful at one point or another.
Yes and no. If all you care about is more money, I have news for you: people like me are going to beat your at your job every time. You don't care about that part because you're just in it for the money, but hey, we'll probably beat you at the pay part, too. We do better work because we like our work.
People have hobbies. Some people like to fish. Some like muscle cars. Others like to spend their time in bars shooting the shit. Me? I like to change my perception about the world. Try it sometime. For me, school is a worthwhile pasttime in itself, because I enjoy computer science.
Since you are still in school, it is impossible for you to even venture a guess if the time has paid off.
I received my first Bachelors nearly a decade ago. I worked in my field for awhile but decided that it wasn't for me after all, and I decided to pursue a career in computers instead. I went on a long hike and when I came back, I changed jobs.
After I started school, my employer changed my role pretty dramatically, because I was a much more capable employee. I was promoted twice in two years, and I make significantly more than I did when I started. I even have my own staff now. But, you know, whatever. Don't let a lack of information stop you from talking shit.
Not to mention extremely expensive with little real pay off.
Pure BS. I don't dispute that some schools aren't worth shit, but I'm now working on my second Bachelor's, in Computer Science, just like the poster, and I wouldn't trade it for anything. I'll be graduating pretty soon.
Here's the thing I noticed the first time around, as a Philosophy major. Take your average community college, and, say Harvard. Have a look at your typical philosophy class. Say, Critique of Pure Reason, or Platonic dialogs. Same. Fucking. Books. So what sets them apart? Well, it _should_ be the quality of the professor, right?
But this gap isn't as big as you'd think. Assuming you get a PhD teaching your class, you've got someone who point quite a lot of time into becoming an expert in that subject. Not to mention-- it's _your_ attitude that matters anyhow. Any sufficiently motivated student will have a good experience no matter who their professors are. I say this now having been through the classically horrible science-professor experience.
I've also supplemented my in-class experience by watching the CS lectures on MIT's OpenCourseWare. I would say that, in general, these guys are perhaps better computer scientists, but whether they are better teachers is in question. So this reinforces my opinion.
The bottom line is that you go back to school because you love the subject. If you think computers are cool, and you want to know more, go for it. Computer science has been the same mind-bending experience that my first degree was. This time I'm a bit more mature-- homework always gets done, and-- shit-- I'm paying for this degree out of my own pocket, so I'd better make the best of it. At work, my CS knowledge has greatly expanded my capabilities and my enjoyment of the job.
"metadata" is a bit of a catch-all, but generally filesystem designers consider "metadata" to be the on-disk structures that tell the OS where to find files. These structures sometimes contain information about the file, but what's more important is that it tells the OS where it is in relation to a directory. The free list (what you're talking about) is also metadata. The interesting thing is that the free list itself is sometimes actually stored in the free blocks, which means that when you use a block, you take it out of the free list, so your storage cost for the list is essentially nothing.
It makes a filesystem more complicated, because now you need additional space to do the copy-on-write. In the past, there was a reasonable expectation that a user might keep a disk near full, but now that storage is cheap, that expectation is no longer valid. Also, making this fast is non-trivial. You probably want your write routine to write to the first available (and acceptable) blocks it finds, but now you need to go back to your original inode and tell it where all of the chunks of the file are. Reading that file back now could be very slow, because your file is no longer/less contiguous. So you want to be very careful about where you put things.
Most filesystems are a balance of things done for data integrity and things done for speed. Copy-on-write was probably not considered to be worth the effort until people started to realize how important their data was...
Accidentally put a user's file in the wrong directory, or don't link it at all, and I can assure you that they will care about metadata consistency. A file is pretty much useless in a modern filesystem unless the directory points to its inode.
Point being-- it is much faster to buffer the write, write the metadata first, unblock the process, and do the rest of the writing in the background. At this point, everyone is well aware of the tradeoffs. Journaling just makes sure you don't end up in an inconsistent state as far as the FS is concerned-- this does NOT mean that your data won't be corrupted! Along the lines of what GP says, if you need certain characteristics, like atomicity, then you probably also need the rest of the ACID stuff, too, so you're probably better off with a real database.
This is simply not true. Say you notice that one of the load-bearing walls in your house is starting to crumble. Do you:
1) Spend money to repair it, or
2) Save money by doing nothing
Rational people would do #1, even though it costs more in the short term, because it will save you more in the long term-- i.e., you do not have to build a new house when it collapses.
Preventative health care is the same thing. E.g., a gym membership can save you a significant amount of money in the long run, not to mention, by maintaining your health, you increase the quality of your life.
Why are people so opposed to knowing the answers around here? I'd pay to know this information about myself.
It matters less how much it costs. What matters more is how much it saves you.
Let's take fluoridation. It costs (according to Wikipedia) $0.93 per person per year. There are roughly 250 million people in the United States, so-- this is a lot of money!
But fluoridation also saves an enormous amount of money (about $60 per tooth surface, in the worst case, again according to Wikipedia), when you factor in money spent repairing cavity-ridden teeth, productivity lost due to said teeth, and personal suffering because of those teeth. It also allows us eat a much wider variety of foods (generally foods of lesser nutritional value) because we don't have to worry about our teeth immediately rotting out of our heads when we eat them.
My gut feeling is that there is no question that, in the long run, constant health monitoring would save us money. We're scientists here, right? More data is always better than less.
Hey, if that means that greed motivates someone to make sure, e.g., my couch doesn't explode into an uncontrolled bonfire when I drop my cigarette onto it, then that's a good thing, isn't it? Ever heard of Underwriters Laboratories? You generally don't think of those guys are trying to fuck us over, but this is in fact an organization whose sole purpose is to reduce insurance claims through better engineering.
Sure, insurance companies are scumbags, but aren't you thinking about this a little bit backward?
we have our nervous system - a far more comprehensive, sensitive monitoring and fault detection system than covers most industrial plants.
The nervous system is good, but-- it could be better. It can't tell you that chronic high blood pressure is an indicator of coronary artery disease. It can't tell you that chronic high blood sugar levels will cause diabetes. It can't tell you when you have HIV.
I think constant monitoring would be most important when it comes to public health research. Right now, we have a model of the human body that is based on periodic testing, and, when you die, pathology on your dead tissue. We haven't been able to constantly monitor a living human being, let alone, a large population of human beings. This will surely give us useful information to adjust our human model.
For instance, we know that both genetics and behavior have big consequences on a person's health, and that understanding how the two interact will allow us to develop medicines tailored to a specific person's biology. Without constant monitoring, coming up with sufficient data is going to be a much lengthier process.
Surely modern medicine has given us great personal freedom. We can travel the entire United States, Europe and many other places without fear of being infected by many diseases, because, by and large, our population has herd immunity to many of the worst diseases. We don't need to worry about our diets, because, we generally understand nutrition-- you're not, e.g., going to get scurvy the next time you take a cruise to the Bahamas. IMHO, knowing the details of your body can only help you live your life to its fullest potential.
The people working at those jobs aren't doing anything PRODUCTIVE.
You seriously believe this? These corporations employ tens of thousands, and sometimes, hundreds of thousands of employees. The people making the absurd wages are only the tiniest fraction of the people working at these companies. Maybe you don't work for a large corporation (I do), so I can't fault you for not knowing that your typical business employs mail room clerks, receptionist, IT people and programmers (like me), repairmen, deliverymen, copyeditors, and on, and on... And these jobs indirectly support an entire ecosystem of other jobs. Food service people, courier services, telecommunications and utility workers. Heck, even police departments make money doing special details when your typical corporation wants to use public spaces.
Now, I'm not saying that all these corporations are worth saving. Bankruptcy is the right solution for many companies. But as a mental exercise, just try, for a moment, to imagine how many people would be affected by GM and Chrysler going under.
I entirely agree with you-- sometimes you need to cut off the toe. But you also want to make sure that when you do it your glasses are on straight. You might in fact be cutting off your leg.
Choice quote from Wikipedia article: "The jury has the right to judge both the law as well as the fact in controversy." -- John Jay, first Chief Justice of the United States
It's not your decision as a juror to decide what laws apply to the case, or what evidence should be presented.
Actually, it is. Jury nullification is a widely-recognized right whereby a jury can disregard not only instructions from the judge, but can also disregard the pertinent statute. This is, in fact, the entire point of the jury trial: not just that you may be judged by your peers, but that the law itself is judged by your peers. This is how the jury trial is a defense against tyranny.
The right of jury nullification has not been confirmed by the US Supreme Court, but it has been confirmed by many appellate courts. On the other hand, most courts have concluded that the defense can be prohibited (and penalized) from instructing a jury about their right to nullification.
Anyhow, point being, juries are quite powerful. They are, in fact, the ones making the decisions in a court of law, even if they are ignorant of this fact.
Based on my experience, I can't say that's true. Now, I'm no audiophile, but I have done radio production work, and so I've been exposed to some fairly high-end stuff. I've used lots of Sennheiser gear in the past, and well... Sennheiser has a tendency to dilute their brandname at the low end of the price scale. Their high-end stuff is generally pretty good.
I own a pair of Shure E2C's. Earbuds. This is not considered a real high-end product. But for me, they sound shockingly good. Better than any of the gear we had at the radio station. Good enough that I can pick out the flaws in my sound card (though that may not be saying much), and occasionally encoding artifacts. It's real easy to pick out bad studio engineering.
Anyway, I consider them to have been a good deal, and I am very happy with them. They lack a little on the bass end, but then again, I don't need the bowels shaken out of me while I code at work.
That is indeed important to keep in mind. However, we must also remember that municipalities presently spray the fuck out of their neighborhoods to kill these things, taking down entire ecosystems with them. So lasers might actually be an improvement-- in Maine, for example, they banned spraying for black flies a couple years ago. As a result, freshwater fish stocks are way up. A system like this could be localized, protecting just your house, for instance.
It's funny, though-- as an avid outdoorsman, I've fantasized about such a system for a long time. I would love to see mosquitoes go down in flames. Dear inventors, please include horseflies in your plans!
Yeah, no kidding. Both Lost and Heroes are on prime time TV. Comic book movies come out faster than we can watch them. And now they change their name? That decision's right up there with canceling MST3K. Time to start looking forward to more Kevin Sorbo!
It's not that simple in a corporate environment (i.e., LAN). We do packet filtering and proxy at our ingress and egress points, we stay up-to-date with patches (WSUS), and AV (ESET), and we've disabled a number of unnecessary Windows services, but still, occasionally infections get through. Sometimes this is because a consultant or freelancer walks through the door and plugs into our network; sometimes it's because a laptop user brings something back with them. Sometimes, yes, it's our own users who are stupid, and the defenses we have in place do not catch them. So far, we've been able to limit damage, but as for stopping it completely-- this has been hard to achieve. As far as we can tell, the only way to accomplish this is to ditch Windows.
Besides, if you don't run AV, how do you know you don't have something? Do you trawl your firewall logs daily? At the moment, Conficker is pretty much just sitting there, waiting to do something. You might not even know you have it.
Slashdot Mirror
In theory a SCSI disk should not be much better than ATA but the reality is the best made, more reliable parts go to the high end more profitable products.
There are, typically, huge differences between your average SCSI and ATA disk beyond just manufacturing quality. USENIX paper here. Differences range from disk interface command richness, to reliability under wider operating conditions, to materials and assembly. They do the same thing, but they aren't even close to being the same thing.
Enjoy the girls who are put off by boys who are still trying to show off.
That is, if you're not married. *Gah!*
You sort of disproved your own point in the second paragraph! While assigning IRQs may be irrelevant now, knowing what an interrupt is and why it is important is not. God, I can't believe I'm putting myself in the old-timer group, but my first computer was a 3MHz, 16-bit pre-IBM PC machine, so I guess I am to some degree. I've come to the conclusion over the years that any knowledge about computers, however obscure or seemingly irrelevant, will definitely be useful at one point or another.
School is just another investment.
Yes and no. If all you care about is more money, I have news for you: people like me are going to beat your at your job every time. You don't care about that part because you're just in it for the money, but hey, we'll probably beat you at the pay part, too. We do better work because we like our work.
People have hobbies. Some people like to fish. Some like muscle cars. Others like to spend their time in bars shooting the shit. Me? I like to change my perception about the world. Try it sometime. For me, school is a worthwhile pasttime in itself, because I enjoy computer science.
Since you are still in school, it is impossible for you to even venture a guess if the time has paid off.
I received my first Bachelors nearly a decade ago. I worked in my field for awhile but decided that it wasn't for me after all, and I decided to pursue a career in computers instead. I went on a long hike and when I came back, I changed jobs.
After I started school, my employer changed my role pretty dramatically, because I was a much more capable employee. I was promoted twice in two years, and I make significantly more than I did when I started. I even have my own staff now. But, you know, whatever. Don't let a lack of information stop you from talking shit.
Not to mention extremely expensive with little real pay off.
Pure BS. I don't dispute that some schools aren't worth shit, but I'm now working on my second Bachelor's, in Computer Science, just like the poster, and I wouldn't trade it for anything. I'll be graduating pretty soon.
Here's the thing I noticed the first time around, as a Philosophy major. Take your average community college, and, say Harvard. Have a look at your typical philosophy class. Say, Critique of Pure Reason, or Platonic dialogs. Same. Fucking. Books. So what sets them apart? Well, it _should_ be the quality of the professor, right?
But this gap isn't as big as you'd think. Assuming you get a PhD teaching your class, you've got someone who point quite a lot of time into becoming an expert in that subject. Not to mention-- it's _your_ attitude that matters anyhow. Any sufficiently motivated student will have a good experience no matter who their professors are. I say this now having been through the classically horrible science-professor experience.
I've also supplemented my in-class experience by watching the CS lectures on MIT's OpenCourseWare. I would say that, in general, these guys are perhaps better computer scientists, but whether they are better teachers is in question. So this reinforces my opinion.
The bottom line is that you go back to school because you love the subject. If you think computers are cool, and you want to know more, go for it. Computer science has been the same mind-bending experience that my first degree was. This time I'm a bit more mature-- homework always gets done, and-- shit-- I'm paying for this degree out of my own pocket, so I'd better make the best of it. At work, my CS knowledge has greatly expanded my capabilities and my enjoyment of the job.
When whooshing, it helps if the GP was actually making a joke.
"metadata" is a bit of a catch-all, but generally filesystem designers consider "metadata" to be the on-disk structures that tell the OS where to find files. These structures sometimes contain information about the file, but what's more important is that it tells the OS where it is in relation to a directory. The free list (what you're talking about) is also metadata. The interesting thing is that the free list itself is sometimes actually stored in the free blocks, which means that when you use a block, you take it out of the free list, so your storage cost for the list is essentially nothing.
It makes a filesystem more complicated, because now you need additional space to do the copy-on-write. In the past, there was a reasonable expectation that a user might keep a disk near full, but now that storage is cheap, that expectation is no longer valid. Also, making this fast is non-trivial. You probably want your write routine to write to the first available (and acceptable) blocks it finds, but now you need to go back to your original inode and tell it where all of the chunks of the file are. Reading that file back now could be very slow, because your file is no longer/less contiguous. So you want to be very careful about where you put things.
Most filesystems are a balance of things done for data integrity and things done for speed. Copy-on-write was probably not considered to be worth the effort until people started to realize how important their data was...
Accidentally put a user's file in the wrong directory, or don't link it at all, and I can assure you that they will care about metadata consistency. A file is pretty much useless in a modern filesystem unless the directory points to its inode.
Point being-- it is much faster to buffer the write, write the metadata first, unblock the process, and do the rest of the writing in the background. At this point, everyone is well aware of the tradeoffs. Journaling just makes sure you don't end up in an inconsistent state as far as the FS is concerned-- this does NOT mean that your data won't be corrupted! Along the lines of what GP says, if you need certain characteristics, like atomicity, then you probably also need the rest of the ACID stuff, too, so you're probably better off with a real database.
This is simply not true. Say you notice that one of the load-bearing walls in your house is starting to crumble. Do you: 1) Spend money to repair it, or 2) Save money by doing nothing Rational people would do #1, even though it costs more in the short term, because it will save you more in the long term-- i.e., you do not have to build a new house when it collapses.
Preventative health care is the same thing. E.g., a gym membership can save you a significant amount of money in the long run, not to mention, by maintaining your health, you increase the quality of your life.
Why are people so opposed to knowing the answers around here? I'd pay to know this information about myself.
It matters less how much it costs. What matters more is how much it saves you.
Let's take fluoridation. It costs (according to Wikipedia) $0.93 per person per year. There are roughly 250 million people in the United States, so-- this is a lot of money!
But fluoridation also saves an enormous amount of money (about $60 per tooth surface, in the worst case, again according to Wikipedia), when you factor in money spent repairing cavity-ridden teeth, productivity lost due to said teeth, and personal suffering because of those teeth. It also allows us eat a much wider variety of foods (generally foods of lesser nutritional value) because we don't have to worry about our teeth immediately rotting out of our heads when we eat them.
My gut feeling is that there is no question that, in the long run, constant health monitoring would save us money. We're scientists here, right? More data is always better than less.
Hey, if that means that greed motivates someone to make sure, e.g., my couch doesn't explode into an uncontrolled bonfire when I drop my cigarette onto it, then that's a good thing, isn't it? Ever heard of Underwriters Laboratories? You generally don't think of those guys are trying to fuck us over, but this is in fact an organization whose sole purpose is to reduce insurance claims through better engineering.
Sure, insurance companies are scumbags, but aren't you thinking about this a little bit backward?
we have our nervous system - a far more comprehensive, sensitive monitoring and fault detection system than covers most industrial plants.
The nervous system is good, but-- it could be better. It can't tell you that chronic high blood pressure is an indicator of coronary artery disease. It can't tell you that chronic high blood sugar levels will cause diabetes. It can't tell you when you have HIV.
I think constant monitoring would be most important when it comes to public health research. Right now, we have a model of the human body that is based on periodic testing, and, when you die, pathology on your dead tissue. We haven't been able to constantly monitor a living human being, let alone, a large population of human beings. This will surely give us useful information to adjust our human model.
For instance, we know that both genetics and behavior have big consequences on a person's health, and that understanding how the two interact will allow us to develop medicines tailored to a specific person's biology. Without constant monitoring, coming up with sufficient data is going to be a much lengthier process.
Surely modern medicine has given us great personal freedom. We can travel the entire United States, Europe and many other places without fear of being infected by many diseases, because, by and large, our population has herd immunity to many of the worst diseases. We don't need to worry about our diets, because, we generally understand nutrition-- you're not, e.g., going to get scurvy the next time you take a cruise to the Bahamas. IMHO, knowing the details of your body can only help you live your life to its fullest potential.
The people working at those jobs aren't doing anything PRODUCTIVE.
You seriously believe this? These corporations employ tens of thousands, and sometimes, hundreds of thousands of employees. The people making the absurd wages are only the tiniest fraction of the people working at these companies. Maybe you don't work for a large corporation (I do), so I can't fault you for not knowing that your typical business employs mail room clerks, receptionist, IT people and programmers (like me), repairmen, deliverymen, copyeditors, and on, and on... And these jobs indirectly support an entire ecosystem of other jobs. Food service people, courier services, telecommunications and utility workers. Heck, even police departments make money doing special details when your typical corporation wants to use public spaces.
Now, I'm not saying that all these corporations are worth saving. Bankruptcy is the right solution for many companies. But as a mental exercise, just try, for a moment, to imagine how many people would be affected by GM and Chrysler going under.
I entirely agree with you-- sometimes you need to cut off the toe. But you also want to make sure that when you do it your glasses are on straight. You might in fact be cutting off your leg.
The jury is to determine the facts, not to apply their own understanding of the law.
This is incorrect. See this post.
Choice quote from Wikipedia article: "The jury has the right to judge both the law as well as the fact in controversy." -- John Jay, first Chief Justice of the United States
It's not your decision as a juror to decide what laws apply to the case, or what evidence should be presented.
Actually, it is. Jury nullification is a widely-recognized right whereby a jury can disregard not only instructions from the judge, but can also disregard the pertinent statute. This is, in fact, the entire point of the jury trial: not just that you may be judged by your peers, but that the law itself is judged by your peers. This is how the jury trial is a defense against tyranny.
The right of jury nullification has not been confirmed by the US Supreme Court, but it has been confirmed by many appellate courts. On the other hand, most courts have concluded that the defense can be prohibited (and penalized) from instructing a jury about their right to nullification.
Anyhow, point being, juries are quite powerful. They are, in fact, the ones making the decisions in a court of law, even if they are ignorant of this fact.
earbuds are crap, period.
Based on my experience, I can't say that's true. Now, I'm no audiophile, but I have done radio production work, and so I've been exposed to some fairly high-end stuff. I've used lots of Sennheiser gear in the past, and well... Sennheiser has a tendency to dilute their brandname at the low end of the price scale. Their high-end stuff is generally pretty good.
I own a pair of Shure E2C's. Earbuds. This is not considered a real high-end product. But for me, they sound shockingly good. Better than any of the gear we had at the radio station. Good enough that I can pick out the flaws in my sound card (though that may not be saying much), and occasionally encoding artifacts. It's real easy to pick out bad studio engineering.
Anyway, I consider them to have been a good deal, and I am very happy with them. They lack a little on the bass end, but then again, I don't need the bowels shaken out of me while I code at work.
Maybe I just like crap!
That is indeed important to keep in mind. However, we must also remember that municipalities presently spray the fuck out of their neighborhoods to kill these things, taking down entire ecosystems with them. So lasers might actually be an improvement-- in Maine, for example, they banned spraying for black flies a couple years ago. As a result, freshwater fish stocks are way up. A system like this could be localized, protecting just your house, for instance.
It's funny, though-- as an avid outdoorsman, I've fantasized about such a system for a long time. I would love to see mosquitoes go down in flames. Dear inventors, please include horseflies in your plans!
Yeah, no kidding. Both Lost and Heroes are on prime time TV. Comic book movies come out faster than we can watch them. And now they change their name? That decision's right up there with canceling MST3K. Time to start looking forward to more Kevin Sorbo!
Really? I guess I'm not a real science fiction fan, then. You guys must have HUGE bookshelves!
It's not that simple in a corporate environment (i.e., LAN). We do packet filtering and proxy at our ingress and egress points, we stay up-to-date with patches (WSUS), and AV (ESET), and we've disabled a number of unnecessary Windows services, but still, occasionally infections get through. Sometimes this is because a consultant or freelancer walks through the door and plugs into our network; sometimes it's because a laptop user brings something back with them. Sometimes, yes, it's our own users who are stupid, and the defenses we have in place do not catch them. So far, we've been able to limit damage, but as for stopping it completely-- this has been hard to achieve. As far as we can tell, the only way to accomplish this is to ditch Windows.
Besides, if you don't run AV, how do you know you don't have something? Do you trawl your firewall logs daily? At the moment, Conficker is pretty much just sitting there, waiting to do something. You might not even know you have it.
If he's using SVN, then he's already got that covered, at least, for the paying stuff.
And yes, I have statistics and anecdotal evidence both on my side.
Link?
But more importantly, it reverts back to plutoid after it passes over.
I say go for it. Astronomy really was too simple before.
Ha ha. I have coworkers who, as far as I can tell, hate the Awesome Bar because the "community" does. I'm with you-- it works for me.
That said, I think a feature that causes so much divisiveness might be a cause for a checkbox somewhere.