The job of the lawyer is to know the law inside out so that they can assist their client. The job of the legislator is to draft laws and regulations that have as few loopholes and weaknesses as possible.
If blame is to be assigned, it goes to the lawmakers.
The law shouldn't be about some cat-and-mouse game to find and close loopholes. It should be about legislating what most citizens would consider right and wrong.
Courts should (and might already) find that willful sidestepping around the intent of a law is equivalent to a violation of the law. The law should be simple - you can't sponsor a green card if Americans are able and willing to do the job. It might give some basic minimums like some basic steps to trying to hire Americans, but in the end if you sponsor green cards when Americans are in fact able and willing to do a job, then you should be punished for breaking the law.
If it really is unclear whether Congress intended to make a practice illegal, that is one thing. However, if you show a practice on TV and 99.99% of the voting public is positive that the practice violates the intent of the law, then it really shouldn't be up for debate whether some loophole around the law exists. If you're 51/49% or maybe even 70/30% that might be one thing, but when EVERYBODY thinks you're sleazy then something probably should be done...
If Intel becomes a monopoly on high-end fabs one possible outcome would be to require them to divest their fabs to an independant company, which would then be free to sell services to anyone. It isn't an unheard-of tactic with monopolies...
Payments were deferred until then because 2008 was the year the district was scheduled to finish making state loan repayments under its previous loan plan
The STATE just finished making the district repay loans for 15 years. So, when the government is owned money you better pay up, but when IBM is owed money the state turns around and recommends forgiveness?
Actually, small businesses might stay open long after large concerns would close shop. A few reasons:
1. The owners may be inclined to stay in the area and tend the shop, so it doesn't matter that the capital could be better used elsewhere.
2. The owners can't just ship the DVDs to their 500 other stores with minimal loss. If they close shop they must liquidate probably for pennies on the dollar.
3. The owners may be able to use dodgy practices to reduce their costs, without the liabilities a major concern faces.
4. The small business probably has less overhead.
Now, in a hot market the small business will get killed by the corporation, but the small guy may stick around long after the corporation leaves - if for no other reason than they don't have much choice...
I tend to doubt this would work. The costs of these projects are astronomical - so in order to recoup them the license costs would have to be VERY high. And the way people are treating drug patents these days, who is going to want to invest $5B in solving the energy crisis when the American public is probably going to just given them a token compulsory license fee instead of the 10% tax on all energy use for a decade that the invention might be worth?
These are very long-term, high-risk investments. Unless the payoff is large and likely to happen, you won't see private investment. That doesn't mean that we can't try to encourage this, but until lots of people are already making money off of this kind of investment you're not going to see a lot of private cash flowing in...
Suddenly I have a mental image of the mythbusters creating a cooking robot to try to make breakfast while another robot sets the cruise control while taking its hands off the wheel...
After all, nobody could envision the car wandering off the road under this scenario without a proper test, could they?
The best is when IT security meets moderately-high-end science. I've heard of cases where operating systems have be desupported (think NT), and the knee-jerk reaction of IT security is to ban the use of the OS. The problem is that many equipment vendors didn't have software that ran on newer OSes. The knee-jerk response? "Well, just buy new equipment!" The only problem is that we are in some cases talking about $500k+ scientific instruments, whose purchase was based on an operational lifetime of a decade or two - not a few years of OS support.
The other problem is when IT isn't invited to the table at all when decisions are made to purchase said equipment in the first place. The only considerations are cost and scientific benefit, and as a result systems are often difficult to maintain after a few years when their attached computers become relics. That 286 might work fine, but if you can't attach it to the network to run backups you're suddenly increasing the costs of administrating the system significantly...
The most outrageous thing I've ever heard of are "fair share" fees. Those are payroll deductions for non-union-members, used to support the negotiating tactics of the union. Where teachers unions are concerned they are often mandated by law.
I'm sorry, it is not the place of law to establish one particular private organization in a position of power. If some workers want to unionize that is fine, but if an employer wants to hire non-union it shouldn't be the place of government to forbid this.
The same applies to non-governmental standards-bodies established in law as well.
If something needs to be regulated then create a government body to regulate it, complete with democratic process, access laws, etc. If it isn't that important, then just leave it alone. Private organizations shouldn't operate under the color of law...
And no doubt in true knee-jerk reaction the speed limit was set to 15mph on every road within 2 miles of the accident.
People who drive 65 in a residential area should be fined. People who drive 35 on a decent-size street that happens to have house on it should not. It seems like most speed limits are far below what would be reasonable from a safety standpoint...
And people without "healthcare" still get health care. They can pay cash to go to the doctor (it isn't THAT expensive), and even the generic medications are more effective than state of the art was 30 years ago. For acute conditions they can go to the hospital and get state-of-the-art care until they are stabilized.
Obviously you don't get great care by today's standards if you don't have insurance (in the US), but you do get decent care by the standards of a few decades ago.
I'm not saying that things can't be done to improve the situation in the US. However, the general call for everybody to get "the best" healthcare possible is simply impossible to achieve - if you spent more money it would get better, and you have to draw the line somewhere...
Uh, you realize that if you had said that 20 years ago it would probably be fulfilled today in the US without any reform at all, right?
Or do you mean that everybody should get at least average healthcare today, and 10 years from now people should get the average care afforded at that time.
If you do mean the latter, then you are missing concept of what an average actually is. It is impossible for more than 50% of society to get better than median care, and on a weighted basis half of everybody is going to be below average. The average just moves the more you spend on those who are "below average". The only thing you can influence is the distribution, and that can come with some steep costs...
but western society at least has decreed that up to a certain limit (ie smokers should not be given lung transplants)the worth of human life is immutable
Uh, do you realize you just contradicted yourself in a single sentence.
That is like saying that up to a certain limit ($8000) my car is of infinite worth.
If the worth of human life was truly infinite in the value of society, then we'd all be gladly be paying 95% taxes to give bounties to those who go into medicine to encourage more people to do so. You wouldn't have to wait at all to see a doctor - there would be at least one doctor per sick person. After all, if life is truly infinite in value, then we should spare no expense at all in prolonging it.
The fact is that people put a price on life EVERY day. They just don't like to talk about it...
Additionally, the REAL harm argument is pretty vague. What REAL harm do you suffer if I steal a couple of rocks out of your bedroom - it isn't like diamonds are edible, provide warmth, or provide shelter? So, you don't really suffer "REAL" harm just because I stole $50k worth of them.
If I possess something that can be legally traded for cash, and I'm deprived of it, then I've suffered harm. Now, there might be an argument as to whether the person who caused the loss had a duty to care for your stuff. If you deposit $500 in a bank they obviously must care for it. If you check a coat with a pocket full of diamonds while you're at dinner and you don't mention the diamonds to anybody in charge, then you're going to be up the creek if those diamonds disappear (the owner could argue that if they had known about the diamonds that they'd either refuse to store the coat, or they might have given it greater protection).
In this case, the Second Life game expressly encourages real-world-cash transactions, which means they are liable for running the economy in a way that provides some stability to such transactions.
How do you know if ANYTHING has value? Simple - if somebody is willing to give you money for it, there is value.
An autograph can have value greater than the paper and the ink. A baseball hit over the center field wall can have more value than the baseball that was fouled out by the previous pitch to the same batter.
Even in online games that forbid trading of characters, such trading occurs, and therefore such characters have value. Now, you can potentially argue whether players had been given advance warning that the game was not going to be managed in a way to preserve the economy. However, that doesn't mean that virtual property has no value.
How about domain names? If I bribe a registrar to assign me microsoft.com, is Microsoft's only remedy to sue me to reimburse their $7 registration fee?
I don't think the virtual nature of the transactions make it any less susceptible to charges of fraud, if such fraud actually occurred...
I actually wonder if such a screen puts the distributor in violation of the GPL.
The GPL requires that users be given the software under the terms of the GPL, with no further restrictions.
The GPL does not require users to accept it to use the software.
Software which DOES require the users to accept the GPL to use the software is enforcing a restriction not listed in the GPL.
Therefore, the software-mandated license acceptance is in violation of the GPL.
Not sure if this argument would apply if the GPL were taken apart in detail with regard to its restrictions against adding restrictions. I just thought it was an interesting concept...
It is very hard to follow the posts you reference. Were the systems you benchmarked of identical cost? Did they have identical RAID configuration?
I'm not debating that RAID-10 can be faster than RAID-5, and that a true hardware RAID can use less CPU than a software RAID.
However, I would argue that hardware RAID does have downsides, and that in general you are often better spending your money on other system improvements.
You seem to have issue with software RAID using 3-11% of the CPU. Well, if you got a faster CPU it would probably not cost you $250 more (over the cost of whatever CPU you currently have, unless you're really cutting-edge) to drop that usage significantly. Now, if this is a $5000 server with the best CPU available and it isn't fast enough then I guess you have no choice but to spend elsewhere.
Also - talking about 1100% increases in CPU use is a bit misleading. It is like saying that putting a drop of bleach in a swimming pool full of distilled water causes a 100,000% increase in chlorine levels. Relative percentages don't mean much at low levels.
Finally, if the software RAID is adequate for the intended use then it is a waste of money to spend on hardware RAID. Likewise, if a 1.5GHz Semperon does the job it is foolish to buy a Core2Duo unless you expect to need the extra CPU later. We're not talking about a production corporate database server - we're talking about a home media server. Likewise, while a BMW might provide better performance than a Honda Civic, I wouldn't purchase the former for my kids to drive around in.
My point is that anytime you spend money on something you are choosing not to spend it on something else. The question is where is the money best spent?
You raise a VERY good point, which many on this discussion ignore.
Sure, a $100k disk-storage solution will perform better than a $500 solution, but is that the best way to spend your money?
If you're setting up a file server for home use that probably connects to a 100Mhz network you don't need to buy 14 10k SCSI hard drives, a $2000 RAID card, an enclosure to hold it, and a 42U rack to hold that. You're probably better off with software RAID-5, and if you're concerned about speed spend an extra few hundred dollars on RAM/CPU - you'll probably get better overall speed than getting an extra 10% performance by spending $2k more on storage.
Under most workloads the software RAID will outperform hardware - as the host CPU will have spare capacity (especially if you spend a fraction of what you save on a faster CPU). Sure, maybe the latest and greatest $2000 adaptec card will do a little better, but is this a 16-CPU DB server?
The key is getting the performance you need with the capacity you need for the right price. Any idiot can just call EMC and ask them for a datacenter SAN solution for their DVR if they have money to burn and they'll get very good performance and scalability - but why waste your time reading slashdot if you have that much cash burning a hole in your pocket?
The latter will lockup if a single drive crashes (low uptime)
That would depend on your controller. I think that most SATA controllers can handle a drive failure without disrupting the other drives. With IDE you might or might not lose the other drive on the same interface depending on what goes wrong.
And even if the system dies you won't lose any data.
What I laugh at is all the posts here suggesting that the only way to have 500GB of reliable storage is to spend $3000. A simple RAID-5 using md gives later expandability and reliability, and you only waste the capacity of a single drive.
Sure, RAID-10 might give you better performance, but you double your costs. A hardware card might give better performance in some cases, but you add $1k to the price and it is less flexible.
A few other comments. I'm running md on linux and love it. Not much of a performance hit, and you get much better use of your drives compared to mirroring.
Also - if you have odd-sized drives just chop them into even partitions. Then stripe multiple RAID1/5s across the partitions and then put all the RAIDs into an lvm2 volume group so that their space is merged.
If you add more drives, you could create more RAIDs and use lvm2 to merge the space. Or, you could grow the existing RAID - linux supports this, in theory even if the RAID is active. Not sure how safe it is though. Probably OK, but like anything...
I've read bad things about true hardware RAID (with the $1000 cards) - if the card dies you can be really up the creek as the drives probably won't work with anything but an identical card. You're also much more limited compared to the capabilities of linux software RAID...
Relax - there isn't anything that you can do with GPLv2 that you can't do with GPLv3, except try to restrict the rights that GPLv2 was created to secure in the first place.
If you don't like it use some other OS, or write your own from scratch. Perhaps it is the GPL that is the reason that so many companies are embracing linux in the first place - after all they could have just used BSD from the beginning.
Many companies find the GPLv3 to be just fine, or even an improvement. After all, if their business model doesn't consist of trying to restrict their user's GPL rights then they're not giving up anything, and in fact they stand to benefit from the elimination of competitors who do have this business model...
only 3 of them represent you. Allowing people to buy additional representation is directly against what a representative democracy should be about.
The problem is that no major corporation would set up shop in small states if this sort of policy were enforced.
The same sort of logic could be used to just get rid of Federal government entirely - after all, why should some representative that lives 5000 miles away from me be able to pass legislation that impacts my life?
I think the better approach would be to move to a parliamentary system like 95% of the rest of the democratic world. Then there are no local representatives, and national issues are decided in a national forum, with more than two parties holding power.
That's OK, if these guys worked for digital they'd mount it on their server rack at about hip level or elbow level facing outwards.
We had a VAX at work and the halt button was positioned such that just about ANYBODY working in the server room would manage to bump it. The thing probably cost more than a car and nobody thought to put a guard over the front-panel switches or at least recess them...
I think part of the problem is the obsession among scientists of being the first to do something. Just think of all the grad students who spend 3 years on a project, only to get scooped by somebody else, and having to switch gears. Is their work somehow less original because it was done independently by somebody else ahead of them?
The same sorts of issues exist with negative results. Nobody cares about them, but they're just as important as positive results. Especially when everybody keeps reinventing the wheel because nobody bothers to publish that a given approach won't work.
I remember attending science classes at a smaller university and a large top-10 university. One thing that struck me as being different between the two was an emphasis on names. At the smaller university you'd just talk about the science. At the larger one every time a concept was mentioned it would also be mentioned who came up with it. I always figured that at the larger university it was due to a subconscious hope that the person doing the teaching would one day end up in one of those textbooks themselves. That and the fact that at larger universities there tends to be a lot more networking going on. Chances are that a professor at a small university will just do independant work and not collaborate with 14 other research groups (who all review each other's papers and grant applications).
Often I think the politics of large-scale science gets in the way. The same problem exists in private industry, but the desire to make a buck can cut through the red tape pretty quickly when the payoff is big enough...
Basically, the issue was that the human genome project was operating under a rather long timeline (mostly based on the state of technology when the project started). Venter thought that using a massively-parallel approach to the problem and using computers to assemble the resulting mess of data would get the results faster, but with some gaps in the final data that would require follow-up. He started Celera to implement this idea.
The business model was simple. All the data would be released publicly, but before it was patents would be secured on medical uses for a few hundred new genes that looked to be important. They also offered some database services that in theory would be superior to what was available publicly (although they did publicize their sequences in Genbank/etc).
Suffice it to say the world of academics wasn't too happy about this, as the math on paper suggested that with the amount of capital Celera was commiting he would indeed overtake the HGP (on its original timeline) with time to spare. The reason for this is similar to what you see in distributed computing projects like the RC-5 cracking attempts. If a project stretches 5 years you'll probably see that half of the project got done in the last six months - because of the ever-growing computing power available to it. The whole project could be repeated in 1 year instead of the original 5. The same applied to DNA sequencing technology at the time - the decade-plus-long HGP could be completed in a year or two if started today with the same level of funding. So, starting late didn't really slow down Celera all that much.
In the end Venter's idea essentially paid off. Sure, some complain that the random sequencing technique leaves gaps, but they can be closed conventionally. Also, if you look at most intended uses for sequence data, having 99% of it is just about as good as having 100% of it. And when you look at cost it is probably better to have 99% of 10 organisms sequenced to 100% of one.
The massively-parallel approach to sequencing would probably be of interest to many in the/. crowd - it is essentially an IT solution to an otherwise-expensive problem.
Ok, normally I resist feeding the trolls, but the assertion was that an expensive drug was as useless as an uninvented drug. I pointed out that this was not true.
An expensive drug is immediately useful to rich people, and useful to all people 10 years down the road. I'll agree that poor people potentially just die in their conditions without treatment, but I'd argue that few simply can't afford drugs at all - they just have to give up lifestyle to accomplish this. And most pharma companies do have patient assistance programs of various forms for cases where some form of socialized medicine isn't available.
An uninvented drug is of no use to anyone at all - now or 10 years from now. Rich or poor.
Now, you might not like the fact that not all people are treated equally without regard to wealth or social status, but there isn't and has never been a society on earth that treats all people equally. I'm sure that in any fine example of a socialized medical system that if the prime minister and a homeless person both need an organ, and there is only one to go around, that the prime minister will turn out to be the slightly better match (or whatever the excuse ends up being).
In any case, the bottom line is that somebody needs to pay for drug development if you want it to happen. You can't require clinical trials that cost hundreds of millions of euros, and then expect companies to just find a way to pay for it while only receiving marginal cost for the goods they sell. Now, potentially these costs could be borne by a government agency of some sort, and the resulting products being made available free of licensing costs. However, that won't really change the cost of developing drugs. So, whether governments just pay high prices for drugs, or they pay high prices for drug R&D and low prices for the actual pills, you're not going to escape the cost of clinical trials...
If governments really want to try license-free drugs they should go ahead and try to fund a few - either offer bounties, or do the work themselves. Go ahead and let private industry continue to maintain the status quo, and then offer cheap brand-new drugs using government funding. Then society will have all the drugs it would have had otherwise and more, and the true costs of both systems can be compared. If private companies don't provide value for the money nobody will buy their products. If they do then they'll do just fine.
But, I'm pretty sure that no matter how you slice it, all those MDs will still want to be paid for running their clinical trials, and volunteers aren't cheap either when you factor in all the tests you have to run and the data coordination...
Slashdot Mirror
The job of the lawyer is to know the law inside out so that they can assist their client. The job of the legislator is to draft laws and regulations that have as few loopholes and weaknesses as possible.
If blame is to be assigned, it goes to the lawmakers.
The law shouldn't be about some cat-and-mouse game to find and close loopholes. It should be about legislating what most citizens would consider right and wrong.
Courts should (and might already) find that willful sidestepping around the intent of a law is equivalent to a violation of the law. The law should be simple - you can't sponsor a green card if Americans are able and willing to do the job. It might give some basic minimums like some basic steps to trying to hire Americans, but in the end if you sponsor green cards when Americans are in fact able and willing to do a job, then you should be punished for breaking the law.
If it really is unclear whether Congress intended to make a practice illegal, that is one thing. However, if you show a practice on TV and 99.99% of the voting public is positive that the practice violates the intent of the law, then it really shouldn't be up for debate whether some loophole around the law exists. If you're 51/49% or maybe even 70/30% that might be one thing, but when EVERYBODY thinks you're sleazy then something probably should be done...
If Intel becomes a monopoly on high-end fabs one possible outcome would be to require them to divest their fabs to an independant company, which would then be free to sell services to anyone. It isn't an unheard-of tactic with monopolies...
The STATE just finished making the district repay loans for 15 years. So, when the government is owned money you better pay up, but when IBM is owed money the state turns around and recommends forgiveness?
Actually, small businesses might stay open long after large concerns would close shop. A few reasons:
1. The owners may be inclined to stay in the area and tend the shop, so it doesn't matter that the capital could be better used elsewhere.
2. The owners can't just ship the DVDs to their 500 other stores with minimal loss. If they close shop they must liquidate probably for pennies on the dollar.
3. The owners may be able to use dodgy practices to reduce their costs, without the liabilities a major concern faces.
4. The small business probably has less overhead.
Now, in a hot market the small business will get killed by the corporation, but the small guy may stick around long after the corporation leaves - if for no other reason than they don't have much choice...
I tend to doubt this would work. The costs of these projects are astronomical - so in order to recoup them the license costs would have to be VERY high. And the way people are treating drug patents these days, who is going to want to invest $5B in solving the energy crisis when the American public is probably going to just given them a token compulsory license fee instead of the 10% tax on all energy use for a decade that the invention might be worth?
These are very long-term, high-risk investments. Unless the payoff is large and likely to happen, you won't see private investment. That doesn't mean that we can't try to encourage this, but until lots of people are already making money off of this kind of investment you're not going to see a lot of private cash flowing in...
Suddenly I have a mental image of the mythbusters creating a cooking robot to try to make breakfast while another robot sets the cruise control while taking its hands off the wheel...
After all, nobody could envision the car wandering off the road under this scenario without a proper test, could they?
The best is when IT security meets moderately-high-end science. I've heard of cases where operating systems have be desupported (think NT), and the knee-jerk reaction of IT security is to ban the use of the OS. The problem is that many equipment vendors didn't have software that ran on newer OSes. The knee-jerk response? "Well, just buy new equipment!" The only problem is that we are in some cases talking about $500k+ scientific instruments, whose purchase was based on an operational lifetime of a decade or two - not a few years of OS support.
The other problem is when IT isn't invited to the table at all when decisions are made to purchase said equipment in the first place. The only considerations are cost and scientific benefit, and as a result systems are often difficult to maintain after a few years when their attached computers become relics. That 286 might work fine, but if you can't attach it to the network to run backups you're suddenly increasing the costs of administrating the system significantly...
The most outrageous thing I've ever heard of are "fair share" fees. Those are payroll deductions for non-union-members, used to support the negotiating tactics of the union. Where teachers unions are concerned they are often mandated by law.
I'm sorry, it is not the place of law to establish one particular private organization in a position of power. If some workers want to unionize that is fine, but if an employer wants to hire non-union it shouldn't be the place of government to forbid this.
The same applies to non-governmental standards-bodies established in law as well.
If something needs to be regulated then create a government body to regulate it, complete with democratic process, access laws, etc. If it isn't that important, then just leave it alone. Private organizations shouldn't operate under the color of law...
And no doubt in true knee-jerk reaction the speed limit was set to 15mph on every road within 2 miles of the accident.
People who drive 65 in a residential area should be fined. People who drive 35 on a decent-size street that happens to have house on it should not. It seems like most speed limits are far below what would be reasonable from a safety standpoint...
And people without "healthcare" still get health care. They can pay cash to go to the doctor (it isn't THAT expensive), and even the generic medications are more effective than state of the art was 30 years ago. For acute conditions they can go to the hospital and get state-of-the-art care until they are stabilized.
Obviously you don't get great care by today's standards if you don't have insurance (in the US), but you do get decent care by the standards of a few decades ago.
I'm not saying that things can't be done to improve the situation in the US. However, the general call for everybody to get "the best" healthcare possible is simply impossible to achieve - if you spent more money it would get better, and you have to draw the line somewhere...
Uh, you realize that if you had said that 20 years ago it would probably be fulfilled today in the US without any reform at all, right?
Or do you mean that everybody should get at least average healthcare today, and 10 years from now people should get the average care afforded at that time.
If you do mean the latter, then you are missing concept of what an average actually is. It is impossible for more than 50% of society to get better than median care, and on a weighted basis half of everybody is going to be below average. The average just moves the more you spend on those who are "below average". The only thing you can influence is the distribution, and that can come with some steep costs...
but western society at least has decreed that up to a certain limit (ie smokers should not be given lung transplants)the worth of human life is immutable
Uh, do you realize you just contradicted yourself in a single sentence.
That is like saying that up to a certain limit ($8000) my car is of infinite worth.
If the worth of human life was truly infinite in the value of society, then we'd all be gladly be paying 95% taxes to give bounties to those who go into medicine to encourage more people to do so. You wouldn't have to wait at all to see a doctor - there would be at least one doctor per sick person. After all, if life is truly infinite in value, then we should spare no expense at all in prolonging it.
The fact is that people put a price on life EVERY day. They just don't like to talk about it...
Additionally, the REAL harm argument is pretty vague. What REAL harm do you suffer if I steal a couple of rocks out of your bedroom - it isn't like diamonds are edible, provide warmth, or provide shelter? So, you don't really suffer "REAL" harm just because I stole $50k worth of them.
If I possess something that can be legally traded for cash, and I'm deprived of it, then I've suffered harm. Now, there might be an argument as to whether the person who caused the loss had a duty to care for your stuff. If you deposit $500 in a bank they obviously must care for it. If you check a coat with a pocket full of diamonds while you're at dinner and you don't mention the diamonds to anybody in charge, then you're going to be up the creek if those diamonds disappear (the owner could argue that if they had known about the diamonds that they'd either refuse to store the coat, or they might have given it greater protection).
In this case, the Second Life game expressly encourages real-world-cash transactions, which means they are liable for running the economy in a way that provides some stability to such transactions.
How do you know if ANYTHING has value? Simple - if somebody is willing to give you money for it, there is value.
An autograph can have value greater than the paper and the ink. A baseball hit over the center field wall can have more value than the baseball that was fouled out by the previous pitch to the same batter.
Even in online games that forbid trading of characters, such trading occurs, and therefore such characters have value. Now, you can potentially argue whether players had been given advance warning that the game was not going to be managed in a way to preserve the economy. However, that doesn't mean that virtual property has no value.
How about domain names? If I bribe a registrar to assign me microsoft.com, is Microsoft's only remedy to sue me to reimburse their $7 registration fee?
I don't think the virtual nature of the transactions make it any less susceptible to charges of fraud, if such fraud actually occurred...
I actually wonder if such a screen puts the distributor in violation of the GPL.
The GPL requires that users be given the software under the terms of the GPL, with no further restrictions.
The GPL does not require users to accept it to use the software.
Software which DOES require the users to accept the GPL to use the software is enforcing a restriction not listed in the GPL.
Therefore, the software-mandated license acceptance is in violation of the GPL.
Not sure if this argument would apply if the GPL were taken apart in detail with regard to its restrictions against adding restrictions. I just thought it was an interesting concept...
It is very hard to follow the posts you reference. Were the systems you benchmarked of identical cost? Did they have identical RAID configuration?
I'm not debating that RAID-10 can be faster than RAID-5, and that a true hardware RAID can use less CPU than a software RAID.
However, I would argue that hardware RAID does have downsides, and that in general you are often better spending your money on other system improvements.
You seem to have issue with software RAID using 3-11% of the CPU. Well, if you got a faster CPU it would probably not cost you $250 more (over the cost of whatever CPU you currently have, unless you're really cutting-edge) to drop that usage significantly. Now, if this is a $5000 server with the best CPU available and it isn't fast enough then I guess you have no choice but to spend elsewhere.
Also - talking about 1100% increases in CPU use is a bit misleading. It is like saying that putting a drop of bleach in a swimming pool full of distilled water causes a 100,000% increase in chlorine levels. Relative percentages don't mean much at low levels.
Finally, if the software RAID is adequate for the intended use then it is a waste of money to spend on hardware RAID. Likewise, if a 1.5GHz Semperon does the job it is foolish to buy a Core2Duo unless you expect to need the extra CPU later. We're not talking about a production corporate database server - we're talking about a home media server. Likewise, while a BMW might provide better performance than a Honda Civic, I wouldn't purchase the former for my kids to drive around in.
My point is that anytime you spend money on something you are choosing not to spend it on something else. The question is where is the money best spent?
You raise a VERY good point, which many on this discussion ignore.
Sure, a $100k disk-storage solution will perform better than a $500 solution, but is that the best way to spend your money?
If you're setting up a file server for home use that probably connects to a 100Mhz network you don't need to buy 14 10k SCSI hard drives, a $2000 RAID card, an enclosure to hold it, and a 42U rack to hold that. You're probably better off with software RAID-5, and if you're concerned about speed spend an extra few hundred dollars on RAM/CPU - you'll probably get better overall speed than getting an extra 10% performance by spending $2k more on storage.
Under most workloads the software RAID will outperform hardware - as the host CPU will have spare capacity (especially if you spend a fraction of what you save on a faster CPU). Sure, maybe the latest and greatest $2000 adaptec card will do a little better, but is this a 16-CPU DB server?
The key is getting the performance you need with the capacity you need for the right price. Any idiot can just call EMC and ask them for a datacenter SAN solution for their DVR if they have money to burn and they'll get very good performance and scalability - but why waste your time reading slashdot if you have that much cash burning a hole in your pocket?
The latter will lockup if a single drive crashes (low uptime)
That would depend on your controller. I think that most SATA controllers can handle a drive failure without disrupting the other drives. With IDE you might or might not lose the other drive on the same interface depending on what goes wrong.
And even if the system dies you won't lose any data.
What I laugh at is all the posts here suggesting that the only way to have 500GB of reliable storage is to spend $3000. A simple RAID-5 using md gives later expandability and reliability, and you only waste the capacity of a single drive.
Sure, RAID-10 might give you better performance, but you double your costs. A hardware card might give better performance in some cases, but you add $1k to the price and it is less flexible.
A few other comments. I'm running md on linux and love it. Not much of a performance hit, and you get much better use of your drives compared to mirroring.
Also - if you have odd-sized drives just chop them into even partitions. Then stripe multiple RAID1/5s across the partitions and then put all the RAIDs into an lvm2 volume group so that their space is merged.
If you add more drives, you could create more RAIDs and use lvm2 to merge the space. Or, you could grow the existing RAID - linux supports this, in theory even if the RAID is active. Not sure how safe it is though. Probably OK, but like anything...
I've read bad things about true hardware RAID (with the $1000 cards) - if the card dies you can be really up the creek as the drives probably won't work with anything but an identical card. You're also much more limited compared to the capabilities of linux software RAID...
Relax - there isn't anything that you can do with GPLv2 that you can't do with GPLv3, except try to restrict the rights that GPLv2 was created to secure in the first place.
If you don't like it use some other OS, or write your own from scratch. Perhaps it is the GPL that is the reason that so many companies are embracing linux in the first place - after all they could have just used BSD from the beginning.
Many companies find the GPLv3 to be just fine, or even an improvement. After all, if their business model doesn't consist of trying to restrict their user's GPL rights then they're not giving up anything, and in fact they stand to benefit from the elimination of competitors who do have this business model...
only 3 of them represent you. Allowing people to buy additional representation is directly against what a representative democracy should be about.
The problem is that no major corporation would set up shop in small states if this sort of policy were enforced.
The same sort of logic could be used to just get rid of Federal government entirely - after all, why should some representative that lives 5000 miles away from me be able to pass legislation that impacts my life?
I think the better approach would be to move to a parliamentary system like 95% of the rest of the democratic world. Then there are no local representatives, and national issues are decided in a national forum, with more than two parties holding power.
That's OK, if these guys worked for digital they'd mount it on their server rack at about hip level or elbow level facing outwards.
We had a VAX at work and the halt button was positioned such that just about ANYBODY working in the server room would manage to bump it. The thing probably cost more than a car and nobody thought to put a guard over the front-panel switches or at least recess them...
I think part of the problem is the obsession among scientists of being the first to do something. Just think of all the grad students who spend 3 years on a project, only to get scooped by somebody else, and having to switch gears. Is their work somehow less original because it was done independently by somebody else ahead of them?
The same sorts of issues exist with negative results. Nobody cares about them, but they're just as important as positive results. Especially when everybody keeps reinventing the wheel because nobody bothers to publish that a given approach won't work.
I remember attending science classes at a smaller university and a large top-10 university. One thing that struck me as being different between the two was an emphasis on names. At the smaller university you'd just talk about the science. At the larger one every time a concept was mentioned it would also be mentioned who came up with it. I always figured that at the larger university it was due to a subconscious hope that the person doing the teaching would one day end up in one of those textbooks themselves. That and the fact that at larger universities there tends to be a lot more networking going on. Chances are that a professor at a small university will just do independant work and not collaborate with 14 other research groups (who all review each other's papers and grant applications).
Often I think the politics of large-scale science gets in the way. The same problem exists in private industry, but the desire to make a buck can cut through the red tape pretty quickly when the payoff is big enough...
Ah, I remember those days.
/. crowd - it is essentially an IT solution to an otherwise-expensive problem.
Basically, the issue was that the human genome project was operating under a rather long timeline (mostly based on the state of technology when the project started). Venter thought that using a massively-parallel approach to the problem and using computers to assemble the resulting mess of data would get the results faster, but with some gaps in the final data that would require follow-up. He started Celera to implement this idea.
The business model was simple. All the data would be released publicly, but before it was patents would be secured on medical uses for a few hundred new genes that looked to be important. They also offered some database services that in theory would be superior to what was available publicly (although they did publicize their sequences in Genbank/etc).
Suffice it to say the world of academics wasn't too happy about this, as the math on paper suggested that with the amount of capital Celera was commiting he would indeed overtake the HGP (on its original timeline) with time to spare. The reason for this is similar to what you see in distributed computing projects like the RC-5 cracking attempts. If a project stretches 5 years you'll probably see that half of the project got done in the last six months - because of the ever-growing computing power available to it. The whole project could be repeated in 1 year instead of the original 5. The same applied to DNA sequencing technology at the time - the decade-plus-long HGP could be completed in a year or two if started today with the same level of funding. So, starting late didn't really slow down Celera all that much.
In the end Venter's idea essentially paid off. Sure, some complain that the random sequencing technique leaves gaps, but they can be closed conventionally. Also, if you look at most intended uses for sequence data, having 99% of it is just about as good as having 100% of it. And when you look at cost it is probably better to have 99% of 10 organisms sequenced to 100% of one.
The massively-parallel approach to sequencing would probably be of interest to many in the
Ok, normally I resist feeding the trolls, but the assertion was that an expensive drug was as useless as an uninvented drug. I pointed out that this was not true.
An expensive drug is immediately useful to rich people, and useful to all people 10 years down the road. I'll agree that poor people potentially just die in their conditions without treatment, but I'd argue that few simply can't afford drugs at all - they just have to give up lifestyle to accomplish this. And most pharma companies do have patient assistance programs of various forms for cases where some form of socialized medicine isn't available.
An uninvented drug is of no use to anyone at all - now or 10 years from now. Rich or poor.
Now, you might not like the fact that not all people are treated equally without regard to wealth or social status, but there isn't and has never been a society on earth that treats all people equally. I'm sure that in any fine example of a socialized medical system that if the prime minister and a homeless person both need an organ, and there is only one to go around, that the prime minister will turn out to be the slightly better match (or whatever the excuse ends up being).
In any case, the bottom line is that somebody needs to pay for drug development if you want it to happen. You can't require clinical trials that cost hundreds of millions of euros, and then expect companies to just find a way to pay for it while only receiving marginal cost for the goods they sell. Now, potentially these costs could be borne by a government agency of some sort, and the resulting products being made available free of licensing costs. However, that won't really change the cost of developing drugs. So, whether governments just pay high prices for drugs, or they pay high prices for drug R&D and low prices for the actual pills, you're not going to escape the cost of clinical trials...
If governments really want to try license-free drugs they should go ahead and try to fund a few - either offer bounties, or do the work themselves. Go ahead and let private industry continue to maintain the status quo, and then offer cheap brand-new drugs using government funding. Then society will have all the drugs it would have had otherwise and more, and the true costs of both systems can be compared. If private companies don't provide value for the money nobody will buy their products. If they do then they'll do just fine.
But, I'm pretty sure that no matter how you slice it, all those MDs will still want to be paid for running their clinical trials, and volunteers aren't cheap either when you factor in all the tests you have to run and the data coordination...