You've certainly nailed one of the biggest issues. The ability to control your data, have a deletion policy that is then subpoena-free (including backup destruction), etc. is certainly a deal breaker for most larger companies.
Not optional for any public company in the US, so a non-issue.
Uh, I don't follow you. For a large company good control of deletions is mandatory, but how that makes this a non-issue escapes me. Does gmail provide for guaranteed deletion?
Pretty much every large company is de-facto required to rigorously delete materials that aren't required by law to be kept. The penalty for not doing so is to be buried alive in discovery whenever you are sued (which probably happens about once a week for a big company).
If a law says that the accounting records used to create a quarterly statement needs to be kept for three years, then you want to keep it for three years (with no meaningful chance of loss), and then make every remote trace of them disappear completely one day later (or as close to this as possible).
This has nothing to do with covering up wrongful behavior. The fact is that you can comply with the letter of the law today, and then be second-guessed about some decision you made with perfect hindsight 10 years from now. So, unless the law states that those records must be preserved for 10 years, then you don't want those records.
As an analogy from the sysadmin world - consider system access logs at an ISP (something here most people would intuitively understand). If you are an ISP you want to keep your logs long enough to handle billing disputes or to be able to identify abuses, but you don't want a 10 year record of everything every one of your customers have ever done online. Then, when some man files a subpoena to find out when a relationship between his soon-to-be-ex-wife and her boyfriend started you can just say that your records only go back 30 days so that you don't have anything, instead of having to go digging through the vaults to find backup tapes, and then show up in court to testify about how those records were maintained (a cost you are not really reimbursed for except maybe a token fee).
Sure, it would be fairly cheap for a big company to save every email ever sent between two people who worked there. However, then whenever somebody gets injured by the company's products there will be some email somewhere between two people joking (or debating) about the product's safety and that will be the "smoking gun" that proves the company covered up the dangers. In every place I've worked the fact is that anytime a decision is made there is somebody who doesn't think it is a good idea - and these people are always praised as prophets when things go wrong. However, if you wait to have unanimous consensus whenever you make a decision nothing would ever get done. A jury has the luxury of not needing to make a profit, and usually fails to find the right risk balance.
Since the TFA doesn't go into the study design I can't be sure how controlled this study was. It sounds like the subjects could choose whether to swear or to chant a neutral word. So, perhaps all they've proven is that the type of personality that makes one endure pain better also makes you more prone to swearing. If they picked 100 people at random, and then randomly assigned each one to the swearing vs neutral groups this would be less of an issue. Then again, it probably wouldn't hurt to do it both ways - and see what the results are.
These kinds of psychological tests can be very difficult to control.
Here's another test - test the subjects independently, and give each one the exact same word to shout. Tell half the subjects that the word is a really nasty expression in some-obscure-language, and tell half of them that the word just means soda. That would better test whether it is the idea of breaking a social norm that provides the relief.
You probably couldn't do this with a small molecule (though that isn't impossible), but some kind of passive membrane H+ transporter that was highly temperature sensitive would probably do the trick. Of course, it would also need to make it into the mitocondria. It should also have a limited half-life so that it could be controlled via dosing. Such a treatment would essentially have a built-in safety valve.
Control the impulses of your own intellect; else, you are of no greater worth to me than some mindless insect.
Your argument boils down to either:
1. I can control my ability to eat, therefore all humans must be able to control their abilities to eat. After all, we're all exactly the same.
2. I can control my ability to eat. Maybe you can't. That means you're less valuable than me. In fact, you're about as valuable as a mindless insect.
Have you considered that it is possible that there are some things that you can do that others cannot? Have you also considered that this does not actually endow you with greater value than those who cannot?
I think that just about any biologist would agree that it is fairly normal for populations of any creature to have a fairly diverse set of traits, because even some traits that appear debilitating can become a survival advantage in the right circumstances. As others have suggested, overeating only is a problem when food is plentiful and cheap - and that has only been true since the advent of refrigeration and mechanized farming - even 100 years ago food was much more expensive than it is today. Perhaps the reason you're surrounded by so many mindless insects is that 300 years ago their ancestors were doing fine while yours were barely scraping by (passing on genes that would eventually be more useful when circumstances changed).
RAID is a partially-effective data protection measure - it does protect against certain scenarios for data loss (namely loss of a physical device). Your data is generally safer on a RAID than not (although it does introduce a new point of failure - the RAID software/hardware itself). However, I do agree with what you're getting at - it certainly shouldn't be used as the sole method of protecting important data.
Not having read the source to the worm I can't be sure, but I suspect that if the worm writes out a string of u's to ever sector on the systems "hard drive" the RAID controller will dutifully replicate that data across all the redundant drives wiping out all traces of the original data. RAID is useless for protecting against logical data loss.
The only true backup solution involving RAID would be having a rotating stack of raid-1 mirror devices that are rebuilt and then taken offline. That really is just a fancy way of doing a physical-level drive mirroring operation. It isn't necessarily a bad approach either since it solves the problem of how to replicate the system without taking it down. On the other hand, it isn't guaranteed to be corruption-free unless at the end of the backup operation something is done to force all open applications to clean themselves up (disk-wise) right before the mirror is severed. In reality there are other approaches that work better.
I wasn't referring to selling API and having the pharmacists compound it - but instead just having one pill = one drug (at least as an option).
Compounding pharmacies simply aren't necessary and are an extra source of risk 99.9% of the time. Don't get me wrong - they certainly should exist because there are cases where they're a medical necessity. However, one per city (plus mail order) is probably plenty.
I suspect that most pharmacists had to take at least some kind of a class on compounding. The pharmacist also has the advantage that the formulated product doesn't need to have a 3-year shelf life so. However, it is a skill that I doubt most pharmacists are actually proficient in unless they're employed by pharmaceutical companies (or the very rare compounding pharmacy).
I tend to agree, but I am not a statistician. It seems like we're almost arguing a statement vs its contrapositive, when the truth of one is completely equivalent to the truth of the other.
Speaking of which, according to the US news, I had thought Canadas socialized medical system prevented this kind of innovation? How is it Canada and their inferior socialized medicine is able to develop this kind of drug?
The same way that all drug companies innovate without regard to their geographic location - they sell their drugs in the US where they can make lots of money (or they are taxpayer-funded). The fact that non-US-based drug companies do well really has nothing to do with socialized healthcare in their own nations - they make their money in the US the same as everybody else. If the US institutes price controls on patented drugs you'll see the Sanofi-Aventises and GSKs of the world collapse just as quickly as the Pfizers. Some smaller companies do more generic business (Apotex comes to mind since you used Canada as your illustration) - they could continue to do fine, but generic drugs and branded drugs are really two different businesses. You won't see much in the way of new drugs, however, unless they're completely paid for by taxpayers or otherwise incentivized (which is really no different from what we have now but by a different name).
The other thing that I should point out is that this drug hasn't gone through any of the expensive parts of the drug development process. They're only starting Phase I trials now, and that is where things just start to get expensive. The most expensive part of developing drugs is also typically the least creative part - the clinical trials.
Don't get me wrong - I'm all for changing the model by which drug research is paid for. However, don't think for a minute that the same thing can be done for 1/10th the cost or anything like that. The biggest expense in developing drugs is the clinical trials, and the only way that will get cheaper is if the state forces people (especially doctors) to participate in them.
1. Your database needs normalization. Almost all of that data should have been in one table, using fields to indicate what kind of statements they were.
2. Your use of random() isn't going to work - unless the random number generator just happens to generate a number that just happens to be the ID of one of your records. The typical way to do what you're doing is to order by random() and limit the query to one record.
True, but I think the point is that you STILL have easy access to the data - at least for reads.
I've seen plenty of data migration headaches dealing with proprietary software. Trust me - they're always 10X easier if the data is in a database. Otherwise you spend more time just trying to figure out what's in some proprietary data store than it would otherwise take to just migrate all the data in the database.
Often with a database you don't even need to migrate most of the data. In many cases you can just write up a few reports and point it at the final read-only snapshot of the database. This takes amost no effort.
I can see why Google needs BigTable - they had one problem with an ENORMOUS scale and their whole company was built on being able to solve that one problem. If they're using that same solution to run their payroll system, I'd be concerned. Ditching the RDBMS is a viable solution to certain problems, but those are probably only 0.00001% of the problems a company will encounter.
Why should I not be able to buy medication because 21 years ago you were unable to resist peer-pressure? I think that banning medication is an over-reaction to what is essentially a personal problem. With cigarettes there is at least the concern for secondhand smoke, which I think does justify some kind of intervention. I don't think that concern exists for medications other than antibiotics (which should be more controlled to prevent resistence).
I agree completely that if anybody could buy drugs that there would be all kinds of abuse. However, I think there will also be all kinds of benefits. The benefits extend to all of mankind, and the abuse only impacts those who willfully choose to abuse the drugs.
Don't get me wrong - I have nothing but compassion for your plight. However, I disagree that young people aren't capable of making good decisions, or that they shouldn't suffer the consequences of their decisions. I'm all for trying to better educate them. That's an entirely different topic, however. For starters, I think the artificial social structure we've created for children is harmful to them - I'd like to see children spending more time with adults than with their peers (which is almost certainly how a typical child's life worked until about 100 years ago). If you go back further it wasn't uncommon for 13 year olds to be full adults with all the rights/responsibilities this affords - even to the point of being sovereign heads of state. Maybe if we stop treating 16 year olds like they are in a nursery they'll stop acting like they are...
Frankly, I believe that all drugs should be available in single-compound formulations. Then doctors can work with their patients to mix and match as appropriate. I'm fine with having convenience formulations as well to make things easier, but that isn't a complete substitute to having single drugs.
As far as drug abuse goes - if I had my way every drug of every kind would be sold on the shelves of your local Walmart. If you want to mess up your body more power to you - just don't ask me or your insurer to pay for it.
Should we start requiring a plumbing license to buy drain cleaner because it isn't good to drink?
The law compels YOU, as a driver of a faster and heavier vehicle, to be aware of slower traffic and conduct yourself accordingly.
Absolutely. Of course, bicyclists are also expected to follow the law.
YOU are the jackass, not the cyclists.
Does not follow. Being a jackass has precisely nothing to do with following the law. You can follow the law and be a rude driver, and you can follow the law and be a courteous driver. Being courteous is also not the same as being slow.
While I have no problem paying reasonable taxes to maintain infrastructure, I am vehemently against using taxes to modify behavior!!
I agree completely - provided the definition of infrastructure includes the cost of maintaining bases and wars in the middle east (of course, that tarrif should be applied at the shipping level only on oil from those nations that incur these costs). Infrastructure should also include any real environmental costs incurred by the government to clean up messes that result from oil and its burning.
Taxes are a great way to get rid of externalities - which are a distortion of the market. I'm also fine with stuff like war being charged at a premium to actual costs - the "cost" of blowing up kids in some 3rd world nation can't be accounted for solely in terms of bullets and rations.
Now, if even after paying for all that stuff oil is still cheaper, then we should stick with it.
Couldn't agree with your general sentiments - Firefox is turning far too much into a "platform". Honestly, I could care less for AJAX 95% of the time as well - it tends to break almost all of the time if you aren't using the specific browser and OS versions it was originally targeted at. AJAX is used just so that the site developers can brag about it most of the time.
However, I really don't have any issues with using a SQL-based storage system. I mean, if you need to persistently store all kinds of data on disk, why not use some libraries written by people who've mastered it? The only thing I would do is ask the upstream provider to make the fsyncs a configurable option and then turn it off.
I really wish I could configure my OS to just ignore fsyncs issued by anything that wasn't running as root. I've got a server running quite a bit of stuff on software-based raid5. It works just fine until some program starts doing fsyncs every 15 seconds forcing the drives to start seeking like mad. If I didn't want a write cache I could just disable it at the OS level.
The funny thing is that half the time it LOWERS data-integrity. I use this server as a mythtv backend among other things and when mythtv forces an fsync it ends up blocking the disk, which then causes buffer overruns, which means dropped video data. Once I patched that out of the source it runs smooth as silk. And the only downside is that if the power dies that instead of losing 5 minutes of TV waiting for the power to come back and for everything to boot (from computers to satellite receivers / etc) I lose up to 5.5 minutes instead (anything not committed to disk). Big deal - the system ends up re-recording anything it can, and anything that airs only once is undoubtedly available online somewhere.
Fsync should ONLY be used by things like databases, and then only when you need to coordinate with stuff that is happening on a different box. Things like barriers and transactions are better solutions. The big thing that linux lacks is transactional filesystem operations so that software can just tell the OS what level of isolation/integrity each activity needs and the OS can deliver it in the most efficient way.
The issue is that publishers want to have their cake and eat it too.
If google said "Salt Lake Times - just set up your robots.txt and we'll ignore your site" the last thing they're going to do is block google. They WANT to be indexed and linked. They just also want to be paid for it. I call foul - you can't force people to buy a service for you, and you certainly can't do it by essentially offering it for free and then suing them over it.
If google stopped indexing some newspaper because of complaints that paper would probably collapse - they're obviously not making money in the print world, and if google drops them then they won't be making money online either.
Newspapers can't get over the fact that they can't be google. When somebody wants to know how the Yankees game went last night, they aren't going to go to the NYT website, log in to their paid subscriber account, and then search for the score there. They're going to just go to google. Prodigy, AOL, and Compuserve's proprietary content are gone - get over it...:)
It is just as relevant today in realtime mission-critical operations.
When you write realtime code on a realtime OS you design your application so that it can get a certain amount of work done in a certain amount of time. It is critical to the operation of the machine that those tasks get done exactly on time or sooner (or maybe not even sooner). As a result, you don't use some OS and language that just runs random tasks at random times and that at any given time you have no idea what is running. Instead everything is prioritized, and you know exactly what the cpu will be working on in any phase of your mission. You don't want the garbage collector to decide to wake up one second before touchdown / docking / intercept / etc.
Sure, the Apollo hardware was an extreme case of this - wouldn't be surprised if half of it was pure single-tasking (want to do 2 jobs at once - just have 2 CPUs). It probably used core memory and I wouldn't be surprised if the RAM had as much latency as a hard drive does today. Then again, at least the cores aren't going to be bit-flipped by a cosmic ray (unless it is the kind of ray that feels like you're hit on the head by a golf-ball when it hits you).:)
You can't have race conditions in a mission-critical realtime system - or at least you can't have them where it counts.
I'm not disputing this - but they could have just set the throttles to a fairly low setting and started a moderate descent down to a more manageable altitude and headed for the nearest usable landing strip. There is no reason they'd have to stay way up at FL380+.
Even so - I'm not under any illusions that this wasn't a dangerous situation. The fact is that the flight crew on a long haul flight has a huge amount of experience and it is unlikely that they didn't consider just taking it easy. Perhaps they got caught up in trying to resolve some problem and missed the signs they were going too fast.
To confuse things further - you're not actually using indicated airspeed but true airspeed.:)
The indicated airspeed at those altitudes is often on the order of 300 knots when the plane is really travelling around 500 knots relative to the air and 600 relative to the ground.
Put it this way - in space if you're travelling at mach 20-30 the airspeed indicator would probably read zero. When you hit an air molecule you're moving very fast relative to it, but so few hit the sensor that it reads zero. Anywhere in-between space and sea level the gauge acts accordingly...
Too little airspeed = too little lift and a stall (which is very dangerous on something as big as an airliner, though theoretically recoverable at that altitude granted you'll waste quite a bit of fuel and scare the living daylights out of the passengers).
Too much airspeed = shock waves rip the wings right off the plane. They're not fighters and while those wings actually are pretty strong they can only make them so heavy and be able to carry payload.
Well, the plane won't just fall out of the sky if they slow down a little - they should have erred on the side of slowing down and losing some alititude.
However, that isn't without issues if they don't resolve the problem quickly. At lower altitude they burn more fuel - which means there is a good chance they'll need to divert. That's better than disintegrating over an ocean, but it has risks of its own if you're 3 hours away from land.
While I agree with some of what you say - I don't buy it fully.
Ok, I'm making a smartphone. It should have a simple on-off button - not a 3-way toggle where you get data corruption if you switch it to the 3rd position. It should be hard to bypass the proper shutdown routine (removing the battery counts). So, then the counterproposal is - get rid of the need to do a proper shutdown. Sure, we can do that - no write cache and everything is transaction isolated so that corruptions are impossible. Now the thing needs 3X the hardware horsepower to have the same effective performance, which means the battery has to be 3X bigger to supply power, which means your smartphone is the size of a brick.
Likewise - something like an A320 is a complex beast - it depends on all kinds of machinery to make it work. Computers are just one more machine. It all needs to be properly engineered, but you can't just go back to pulling strings to warp the wings.
Now, I do think that primary instruments need to be operable in the absense of the computers/gyros/etc. At least the backup instruments. There should never be a question as to what the aircraft's speed, altitude, attitude, and heading are.
ALWAYS demand an itemised bill. ALWAYS. No exceptions, EVER
No problem sir - we'll just add on the itemized bill option for $4.95 per month. If you'd like it actually mailed to you instead of buried on a website I can do that too for only $3.95 more.
Oh, since you're concerned about getting raped on minutes you don't intend to use, for a mere $6.95 we'll let you set a limit on your usage so that you won't get billed for unintended calls. No, that won't help with roaming charges. We're looking into an experimental $14.95 service to handle those - would you like to be in our pilot group?
The phone company: all about finding clever ways to charge you for stuff that should be required as a matter of law...
Slashdot Mirror
Uh, I don't follow you. For a large company good control of deletions is mandatory, but how that makes this a non-issue escapes me. Does gmail provide for guaranteed deletion?
Pretty much every large company is de-facto required to rigorously delete materials that aren't required by law to be kept. The penalty for not doing so is to be buried alive in discovery whenever you are sued (which probably happens about once a week for a big company).
If a law says that the accounting records used to create a quarterly statement needs to be kept for three years, then you want to keep it for three years (with no meaningful chance of loss), and then make every remote trace of them disappear completely one day later (or as close to this as possible).
This has nothing to do with covering up wrongful behavior. The fact is that you can comply with the letter of the law today, and then be second-guessed about some decision you made with perfect hindsight 10 years from now. So, unless the law states that those records must be preserved for 10 years, then you don't want those records.
As an analogy from the sysadmin world - consider system access logs at an ISP (something here most people would intuitively understand). If you are an ISP you want to keep your logs long enough to handle billing disputes or to be able to identify abuses, but you don't want a 10 year record of everything every one of your customers have ever done online. Then, when some man files a subpoena to find out when a relationship between his soon-to-be-ex-wife and her boyfriend started you can just say that your records only go back 30 days so that you don't have anything, instead of having to go digging through the vaults to find backup tapes, and then show up in court to testify about how those records were maintained (a cost you are not really reimbursed for except maybe a token fee).
Sure, it would be fairly cheap for a big company to save every email ever sent between two people who worked there. However, then whenever somebody gets injured by the company's products there will be some email somewhere between two people joking (or debating) about the product's safety and that will be the "smoking gun" that proves the company covered up the dangers. In every place I've worked the fact is that anytime a decision is made there is somebody who doesn't think it is a good idea - and these people are always praised as prophets when things go wrong. However, if you wait to have unanimous consensus whenever you make a decision nothing would ever get done. A jury has the luxury of not needing to make a profit, and usually fails to find the right risk balance.
Since the TFA doesn't go into the study design I can't be sure how controlled this study was. It sounds like the subjects could choose whether to swear or to chant a neutral word. So, perhaps all they've proven is that the type of personality that makes one endure pain better also makes you more prone to swearing. If they picked 100 people at random, and then randomly assigned each one to the swearing vs neutral groups this would be less of an issue. Then again, it probably wouldn't hurt to do it both ways - and see what the results are.
These kinds of psychological tests can be very difficult to control.
Here's another test - test the subjects independently, and give each one the exact same word to shout. Tell half the subjects that the word is a really nasty expression in some-obscure-language, and tell half of them that the word just means soda. That would better test whether it is the idea of breaking a social norm that provides the relief.
You probably couldn't do this with a small molecule (though that isn't impossible), but some kind of passive membrane H+ transporter that was highly temperature sensitive would probably do the trick. Of course, it would also need to make it into the mitocondria. It should also have a limited half-life so that it could be controlled via dosing. Such a treatment would essentially have a built-in safety valve.
Control the impulses of your own intellect; else, you are of no greater worth to me than some mindless insect.
Your argument boils down to either:
1. I can control my ability to eat, therefore all humans must be able to control their abilities to eat. After all, we're all exactly the same.
2. I can control my ability to eat. Maybe you can't. That means you're less valuable than me. In fact, you're about as valuable as a mindless insect.
Have you considered that it is possible that there are some things that you can do that others cannot? Have you also considered that this does not actually endow you with greater value than those who cannot?
I think that just about any biologist would agree that it is fairly normal for populations of any creature to have a fairly diverse set of traits, because even some traits that appear debilitating can become a survival advantage in the right circumstances. As others have suggested, overeating only is a problem when food is plentiful and cheap - and that has only been true since the advent of refrigeration and mechanized farming - even 100 years ago food was much more expensive than it is today. Perhaps the reason you're surrounded by so many mindless insects is that 300 years ago their ancestors were doing fine while yours were barely scraping by (passing on genes that would eventually be more useful when circumstances changed).
RAID is a partially-effective data protection measure - it does protect against certain scenarios for data loss (namely loss of a physical device). Your data is generally safer on a RAID than not (although it does introduce a new point of failure - the RAID software/hardware itself). However, I do agree with what you're getting at - it certainly shouldn't be used as the sole method of protecting important data.
Not having read the source to the worm I can't be sure, but I suspect that if the worm writes out a string of u's to ever sector on the systems "hard drive" the RAID controller will dutifully replicate that data across all the redundant drives wiping out all traces of the original data. RAID is useless for protecting against logical data loss.
The only true backup solution involving RAID would be having a rotating stack of raid-1 mirror devices that are rebuilt and then taken offline. That really is just a fancy way of doing a physical-level drive mirroring operation. It isn't necessarily a bad approach either since it solves the problem of how to replicate the system without taking it down. On the other hand, it isn't guaranteed to be corruption-free unless at the end of the backup operation something is done to force all open applications to clean themselves up (disk-wise) right before the mirror is severed. In reality there are other approaches that work better.
I wasn't referring to selling API and having the pharmacists compound it - but instead just having one pill = one drug (at least as an option).
Compounding pharmacies simply aren't necessary and are an extra source of risk 99.9% of the time. Don't get me wrong - they certainly should exist because there are cases where they're a medical necessity. However, one per city (plus mail order) is probably plenty.
I suspect that most pharmacists had to take at least some kind of a class on compounding. The pharmacist also has the advantage that the formulated product doesn't need to have a 3-year shelf life so. However, it is a skill that I doubt most pharmacists are actually proficient in unless they're employed by pharmaceutical companies (or the very rare compounding pharmacy).
I tend to agree, but I am not a statistician. It seems like we're almost arguing a statement vs its contrapositive, when the truth of one is completely equivalent to the truth of the other.
Good point. Are gas turbines street-legal? You'd think that this would be a much more effective design at these kinds of speeds.
Speaking of which, according to the US news, I had thought Canadas socialized medical system prevented this kind of innovation?
How is it Canada and their inferior socialized medicine is able to develop this kind of drug?
The same way that all drug companies innovate without regard to their geographic location - they sell their drugs in the US where they can make lots of money (or they are taxpayer-funded). The fact that non-US-based drug companies do well really has nothing to do with socialized healthcare in their own nations - they make their money in the US the same as everybody else. If the US institutes price controls on patented drugs you'll see the Sanofi-Aventises and GSKs of the world collapse just as quickly as the Pfizers. Some smaller companies do more generic business (Apotex comes to mind since you used Canada as your illustration) - they could continue to do fine, but generic drugs and branded drugs are really two different businesses. You won't see much in the way of new drugs, however, unless they're completely paid for by taxpayers or otherwise incentivized (which is really no different from what we have now but by a different name).
The other thing that I should point out is that this drug hasn't gone through any of the expensive parts of the drug development process. They're only starting Phase I trials now, and that is where things just start to get expensive. The most expensive part of developing drugs is also typically the least creative part - the clinical trials.
Don't get me wrong - I'm all for changing the model by which drug research is paid for. However, don't think for a minute that the same thing can be done for 1/10th the cost or anything like that. The biggest expense in developing drugs is the clinical trials, and the only way that will get cheaper is if the state forces people (especially doctors) to participate in them.
Most people aren't able to diagnose drug psychosis on themselves, much less treat it correctly.
Agreed.
How is this relevant to whether I (who don't abuse drugs) should be able to buy medications without a prescription?
I hate to nitpick, but:
1. Your database needs normalization. Almost all of that data should have been in one table, using fields to indicate what kind of statements they were.
2. Your use of random() isn't going to work - unless the random number generator just happens to generate a number that just happens to be the ID of one of your records. The typical way to do what you're doing is to order by random() and limit the query to one record.
True, but I think the point is that you STILL have easy access to the data - at least for reads.
I've seen plenty of data migration headaches dealing with proprietary software. Trust me - they're always 10X easier if the data is in a database. Otherwise you spend more time just trying to figure out what's in some proprietary data store than it would otherwise take to just migrate all the data in the database.
Often with a database you don't even need to migrate most of the data. In many cases you can just write up a few reports and point it at the final read-only snapshot of the database. This takes amost no effort.
I can see why Google needs BigTable - they had one problem with an ENORMOUS scale and their whole company was built on being able to solve that one problem. If they're using that same solution to run their payroll system, I'd be concerned. Ditching the RDBMS is a viable solution to certain problems, but those are probably only 0.00001% of the problems a company will encounter.
Why should I not be able to buy medication because 21 years ago you were unable to resist peer-pressure? I think that banning medication is an over-reaction to what is essentially a personal problem. With cigarettes there is at least the concern for secondhand smoke, which I think does justify some kind of intervention. I don't think that concern exists for medications other than antibiotics (which should be more controlled to prevent resistence).
I agree completely that if anybody could buy drugs that there would be all kinds of abuse. However, I think there will also be all kinds of benefits. The benefits extend to all of mankind, and the abuse only impacts those who willfully choose to abuse the drugs.
Don't get me wrong - I have nothing but compassion for your plight. However, I disagree that young people aren't capable of making good decisions, or that they shouldn't suffer the consequences of their decisions. I'm all for trying to better educate them. That's an entirely different topic, however. For starters, I think the artificial social structure we've created for children is harmful to them - I'd like to see children spending more time with adults than with their peers (which is almost certainly how a typical child's life worked until about 100 years ago). If you go back further it wasn't uncommon for 13 year olds to be full adults with all the rights/responsibilities this affords - even to the point of being sovereign heads of state. Maybe if we stop treating 16 year olds like they are in a nursery they'll stop acting like they are...
Frankly, I believe that all drugs should be available in single-compound formulations. Then doctors can work with their patients to mix and match as appropriate. I'm fine with having convenience formulations as well to make things easier, but that isn't a complete substitute to having single drugs.
As far as drug abuse goes - if I had my way every drug of every kind would be sold on the shelves of your local Walmart. If you want to mess up your body more power to you - just don't ask me or your insurer to pay for it.
Should we start requiring a plumbing license to buy drain cleaner because it isn't good to drink?
The law compels YOU, as a driver of a faster and heavier vehicle, to be aware of slower traffic and conduct yourself accordingly.
Absolutely. Of course, bicyclists are also expected to follow the law.
YOU are the jackass, not the cyclists.
Does not follow. Being a jackass has precisely nothing to do with following the law. You can follow the law and be a rude driver, and you can follow the law and be a courteous driver. Being courteous is also not the same as being slow.
While I have no problem paying reasonable taxes to maintain infrastructure, I am vehemently against using taxes to modify behavior!!
I agree completely - provided the definition of infrastructure includes the cost of maintaining bases and wars in the middle east (of course, that tarrif should be applied at the shipping level only on oil from those nations that incur these costs). Infrastructure should also include any real environmental costs incurred by the government to clean up messes that result from oil and its burning.
Taxes are a great way to get rid of externalities - which are a distortion of the market. I'm also fine with stuff like war being charged at a premium to actual costs - the "cost" of blowing up kids in some 3rd world nation can't be accounted for solely in terms of bullets and rations.
Now, if even after paying for all that stuff oil is still cheaper, then we should stick with it.
Couldn't agree with your general sentiments - Firefox is turning far too much into a "platform". Honestly, I could care less for AJAX 95% of the time as well - it tends to break almost all of the time if you aren't using the specific browser and OS versions it was originally targeted at. AJAX is used just so that the site developers can brag about it most of the time.
However, I really don't have any issues with using a SQL-based storage system. I mean, if you need to persistently store all kinds of data on disk, why not use some libraries written by people who've mastered it? The only thing I would do is ask the upstream provider to make the fsyncs a configurable option and then turn it off.
I really wish I could configure my OS to just ignore fsyncs issued by anything that wasn't running as root. I've got a server running quite a bit of stuff on software-based raid5. It works just fine until some program starts doing fsyncs every 15 seconds forcing the drives to start seeking like mad. If I didn't want a write cache I could just disable it at the OS level.
The funny thing is that half the time it LOWERS data-integrity. I use this server as a mythtv backend among other things and when mythtv forces an fsync it ends up blocking the disk, which then causes buffer overruns, which means dropped video data. Once I patched that out of the source it runs smooth as silk. And the only downside is that if the power dies that instead of losing 5 minutes of TV waiting for the power to come back and for everything to boot (from computers to satellite receivers / etc) I lose up to 5.5 minutes instead (anything not committed to disk). Big deal - the system ends up re-recording anything it can, and anything that airs only once is undoubtedly available online somewhere.
Fsync should ONLY be used by things like databases, and then only when you need to coordinate with stuff that is happening on a different box. Things like barriers and transactions are better solutions. The big thing that linux lacks is transactional filesystem operations so that software can just tell the OS what level of isolation/integrity each activity needs and the OS can deliver it in the most efficient way.
The issue is that publishers want to have their cake and eat it too.
If google said "Salt Lake Times - just set up your robots.txt and we'll ignore your site" the last thing they're going to do is block google. They WANT to be indexed and linked. They just also want to be paid for it. I call foul - you can't force people to buy a service for you, and you certainly can't do it by essentially offering it for free and then suing them over it.
If google stopped indexing some newspaper because of complaints that paper would probably collapse - they're obviously not making money in the print world, and if google drops them then they won't be making money online either.
Newspapers can't get over the fact that they can't be google. When somebody wants to know how the Yankees game went last night, they aren't going to go to the NYT website, log in to their paid subscriber account, and then search for the score there. They're going to just go to google. Prodigy, AOL, and Compuserve's proprietary content are gone - get over it... :)
It is just as relevant today in realtime mission-critical operations.
When you write realtime code on a realtime OS you design your application so that it can get a certain amount of work done in a certain amount of time. It is critical to the operation of the machine that those tasks get done exactly on time or sooner (or maybe not even sooner). As a result, you don't use some OS and language that just runs random tasks at random times and that at any given time you have no idea what is running. Instead everything is prioritized, and you know exactly what the cpu will be working on in any phase of your mission. You don't want the garbage collector to decide to wake up one second before touchdown / docking / intercept / etc.
Sure, the Apollo hardware was an extreme case of this - wouldn't be surprised if half of it was pure single-tasking (want to do 2 jobs at once - just have 2 CPUs). It probably used core memory and I wouldn't be surprised if the RAM had as much latency as a hard drive does today. Then again, at least the cores aren't going to be bit-flipped by a cosmic ray (unless it is the kind of ray that feels like you're hit on the head by a golf-ball when it hits you). :)
You can't have race conditions in a mission-critical realtime system - or at least you can't have them where it counts.
I'm not disputing this - but they could have just set the throttles to a fairly low setting and started a moderate descent down to a more manageable altitude and headed for the nearest usable landing strip. There is no reason they'd have to stay way up at FL380+.
Even so - I'm not under any illusions that this wasn't a dangerous situation. The fact is that the flight crew on a long haul flight has a huge amount of experience and it is unlikely that they didn't consider just taking it easy. Perhaps they got caught up in trying to resolve some problem and missed the signs they were going too fast.
To confuse things further - you're not actually using indicated airspeed but true airspeed. :)
The indicated airspeed at those altitudes is often on the order of 300 knots when the plane is really travelling around 500 knots relative to the air and 600 relative to the ground.
Put it this way - in space if you're travelling at mach 20-30 the airspeed indicator would probably read zero. When you hit an air molecule you're moving very fast relative to it, but so few hit the sensor that it reads zero. Anywhere in-between space and sea level the gauge acts accordingly...
Both are important.
Too little airspeed = too little lift and a stall (which is very dangerous on something as big as an airliner, though theoretically recoverable at that altitude granted you'll waste quite a bit of fuel and scare the living daylights out of the passengers).
Too much airspeed = shock waves rip the wings right off the plane. They're not fighters and while those wings actually are pretty strong they can only make them so heavy and be able to carry payload.
Well, the plane won't just fall out of the sky if they slow down a little - they should have erred on the side of slowing down and losing some alititude.
However, that isn't without issues if they don't resolve the problem quickly. At lower altitude they burn more fuel - which means there is a good chance they'll need to divert. That's better than disintegrating over an ocean, but it has risks of its own if you're 3 hours away from land.
While I agree with some of what you say - I don't buy it fully.
Ok, I'm making a smartphone. It should have a simple on-off button - not a 3-way toggle where you get data corruption if you switch it to the 3rd position. It should be hard to bypass the proper shutdown routine (removing the battery counts). So, then the counterproposal is - get rid of the need to do a proper shutdown. Sure, we can do that - no write cache and everything is transaction isolated so that corruptions are impossible. Now the thing needs 3X the hardware horsepower to have the same effective performance, which means the battery has to be 3X bigger to supply power, which means your smartphone is the size of a brick.
Likewise - something like an A320 is a complex beast - it depends on all kinds of machinery to make it work. Computers are just one more machine. It all needs to be properly engineered, but you can't just go back to pulling strings to warp the wings.
Now, I do think that primary instruments need to be operable in the absense of the computers/gyros/etc. At least the backup instruments. There should never be a question as to what the aircraft's speed, altitude, attitude, and heading are.
ALWAYS demand an itemised bill. ALWAYS. No exceptions, EVER
No problem sir - we'll just add on the itemized bill option for $4.95 per month. If you'd like it actually mailed to you instead of buried on a website I can do that too for only $3.95 more.
Oh, since you're concerned about getting raped on minutes you don't intend to use, for a mere $6.95 we'll let you set a limit on your usage so that you won't get billed for unintended calls. No, that won't help with roaming charges. We're looking into an experimental $14.95 service to handle those - would you like to be in our pilot group?
The phone company: all about finding clever ways to charge you for stuff that should be required as a matter of law...