You'll have to pardon me if I hold in low regard a distro in which you can be locked out of your system merely for following the install instructions as written. I wanted to use Ubuntu, I really did, but poor design really grates on me, and this a great example of not-so-great design. As for GRUB, not all distros use the same version of it. And if it's error isn't even the right one, that's even worse. I wish I had gone into software engineering rather than real engineering, where low quality is better tolerated.
When I had GRUB installed, the BIOS would load, then it would go into GRUB, have the error, and then refuse further input. The way GRUB could have been written is that when it returns an error, it spits control back to whatever would load in the absence of GRUB having been installed. Don't tell me that's not possible. And if it weren't possible, the fact that GRUB failure can be that catastrophic tells me that, being an engineer, it would warrant much more testing before release. But in all fairness, I work in a field where my solutions have to work, and I can't just shrug when the helicopter crashes.
In that I write computer programs, yes. I'm also an engineer. A real engineer, not a "software engineer".
You're clearly not one that understands bootloaders. How do you recover from every possible error in only 512 bytes?
You toss control back to whatever would otherwise load when it fails. Now, self-proclaimed programmer, let me turn this back around: wouldn't you agree that programs with more catastrophic results upon failure deserve more testing?
Right. I somehow inadvertently gave it a command that says "GRUB: complain about reading a hard drive that works fine and then refuse input". What a goose I am.
When I installed Ubuntu, and chose the option to install GRUB for dual booting, whenever my computer booted, GRUB would try to load and then give "error 25", and then refuse ALL further input, locking me out of BOTH operating systems!!!!! Now, I can sympathize with errors. I cannot sympathize with errors that LOCK ME OUT OF MY OWN GODDAMN BOX. And yes, I have researched the error and asked for help one the Ubuntu forums, and all I got was "shame on you for following the instructions on the web site" and "why didn't you take [precaution X mentioned nowhere in the install instructions]". By the way, Error 25 refers to a hard disk read error. One small problem though: that hard drive works fine and (once I got rid of GRUB and Ubuntu) have been using it to store data without any problems at all. No problems reading it. Hm.
If you want to make Ubuntu user friendly, how about making it so that you can bypass GRUB instead of being locked out of your system just because it has an error? That to me is the epitome of bad programming.
Hm, flamebait... nah, wasn't that. Trolling? No. Redundant? Hardly. Let's see if I can refute the logic! Hm, can't. What to do, what to do. Oh, I know! I'll mod it overrated! That'll show him!!! Never mind that it wasn't rated in the first place.
If the only way you have to defend socialism in inaccurate moddings of people who argue against it, maybe you need to rethink your political philosophy?
Which is stupid, of course; unions are in fact an admirably free-market solution to the problem of employer-employee conflicts. In fact, I'd go so far as to say that asking someone who calls himself a libertarian about his opinion of organized labor is a good way to distinguish between true libertarians on the one hand, and right-wingers who call themselves libertarians because it's fashionable in certain circles on the other.
In theory, you're right. But the only way I can see a union using "free market" means to get its way is to threaten to quit en masse if the employer didn't meet their demands. Can you name *any* union whatsoever that has done only this? I can't. In reality, unions have never used free market methods. They block enterances and refuse to leave. They harass customers and resort to violence. Often times the quitting is in violation of a contract.
So it's true that unions can be a free market method, but any union that did this would be as different from every other union as night and day. (not going to bother your with a clever simile)
And it's simply not true that unions "improved the workplace". First of all, wage growth was much faster in late 19th century America vs. Europe, while the former was virtually devoid of unions, while the latter was heavily unionized. This only makes sense. Would you really want to invest somewhere where your employees will randomly decide to just shut down at a critical moment? This is why employers heavily discount, in calculations, the expected productivity of unionized workforces. Over time, unions scare off investement, like has been done in the Rust Belt, leading to lower wages. (And I'm not going to bother you with the stories people who can't turn off their own light switches because of union regulations.)
Look at it this way -- you buy labor, just as surely as any employer does. Now imagine that a supermarket had a history of jacking up prices right before you checked out, leading you to either pay more, or waste the time you spent collecting the goods for purchase there. Is that going to make you want to shop there? That's exactly what it looks like from the investor's standpoint. Now, you may still go there, but only if they offered lower prices to begin with. This is exactly why unions impede wage growth.
That socialist governments negligently give out social services, and free market economies are far more sustainable and desirable to work in is exactly what I would hope it would prove. If a socialist economy only "works" when all vesitiges of capitalism are eradicated, or it cuts off trade and workers are imprisoned there, what the fuck does that say about socialism?
The problem with banding and the like (from the employee's point of view) is that everything is generalized to the point where excellence cannot be rewarded.
Right, and there's a specific reason this resulted. Numerous discrimination lawsuits, and, more importantly, the fear thereof, make it so they have to have regimented systems that leave little room for incentive payment. If you pay $NON_PRIVELEGED_RACE_GENDER person more than "equal" $PRIVELEGED_RACE_GENDER then, whoa Nellie, get ready to get sued. How do you prove that you genuinely deem one person's work more valuable than another's? You can't. Since a) value is in the mind of the valuer and b) it is hard to justify valuations to outsiders, you're ultimately punished for your thoughts. If you "underpay" someone for a "good reason", that's AOK. If you underpay them for a "bad reason", that's not OK.
So to CYA, you have to have these thoughtless, rote processes for determining pay, which is a simple function of time worked, education (!), and work experience before joining. Any subjective metric can be painted as discriminatory.
A stockholder owns a small piece of a company. A piece of paper. They don't do the work that makes the company go, they don't make decisions that make the company go, and they don't have to buy the product. So why is this piece of paper so damn important?
Yeah, and what's the deal with money, anyway? Why does Walmart give you stuff just for pieces of green cotton paper? You didn't grow the crops. You didn't drive the products to the store. You didn't build the items. So why are they giving away that stuff to people who *never* did any work for them, simply cause they shovel green at them? Or sign little slips with some numbers on them? Why are those green pieces of paper so damn important?
Reductionism: destroying understanding since 200 BC.
I'd explain to you how savings, investment, entrepreneurism, and financial markets work, but come on -- shouldn't you have done that first before saying something so asinine?
Exactly. I'm so glad to find out (based on your post and the moddings) that I'm not alone in thinking this. Excellent example about "AJAX technology". It's ridiculous how people deem any computer program a separate "technology". Despite the reductionist view of the other poster, while that may meet some dictionary's literal definition, it's really not how people use the term and connotes something very different. It's exactly this "all computer programs = new technology" mentality that leads the USPTO to grant software patents. 'Cause hey -- it's "new technology", right?
. It'd be nice to read a calm, clear explanation of what it really does, and how it's different from earlier versions.
He wants a simple summary. If he wanted to wade through the whole thing, he would have done it already. Thanks for completely missing the point of his question.
By not providing such information, Google is leaving folks uncertain - now, honestly, if your data was good you'd release it because it would do good to your stock price. If you aren't, I'd be worried about what else is going on, and that is most definitely not a good sign.
Wow, I never thought I'd see someone who either knew this little, or this much game theory. You wouldn't necessarily release data that was good *for the precise reason that you just gave* -- it causes "no data" to signal "bad". By never giving any data, you provide no signals. If you just provide good data, any time you don't, it will be assumed the data is bad.
If you have "some solid leads, some strong hypotheses, but realize that another ten-year study is in order to truly rule out some key issues" then your study should claim that you "found an interesting correlation", not that you "found that coffee may reduce the risk of damage to health" (or whatever the case may be). That way you can convey information which may be useful to others while not overstating your results.
(And while we're at it, why is all this raw data the exclusive domain of a few people? If all you did was become the first person who, upon searching a database, found a cool correlation, why isn't this same data open to others? Why should we have to wait until a busy scientist gets around to checking out a few correlations before we find out about things like this?)
You seem to have this impression that the job of a scientist is to gather data. It's not. That's the job description of a lab worker. The job of a scientist is to properly design experiments, to come up with innovative ways to rule out alternate hypotheses that a layman can come with in ten seconds, to state precisely the caveats behind the results. (A great example of this in the social sciences was an economist who investigated the impact of women having children on their future income. To control for the possibility that women who want to have children at age X also share income-enhancing or income worsening effects, she focused on a wide array of data from women who had miscarried, and therefore "were the type to have kids" but didn't.)
It's true as you say that science is an ever-growing process, but please -- when the conclusions flip every few years, that should be taken as a signal that your conclusions have been premature and both sides need to revise estimates about how much data they need before calling something a theory. It's also correct that humans need validation, but that's precisely why we should be suspicious of anything that suggest something we want to believe. Coffee is good for you? If it's too good to be true, make damn sure it is.
Any well-funded fool can gather data. Figuring out how to gather the right data... that takes a scientist. Any any scientist should know how unreliable a self-reported tiny change in risk over 30 years is going to be.
You think this is the media's fault? You don't think scientists overstate their pop-wisdon-verifying results, knowing they'll easily be misinterpreted, in order to get more attention?
That's an extreme stretch of the normal use of the term "technology". They thought of systematic way of warning people about phishing sites by compiling a list of them. Good for you. But computer programs, databases, and browsers have existed for a long time. This isn't a "new technology". It's a computer program. I know, you probably think it's a minor point, but keep in mind that Microsoft considers removing its own damn bugs to be "new technology" (NT).
Thinking up ways to warn people about phishing sites isn't "new technology".
When Google went public, they became obligated to the stockholders, regardless of any preexisting 'mantras'.
Actually, to borrow from Fight Club, you choose their level of involvement. If you issue stock with the explicit caveat that you're not going to listen to stockholder pressure for short-term gain, and you're going to limit their ability to control the company, that's exactly what stockholders get, and indeed deserve, for buying in. They have every right to set the terms of their issuance of stock. If you didn't want that kind of deal, you were under no obligation to buy in. No one was bait-and-switched. It's just that Google's operation style does fit in the old-boys' paradigm. (Sorry for using that word, I couldn't help it.)
It's not impossible to show that an attribute (enviornmental or genetic) can affect relative risk for any phenotype (in this case, heart disease.) You need a strong understanding of statistics, a well designed study, and a large enough number of patients to see weaker effects.
True, but did they do that here? How often do they even do that? If they really controlled as well as you claim, you would not see anywhere near the number of reversals as there have been. Plus, whenever they give one of these studies, they always, always add "Oh, but there's always the possibility that this apparent health gain from coffee merely arises from other healthy activities that correlated with coffee drinking." But if you didn't find a very compelling reason to rule this out, your work never should have been published!
And how precisely would you isolate this effect over *thirty years*? You would have to come up with, by my rought estimates, 100,000 people for a remotely useful study that could control for all the other factors and phony correlations that the GP came up with after just a few minutes of thinking. Given the self-reporting they use, the multiple alternate correlations, the number of factors that enter over 30 years, the sample size needed for controls, and the small magnitude of the supposed effect (if you cut in half the risk of heart disease, that's still only 1 in 100,000 fewer cases), it really makes me suspicious of the supposed *confidence* they claim to have in their findings.
The problem is not that scientists change their minds. It's that the correlation was too flimsy to justify publishing in the first place. Every time they publish something like this, they always add "but it's still not clear if this correlation between [engaging in activity thought to be unhealthy] and better heart/longevity is because of the [unhealthy activity] or because people who do that tend to also [engage in separate, healthy activity]". Um, EXCUSE ME? If you didn't find a damn good reason to rule that out, your study should not have passed peer review. Of course you shouldn't wait until all doubt is removed before publishing -- but you should also at least try to look like you did your homework the first time around.
Slashdot Mirror
You'll have to pardon me if I hold in low regard a distro in which you can be locked out of your system merely for following the install instructions as written. I wanted to use Ubuntu, I really did, but poor design really grates on me, and this a great example of not-so-great design. As for GRUB, not all distros use the same version of it. And if it's error isn't even the right one, that's even worse. I wish I had gone into software engineering rather than real engineering, where low quality is better tolerated.
When I had GRUB installed, the BIOS would load, then it would go into GRUB, have the error, and then refuse further input. The way GRUB could have been written is that when it returns an error, it spits control back to whatever would load in the absence of GRUB having been installed. Don't tell me that's not possible. And if it weren't possible, the fact that GRUB failure can be that catastrophic tells me that, being an engineer, it would warrant much more testing before release. But in all fairness, I work in a field where my solutions have to work, and I can't just shrug when the helicopter crashes.
Here's a little secret: Windows was able to load before I installed GRUB. Take that hint, and answer your question.
Are you a programmer?
In that I write computer programs, yes. I'm also an engineer. A real engineer, not a "software engineer".
You're clearly not one that understands bootloaders. How do you recover from every possible error in only 512 bytes?
You toss control back to whatever would otherwise load when it fails. Now, self-proclaimed programmer, let me turn this back around: wouldn't you agree that programs with more catastrophic results upon failure deserve more testing?
Right. I somehow inadvertently gave it a command that says "GRUB: complain about reading a hard drive that works fine and then refuse input". What a goose I am.
When I installed Ubuntu, and chose the option to install GRUB for dual booting, whenever my computer booted, GRUB would try to load and then give "error 25", and then refuse ALL further input, locking me out of BOTH operating systems!!!!! Now, I can sympathize with errors. I cannot sympathize with errors that LOCK ME OUT OF MY OWN GODDAMN BOX. And yes, I have researched the error and asked for help one the Ubuntu forums, and all I got was "shame on you for following the instructions on the web site" and "why didn't you take [precaution X mentioned nowhere in the install instructions]". By the way, Error 25 refers to a hard disk read error. One small problem though: that hard drive works fine and (once I got rid of GRUB and Ubuntu) have been using it to store data without any problems at all. No problems reading it. Hm.
If you want to make Ubuntu user friendly, how about making it so that you can bypass GRUB instead of being locked out of your system just because it has an error? That to me is the epitome of bad programming.
Hm, flamebait... nah, wasn't that. Trolling? No. Redundant? Hardly. Let's see if I can refute the logic! Hm, can't. What to do, what to do. Oh, I know! I'll mod it overrated! That'll show him!!! Never mind that it wasn't rated in the first place.
If the only way you have to defend socialism in inaccurate moddings of people who argue against it, maybe you need to rethink your political philosophy?
Which is stupid, of course; unions are in fact an admirably free-market solution to the problem of employer-employee conflicts. In fact, I'd go so far as to say that asking someone who calls himself a libertarian about his opinion of organized labor is a good way to distinguish between true libertarians on the one hand, and right-wingers who call themselves libertarians because it's fashionable in certain circles on the other.
In theory, you're right. But the only way I can see a union using "free market" means to get its way is to threaten to quit en masse if the employer didn't meet their demands. Can you name *any* union whatsoever that has done only this? I can't. In reality, unions have never used free market methods. They block enterances and refuse to leave. They harass customers and resort to violence. Often times the quitting is in violation of a contract.
So it's true that unions can be a free market method, but any union that did this would be as different from every other union as night and day. (not going to bother your with a clever simile)
And it's simply not true that unions "improved the workplace". First of all, wage growth was much faster in late 19th century America vs. Europe, while the former was virtually devoid of unions, while the latter was heavily unionized. This only makes sense. Would you really want to invest somewhere where your employees will randomly decide to just shut down at a critical moment? This is why employers heavily discount, in calculations, the expected productivity of unionized workforces. Over time, unions scare off investement, like has been done in the Rust Belt, leading to lower wages. (And I'm not going to bother you with the stories people who can't turn off their own light switches because of union regulations.)
Look at it this way -- you buy labor, just as surely as any employer does. Now imagine that a supermarket had a history of jacking up prices right before you checked out, leading you to either pay more, or waste the time you spent collecting the goods for purchase there. Is that going to make you want to shop there? That's exactly what it looks like from the investor's standpoint. Now, you may still go there, but only if they offered lower prices to begin with. This is exactly why unions impede wage growth.
That socialist governments negligently give out social services, and free market economies are far more sustainable and desirable to work in is exactly what I would hope it would prove. If a socialist economy only "works" when all vesitiges of capitalism are eradicated, or it cuts off trade and workers are imprisoned there, what the fuck does that say about socialism?
Ah! So that's what that stuff is!
Er... the whole point of it is to try both, very close to each other. But I appreciate your input and the time you took to read it.
Yeah, I was kinda surprised by that, releasing a product in beta is an unusual move for Google.
The problem with banding and the like (from the employee's point of view) is that everything is generalized to the point where excellence cannot be rewarded.
Right, and there's a specific reason this resulted. Numerous discrimination lawsuits, and, more importantly, the fear thereof, make it so they have to have regimented systems that leave little room for incentive payment. If you pay $NON_PRIVELEGED_RACE_GENDER person more than "equal" $PRIVELEGED_RACE_GENDER then, whoa Nellie, get ready to get sued. How do you prove that you genuinely deem one person's work more valuable than another's? You can't. Since a) value is in the mind of the valuer and b) it is hard to justify valuations to outsiders, you're ultimately punished for your thoughts. If you "underpay" someone for a "good reason", that's AOK. If you underpay them for a "bad reason", that's not OK.
So to CYA, you have to have these thoughtless, rote processes for determining pay, which is a simple function of time worked, education (!), and work experience before joining. Any subjective metric can be painted as discriminatory.
A stockholder owns a small piece of a company. A piece of paper. They don't do the work that makes the company go, they don't make decisions that make the company go, and they don't have to buy the product. So why is this piece of paper so damn important?
Yeah, and what's the deal with money, anyway? Why does Walmart give you stuff just for pieces of green cotton paper? You didn't grow the crops. You didn't drive the products to the store. You didn't build the items. So why are they giving away that stuff to people who *never* did any work for them, simply cause they shovel green at them? Or sign little slips with some numbers on them? Why are those green pieces of paper so damn important?
Reductionism: destroying understanding since 200 BC.
I'd explain to you how savings, investment, entrepreneurism, and financial markets work, but come on -- shouldn't you have done that first before saying something so asinine?
Wow! Brilliant! I can save on labor costs simply by telling employees they're being competitively paid! Why didn't I think of this before!
"I'll fix your toilet for $150."
"Nah, $75 is a competitive payment."
"It is? Oh, well, in that case, $75 is find and dandy."
Oh, that's right, because it doesn't work.
Exactly. Just like how we're running out of horses, which are needed for long-distance travel. As they decline, so does the human race.
Oh, sorry, got stuck in the late 19th century for a second there.
Exactly. I'm so glad to find out (based on your post and the moddings) that I'm not alone in thinking this. Excellent example about "AJAX technology". It's ridiculous how people deem any computer program a separate "technology". Despite the reductionist view of the other poster, while that may meet some dictionary's literal definition, it's really not how people use the term and connotes something very different. It's exactly this "all computer programs = new technology" mentality that leads the USPTO to grant software patents. 'Cause hey -- it's "new technology", right?
Hey moron, did you read what he said?
. It'd be nice to read a calm, clear explanation of what it really does, and how it's different from earlier versions.
He wants a simple summary. If he wanted to wade through the whole thing, he would have done it already. Thanks for completely missing the point of his question.
By not providing such information, Google is leaving folks uncertain - now, honestly, if your data was good you'd release it because it would do good to your stock price. If you aren't, I'd be worried about what else is going on, and that is most definitely not a good sign.
Wow, I never thought I'd see someone who either knew this little, or this much game theory. You wouldn't necessarily release data that was good *for the precise reason that you just gave* -- it causes "no data" to signal "bad". By never giving any data, you provide no signals. If you just provide good data, any time you don't, it will be assumed the data is bad.
If you have "some solid leads, some strong hypotheses, but realize that another ten-year study is in order to truly rule out some key issues" then your study should claim that you "found an interesting correlation", not that you "found that coffee may reduce the risk of damage to health" (or whatever the case may be). That way you can convey information which may be useful to others while not overstating your results.
... that takes a scientist. Any any scientist should know how unreliable a self-reported tiny change in risk over 30 years is going to be.
(And while we're at it, why is all this raw data the exclusive domain of a few people? If all you did was become the first person who, upon searching a database, found a cool correlation, why isn't this same data open to others? Why should we have to wait until a busy scientist gets around to checking out a few correlations before we find out about things like this?)
You seem to have this impression that the job of a scientist is to gather data. It's not. That's the job description of a lab worker. The job of a scientist is to properly design experiments, to come up with innovative ways to rule out alternate hypotheses that a layman can come with in ten seconds, to state precisely the caveats behind the results. (A great example of this in the social sciences was an economist who investigated the impact of women having children on their future income. To control for the possibility that women who want to have children at age X also share income-enhancing or income worsening effects, she focused on a wide array of data from women who had miscarried, and therefore "were the type to have kids" but didn't.)
It's true as you say that science is an ever-growing process, but please -- when the conclusions flip every few years, that should be taken as a signal that your conclusions have been premature and both sides need to revise estimates about how much data they need before calling something a theory. It's also correct that humans need validation, but that's precisely why we should be suspicious of anything that suggest something we want to believe. Coffee is good for you? If it's too good to be true, make damn sure it is.
Any well-funded fool can gather data. Figuring out how to gather the right data
You think this is the media's fault? You don't think scientists overstate their pop-wisdon-verifying results, knowing they'll easily be misinterpreted, in order to get more attention?
That's an extreme stretch of the normal use of the term "technology". They thought of systematic way of warning people about phishing sites by compiling a list of them. Good for you. But computer programs, databases, and browsers have existed for a long time. This isn't a "new technology". It's a computer program. I know, you probably think it's a minor point, but keep in mind that Microsoft considers removing its own damn bugs to be "new technology" (NT).
Thinking up ways to warn people about phishing sites isn't "new technology".
When Google went public, they became obligated to the stockholders, regardless of any preexisting 'mantras'.
Actually, to borrow from Fight Club, you choose their level of involvement. If you issue stock with the explicit caveat that you're not going to listen to stockholder pressure for short-term gain, and you're going to limit their ability to control the company, that's exactly what stockholders get, and indeed deserve, for buying in. They have every right to set the terms of their issuance of stock. If you didn't want that kind of deal, you were under no obligation to buy in. No one was bait-and-switched. It's just that Google's operation style does fit in the old-boys' paradigm. (Sorry for using that word, I couldn't help it.)
In Bush's mind, someone trying to lift off a load of debt and gain some financial independence is a terrorist. ;-)
It's not impossible to show that an attribute (enviornmental or genetic) can affect relative risk for any phenotype (in this case, heart disease.) You need a strong understanding of statistics, a well designed study, and a large enough number of patients to see weaker effects.
True, but did they do that here? How often do they even do that? If they really controlled as well as you claim, you would not see anywhere near the number of reversals as there have been. Plus, whenever they give one of these studies, they always, always add "Oh, but there's always the possibility that this apparent health gain from coffee merely arises from other healthy activities that correlated with coffee drinking." But if you didn't find a very compelling reason to rule this out, your work never should have been published!
And how precisely would you isolate this effect over *thirty years*? You would have to come up with, by my rought estimates, 100,000 people for a remotely useful study that could control for all the other factors and phony correlations that the GP came up with after just a few minutes of thinking. Given the self-reporting they use, the multiple alternate correlations, the number of factors that enter over 30 years, the sample size needed for controls, and the small magnitude of the supposed effect (if you cut in half the risk of heart disease, that's still only 1 in 100,000 fewer cases), it really makes me suspicious of the supposed *confidence* they claim to have in their findings.
The problem is not that scientists change their minds. It's that the correlation was too flimsy to justify publishing in the first place. Every time they publish something like this, they always add "but it's still not clear if this correlation between [engaging in activity thought to be unhealthy] and better heart/longevity is because of the [unhealthy activity] or because people who do that tend to also [engage in separate, healthy activity]". Um, EXCUSE ME? If you didn't find a damn good reason to rule that out, your study should not have passed peer review. Of course you shouldn't wait until all doubt is removed before publishing -- but you should also at least try to look like you did your homework the first time around.