I have a tendency to write really long e-mails, too -- and often for many of the same reasons you mentioned.
Then, I rewrite them to include only the information I really need to send, and save everyone else the trouble of reading half to 3/4's of a page of information they really won't understand and couldn't care less about, even if they could understand it. And do you know what? I've found that doing so has improved communication with my coworkers immensely.
It doesn't matter how improbable it is -- the math does not say that it cannot happen, and therefore my point remains: even if the odds are 1 in 1000000000000000000000000000000000000000 that such an event would occur and that we would happen to look at the right place at the right time, it is still conceivable that we could have done so, whether we looked at one exoplanet or 1000000000000000000000000000000000000000 exoplanets. It may be highly unlikely to have found such an occurrence by exoplanet #370, but that is all the math says -- "wow, you REALLY got lucky this time."
Yes, if all you are concerned about is getting a five (or any other specific number). If you don't care what the first number is, only that it matches whatever you roll, then the chance of both events occurring together is 1 in 1000:
1) Chance of picking some number between 1 and 1000 from a pool of numbers between 1 and 1000: 1 (not 1 in 1000).
2) Chance of rolling some number that matches the number from step 1 on a 1000 sided die : 1 in 1000.
3) Chance of both occurrences happening together: 1 * 1 in 1000 = 1 in 1000.
I perhaps could have worded the original thought experiment better, though.
WiFi AP to a wired router running OpenVPN? It's my preferred method. It still may not count as "not jumping through hoops", but it's about the best I can think of right now.
Did you actually read and comprehend what I wrote? I have no quibble with the scientists findings, nor did I say that finding a planet with such an orbit is highly likely. Rather, I disagreed with the poster's statement that since we only have a 1 in 1000 chance of detecting such an occurrence, but we've only observed 370 exoplanets, something must be wrong because the math says we can't detect an event with a probability of 1 in 1000 after only 370 observations.
That is an invalid leap of logic, and suggests the poster does not understand statistics. I have no problem with the scientists or poster saying that such an event is extremely rare ("1 in 1000 chance that it would be observable now"). I have a problem with the poster taking that statement and expanding it into, and I quote, "We've found less than 1000 exoplanets, so the math doesn't work."
That is like saying the odds of flipping a coin and getting heads is 1 in 2, so I have to flip a coin twice to get heads. Well, the odds of getting tails is also one in two, so do you also have to flip a coin twice to get tails? In that case, what do you get on the first flip? It can't be heads, since we've decided that it takes twice to get heads. It can't be tails, since we've also decided that we have flip the coin twice to get tails. Assuming that a coin flip is either heads or tails (disregarding the extremely rare chance that the coin lands on edge and remains balanced there), a coin flip *has* to be either heads or tails, but we just proved it can't be either heads or tails on the first flip. Therefore, our initial assumption -- that given a 1 in x chance of some event happening, we have to observe x occurrences of that event before we can expect to see the event -- HAS TO BE FALSE, which was all I was trying to say above. An event may only have a 1 in 1000 chance of occurring, but it is entirely possible that you may observe that 1 in 1000 event the very first time.
I've got a motorcycle I am trying to rebuild (the engine was in six boxes when I acquired it...) and my wife is constantly pointing out the fact that it's been several months, and I haven't made a lot of progress (I have one head rebuilt and I've scrapped a lot of unusable parts). The rebuild was a lot harder than I expected and if I had even half a brain, I'd just go to eBay and buy a new engine, but I'm too proud to admit defeat and do the sensible thing, so it sits in the garage and I occasionally go tinker with it for a while. At this rate, when my great, great, great, great, great, great-grandchildren are old enough to ride, it might be finished.
On the other hand, we should be surprised if we catch a one-in-a-billion event after a mere 370 observations.
Does no one take statistics and survey sampling anymore? I'm not a statistician, but even I understand why that is wrong. We should not be surprised when we detect a one-in-a-billion event after only 370 observations. We should be surprised if, on average, we have detected a one-in-a-billion event far more frequently than expected (that is, we conducted one trillion observations and detected significantly more than 100 such occurrences), assuming our sample technique doesn't skew the probability of detecting the one-in-a-billion event.
If the odds of finding a Wasp-18b type planet is one-in-a-billion, what are the odds that the first planet we detect is a Wasp-18b type planet? One in a billion. What about the 370th planet? One in a billion, again. What about the one billionth planet? Yep, one in a billion. We don't have to detect one billion planets -- or even a statistically significant portion of one billion planets -- before we can expect to find such a planet. It could be the first, the 10th, the 1000th or the one billionth. But on average, for every billion planets we examine, we should only detect one Wasp-18b type planet. If we consistently start finding one in a million (or one in a thousand, or...) we should be surprised. But it is not statistically meaningful that the 370th planet we found was that one in a billion; our sample size is not yet large enough to say such a thing.
Furthermore, if our sampling (detecting) technique skews the odds in favor of finding Wasp-18b type planets, then maybe the odds of us detecting such a planet is only 1 in 1000 although the odds of such planets forming are only 1 in a billion.
"Now, my own suspicion is that the universe is not only queerer than we suppose, but queerer than we can suppose...I suspect that there are more things in heaven and earth than are dreamed of, or can be dreamed of, in any philosophy." --J.B.S. Haldane
It can be, I suppose...if you were clever and/or determined enough, you might be able to figure out a way to moderate via wget or "telnet slashdot.org 80" and enter the HTTP commands by hand (I don't have mod points right now, or I would try it).
...there's only a 1 in 1000 chance it would be observable now. We've found less than 1000 exoplanets, so the math doesn't work.
That's not how it works. Suppose I pick a number between 1 and 1000 at random. Then, suppose I roll a 1000 sided die. Odds are 1 in 1000 that those numbers will match, but that doesn't mean I can't roll the die and match those two numbers with less than 1000 throws. I might match the numbers on the very first throw! It just means I probably shouldn't bet my retirement on matching those numbers on any given throw.
Similarly, if the odds of discovering a planet such as Wasp18b are 1 in 1000, that doesn't mean that, "we've found less than 1000 planets, so we couldn't possibly have found such a planet yet." It just means that if I observe 1000 planets, most likely only one of them will be like Wasp18b. It could be the first one I observe, the 99th, the 1000th or any one in between.
CrimsonAvenger also raised a similar objection, pointing out that guns are used far more often for target practice than for killing, which might also be true (historically, including hunting for food? Maybe; I don't know for sure). It's certainly true for me, at least. I own several guns but I have never shot a single living thing with any of them (although I have shot a grouse -- once -- with a bow, but that's slightly off-topic).
It never hurts to flesh out and clarify an argument. That's one of the reasons I love/. -- I get critical analysis of my thinking, which I greatly enjoy (well,usually:). "Steel sharpens steel," and all that. Thanks for showing me where I can do better!
Yeah, I didn't really want to go there. Although I think the principle is as true for guns and software as it is for cars, a lot of people feel that guns are only used for killing, therefore they are inherently evil[1]. Consequently, if I had used guns rather cars in my analogy, I would have potentially harmed my argument.
[1] I believe that this conclusion is false, too. A gun is designed to kill, but I disagree that this is always evil. I would not hesitate for a single second to kill someone who intended to harm my family, nor do I think that using a gun to kill a moose or deer, etc. for food to feed my family is evil. However, there are people who believe that killing is never, ever justified, so I didn't intend to go there in my original argument.
You missed my point. I'm not arguing whether or not this particular piece of software is good or evil; I'm arguing whether or not someone releasing under the GPL a piece of software that is most likely to be used for ill intent makes the GPL itself good or evil.
The argument I am trying to counter goes like this:
1) This software is evil.
2) This software was released under the GPL.
3) Therefore, the GPL is evil.
This is the argument I was attacking, and it is indeed a logical fallacy. The GPL does not take on the characteristics of the software released under it; software released under the GPL takes on the characteristics of the GPL, and those characteristics apply equally to malware and to beneficial software.
In my argument, the GPL, not the software, is like the car. Both the GPL and a car can be used either for legitimate purposes (i.e., Linux/GPL and transportation/car) or illegitimately (i.e., malware/GPL and murder/car).
Incidentally, I hereby release this argument under the GPL so that anyone wishing to use it to counter FUD from big corporations trying to make the association from malware to GPL may be free to do so:)
Your signature, in light of your post, is rather interesting. And I won't even go into the differences between a pure democracy and a representative democracy right now.
I guess I should RTFA -- there are posts below mine that show that he does, in fact, own the copyright to the software. In which case, if the company hires someone to write software and the author of that software then posts it under the GPL (or other FOSS license), then how does that possibly add ammo to the FUD argument about the GPL?
GPL really is a stupid option in my opinion...it will give some more ammo to the FUD that carries some big corporations that GPL is bad.
Assuming the source code is his to give away (certainly not a given!), I have to disagree.
1) GPL is perfect for this, since it essentially says, look -- take this code and modify it, redistribute it, analyze it, re-publish it...do what you want with it, as long as you allow this same freedom to anyone else who gets the software. This is the whole reason the GPL exists in the first place! In this case, this is good because it allows others to take the code apart, figure out what makes it tick and come up with A/V signatures to detect it without worrying about whether or not you are violating a licensing agreement by trying to analyze and reverse engineer the code. It does also allow black hats to rewrite and enhance it for illicit use, but that's one of the problems with freedom -- you can always abuse freedom, if you choose. And for whatever it's worth, I don't think the black hats were going to be too concerned about license restrictions, anyway...
2) Saying that GPL is bad because software that may possibly be used for ill intent is licensed under the GPL is a logical fallacy. Would anyone in their right mind say that, because someone somewhere has used a car to commit a crime (drunk driving? getaway car in a robbery? ran over someone who pissed them off?) that therefore all cars are inherently evil? Of course not, so why would you say that about software?
3) Okay, maybe that's not what you meant by your "more ammo to FUD" argument. Maybe instead you meant that it allows big corporations to worry that their developers might give away their software products by licensing them under the GPL. How is that any different than any other commercially developed GPL'd product (MySQL, RHEL, etc.)? Or, from another angle, how is that any different than any other big company worrying that their developers might give their intellectual property to a competitor, or publish it on-line somewhere? It is *possible* for this to happen whether it's GPL'd, released under other FOSS licenses or simply posted on-line without any kind of license at all.
Of course, if he doesn't really own the rights to the source code, then all bets are off.
Slashdot Mirror
I have a tendency to write really long e-mails, too -- and often for many of the same reasons you mentioned.
Then, I rewrite them to include only the information I really need to send, and save everyone else the trouble of reading half to 3/4's of a page of information they really won't understand and couldn't care less about, even if they could understand it. And do you know what? I've found that doing so has improved communication with my coworkers immensely.
You, sir, are more geekier than I :)
:(
Not only did you beat me to the punch, but you one-upped me as well. I'm ashamed I didn't think of that myself
What, your browser doesn't start searching as soon as you hit the "/" key? That's so....Windows! :D
Sorry -- "Your search - "1000 sided" die - did not match any products."
;-D
[:rolleyes:]
It doesn't matter how improbable it is -- the math does not say that it cannot happen, and therefore my point remains: even if the odds are 1 in 1000000000000000000000000000000000000000 that such an event would occur and that we would happen to look at the right place at the right time, it is still conceivable that we could have done so, whether we looked at one exoplanet or 1000000000000000000000000000000000000000 exoplanets. It may be highly unlikely to have found such an occurrence by exoplanet #370, but that is all the math says -- "wow, you REALLY got lucky this time."
Yes, if all you are concerned about is getting a five (or any other specific number). If you don't care what the first number is, only that it matches whatever you roll, then the chance of both events occurring together is 1 in 1000:
1) Chance of picking some number between 1 and 1000 from a pool of numbers between 1 and 1000: 1 (not 1 in 1000).
2) Chance of rolling some number that matches the number from step 1 on a 1000 sided die : 1 in 1000.
3) Chance of both occurrences happening together: 1 * 1 in 1000 = 1 in 1000.
I perhaps could have worded the original thought experiment better, though.
WiFi AP to a wired router running OpenVPN? It's my preferred method. It still may not count as "not jumping through hoops", but it's about the best I can think of right now.
Exactly.
Did you actually read and comprehend what I wrote? I have no quibble with the scientists findings, nor did I say that finding a planet with such an orbit is highly likely. Rather, I disagreed with the poster's statement that since we only have a 1 in 1000 chance of detecting such an occurrence, but we've only observed 370 exoplanets, something must be wrong because the math says we can't detect an event with a probability of 1 in 1000 after only 370 observations.
That is an invalid leap of logic, and suggests the poster does not understand statistics. I have no problem with the scientists or poster saying that such an event is extremely rare ("1 in 1000 chance that it would be observable now"). I have a problem with the poster taking that statement and expanding it into, and I quote, "We've found less than 1000 exoplanets, so the math doesn't work."
That is like saying the odds of flipping a coin and getting heads is 1 in 2, so I have to flip a coin twice to get heads. Well, the odds of getting tails is also one in two, so do you also have to flip a coin twice to get tails? In that case, what do you get on the first flip? It can't be heads, since we've decided that it takes twice to get heads. It can't be tails, since we've also decided that we have flip the coin twice to get tails. Assuming that a coin flip is either heads or tails (disregarding the extremely rare chance that the coin lands on edge and remains balanced there), a coin flip *has* to be either heads or tails, but we just proved it can't be either heads or tails on the first flip. Therefore, our initial assumption -- that given a 1 in x chance of some event happening, we have to observe x occurrences of that event before we can expect to see the event -- HAS TO BE FALSE, which was all I was trying to say above. An event may only have a 1 in 1000 chance of occurring, but it is entirely possible that you may observe that 1 in 1000 event the very first time.
I feel your pain :)
I've got a motorcycle I am trying to rebuild (the engine was in six boxes when I acquired it...) and my wife is constantly pointing out the fact that it's been several months, and I haven't made a lot of progress (I have one head rebuilt and I've scrapped a lot of unusable parts). The rebuild was a lot harder than I expected and if I had even half a brain, I'd just go to eBay and buy a new engine, but I'm too proud to admit defeat and do the sensible thing, so it sits in the garage and I occasionally go tinker with it for a while. At this rate, when my great, great, great, great, great, great-grandchildren are old enough to ride, it might be finished.
On the other hand, we should be surprised if we catch a one-in-a-billion event after a mere 370 observations.
Does no one take statistics and survey sampling anymore? I'm not a statistician, but even I understand why that is wrong. We should not be surprised when we detect a one-in-a-billion event after only 370 observations. We should be surprised if, on average, we have detected a one-in-a-billion event far more frequently than expected (that is, we conducted one trillion observations and detected significantly more than 100 such occurrences), assuming our sample technique doesn't skew the probability of detecting the one-in-a-billion event.
...) we should be surprised. But it is not statistically meaningful that the 370th planet we found was that one in a billion; our sample size is not yet large enough to say such a thing.
If the odds of finding a Wasp-18b type planet is one-in-a-billion, what are the odds that the first planet we detect is a Wasp-18b type planet? One in a billion. What about the 370th planet? One in a billion, again. What about the one billionth planet? Yep, one in a billion. We don't have to detect one billion planets -- or even a statistically significant portion of one billion planets -- before we can expect to find such a planet. It could be the first, the 10th, the 1000th or the one billionth. But on average, for every billion planets we examine, we should only detect one Wasp-18b type planet. If we consistently start finding one in a million (or one in a thousand, or
Furthermore, if our sampling (detecting) technique skews the odds in favor of finding Wasp-18b type planets, then maybe the odds of us detecting such a planet is only 1 in 1000 although the odds of such planets forming are only 1 in a billion.
"Now, my own suspicion is that the universe is not only queerer than we suppose, but queerer than we can suppose...I suspect that there are more things in heaven and earth than are dreamed of, or can be dreamed of, in any philosophy." --J.B.S. Haldane
Marvin? Is that you?
Don't forget to eat a lot of peanuts first...you will need the salt.
/. of all places should have moderators who recognize PP's reference!
On a side note, "Score:0"?!?!?!
"All these other planets are yours...Attempt no landings there"?
It can be, I suppose...if you were clever and/or determined enough, you might be able to figure out a way to moderate via wget or "telnet slashdot.org 80" and enter the HTTP commands by hand (I don't have mod points right now, or I would try it).
...there's only a 1 in 1000 chance it would be observable now. We've found less than 1000 exoplanets, so the math doesn't work.
That's not how it works. Suppose I pick a number between 1 and 1000 at random. Then, suppose I roll a 1000 sided die. Odds are 1 in 1000 that those numbers will match, but that doesn't mean I can't roll the die and match those two numbers with less than 1000 throws. I might match the numbers on the very first throw! It just means I probably shouldn't bet my retirement on matching those numbers on any given throw.
Similarly, if the odds of discovering a planet such as Wasp18b are 1 in 1000, that doesn't mean that, "we've found less than 1000 planets, so we couldn't possibly have found such a planet yet." It just means that if I observe 1000 planets, most likely only one of them will be like Wasp18b. It could be the first one I observe, the 99th, the 1000th or any one in between.
Point taken.
CrimsonAvenger also raised a similar objection, pointing out that guns are used far more often for target practice than for killing, which might also be true (historically, including hunting for food? Maybe; I don't know for sure). It's certainly true for me, at least. I own several guns but I have never shot a single living thing with any of them (although I have shot a grouse -- once -- with a bow, but that's slightly off-topic).
It never hurts to flesh out and clarify an argument. That's one of the reasons I love /. -- I get critical analysis of my thinking, which I greatly enjoy (well,usually :). "Steel sharpens steel," and all that. Thanks for showing me where I can do better!
Yeah, I didn't really want to go there. Although I think the principle is as true for guns and software as it is for cars, a lot of people feel that guns are only used for killing, therefore they are inherently evil[1]. Consequently, if I had used guns rather cars in my analogy, I would have potentially harmed my argument.
[1] I believe that this conclusion is false, too. A gun is designed to kill, but I disagree that this is always evil. I would not hesitate for a single second to kill someone who intended to harm my family, nor do I think that using a gun to kill a moose or deer, etc. for food to feed my family is evil. However, there are people who believe that killing is never, ever justified, so I didn't intend to go there in my original argument.
You missed my point. I'm not arguing whether or not this particular piece of software is good or evil; I'm arguing whether or not someone releasing under the GPL a piece of software that is most likely to be used for ill intent makes the GPL itself good or evil.
:)
The argument I am trying to counter goes like this:
1) This software is evil.
2) This software was released under the GPL.
3) Therefore, the GPL is evil.
This is the argument I was attacking, and it is indeed a logical fallacy. The GPL does not take on the characteristics of the software released under it; software released under the GPL takes on the characteristics of the GPL, and those characteristics apply equally to malware and to beneficial software.
In my argument, the GPL, not the software, is like the car. Both the GPL and a car can be used either for legitimate purposes (i.e., Linux/GPL and transportation/car) or illegitimately (i.e., malware/GPL and murder/car).
Incidentally, I hereby release this argument under the GPL so that anyone wishing to use it to counter FUD from big corporations trying to make the association from malware to GPL may be free to do so
Your signature, in light of your post, is rather interesting. And I won't even go into the differences between a pure democracy and a representative democracy right now.
Well...it makes it easier for the A/V companies to write a detection algorithm, since they don't have to reverse engineer the binary now.
I guess I should RTFA -- there are posts below mine that show that he does, in fact, own the copyright to the software. In which case, if the company hires someone to write software and the author of that software then posts it under the GPL (or other FOSS license), then how does that possibly add ammo to the FUD argument about the GPL?
GPL really is a stupid option in my opinion...it will give some more ammo to the FUD that carries some big corporations that GPL is bad.
Assuming the source code is his to give away (certainly not a given!), I have to disagree.
1) GPL is perfect for this, since it essentially says, look -- take this code and modify it, redistribute it, analyze it, re-publish it...do what you want with it, as long as you allow this same freedom to anyone else who gets the software. This is the whole reason the GPL exists in the first place! In this case, this is good because it allows others to take the code apart, figure out what makes it tick and come up with A/V signatures to detect it without worrying about whether or not you are violating a licensing agreement by trying to analyze and reverse engineer the code. It does also allow black hats to rewrite and enhance it for illicit use, but that's one of the problems with freedom -- you can always abuse freedom, if you choose. And for whatever it's worth, I don't think the black hats were going to be too concerned about license restrictions, anyway...
2) Saying that GPL is bad because software that may possibly be used for ill intent is licensed under the GPL is a logical fallacy. Would anyone in their right mind say that, because someone somewhere has used a car to commit a crime (drunk driving? getaway car in a robbery? ran over someone who pissed them off?) that therefore all cars are inherently evil? Of course not, so why would you say that about software?
3) Okay, maybe that's not what you meant by your "more ammo to FUD" argument. Maybe instead you meant that it allows big corporations to worry that their developers might give away their software products by licensing them under the GPL. How is that any different than any other commercially developed GPL'd product (MySQL, RHEL, etc.)? Or, from another angle, how is that any different than any other big company worrying that their developers might give their intellectual property to a competitor, or publish it on-line somewhere? It is *possible* for this to happen whether it's GPL'd, released under other FOSS licenses or simply posted on-line without any kind of license at all.
Of course, if he doesn't really own the rights to the source code, then all bets are off.