There are a fair number of posts here that say something like:
This will always be a problem because people are just stupid.
At this point don't you think the "You are an idiot, I'm going to educate you," "awareness raising" security efforts by IT (and HR) people have basically failed? An irritatingly intrusive security approach combined with condescension to the users -- that should work, right? So let's force them to change passwords every month, but then chide them about writing down their passwords anywhere. Good idea. Makes things less secure, but as long as they're more secure in theory...
I'm not sure that makes sense. It's easy to be down on those "arrogant IT people", but what evidence do you have that arrogant IT people are hurting security? Maybe being pushy and mean actually does help people remember to do an unpleasant job, namely paying attention to security. For that matter, what evidence do you have that the basic point about people being stupid is wrong? Without saying it explicitly, you are arguing that passwords could be secure if people were just trained better. I disagree.
See, I think the conclusion which should be drawn from the whole "people are stupid" thing is simply that passwords aren't a security system which works very well. I wouldn't use the word "stupid" to describe people, but I would argue that people will make mistakes. I think that sooner or later someone is bound to use a too-simple password or accidentally give out their password no matter how well they've been trained. Furthermore, the types of passwords which are secure are hard for humans to remember, which will lead to the use of post-it notes wherever you have passwords (a _major_ problem with passwords). Unlike some other security methods, dictionary-type attacks tend to work on passwords (even good ones are easier to brute-force that a 2048 bit DSA key, or a retinal map:). And of course there will probably always be issues with weak/null default passwords (because programmers "screw up" too...)
Really, this whole "people are stupid" thing is just an argument for some other security method. I've argued that passwords just aren't a terribly secure security method. They're a _convenient_ security method. They're cheap and easy to implement, and can be layered on top of other methods (e.g. physical tokens -- think "ATM card"). That's why they're so commonplace. But at the level of theory, they have many flaws, the main one being that they force everyone in an organization to become at least a small part of the security team. Compare this to, say, fingerprint identification, which require only that members of the organization keep their fingers intact.:)
...by simply ignoring the horrific examples of abusive governments throughout history. How can a responsible thinker so easily shrug off the need to protect oneself from the unknown abuses of the future just because one may think things are relatively agreeable at present?"
This is a perfect example of why gun ownership is a good idea and why our forefathers thought the Second Ammendment was a good idea.
Ummm... I suspect that a "responsible" thinker would pick the appropriate weapon, rather than recommending violence as the solution for every problem. Consider this: which of these items will be more useful in helping you prevent violations of your personal privacy:
When I do on-site work, I often have to ask people their passwords. I can't think of one time when anybody refused to tell me.
They probably figure that they're supposed to. I mean, really, does the average office worker want to tell the boss that their $150/hour consultant had to stay extra time while you figured out whether or not you were allowed to give them the information they requested? Sure, they're still making a mistake, but at least there's a reason for it.
I wonder if something similar happened in this UK study. It sounds like everyone in the company participated (up to and including the CEO). I wonder how this interview worked exactly. Did employees get a memo from the CEO saying that they would be interviewed by some mysterious individuals about some unknown topic (you can't _say_ that it's about security, or the workers' behavior will change completely) and that they were expected to participate? Think "Office Space" here, people. Who wouldn't want to cooperate under those circumstances?
Well, considering 1GB CompactFlash cards are around $200, $100 for 1.5GB is pretty cheap for ultra-small storage.
Edit for accuracy:
Well, considering 1GB CompactFlash cards are presently around $200, $100 for 1.5GB will be pretty cheap for ultra-small storage in Fall, which is when they're expected to be available.
Now, I'm not sure that I agree with this statement.
Why does the USPS need to get it's act together? you cite that our already privatzed postal service is the envy of the world, but why say it needs to get its act together? They are efficient, statistically reliable (anecdotes about US mail getting lost are mere, well, anecdotes) and very cheap.
Try sending 2 oz letter 3500 miles for $0.36 in any other country in the world.
I assume that joke here is that there are hardly any other countries in the world which have 3500 miles for a letter to go. Even if the mail in, say, Japan were free, you still couldn't send a letter 3500 miles. I guess nobody got it.:)
"I would be a lot less anti-Microsoft if they actually put forth any effort at all to be compatible and/or interoperate with other OSes. I too am sick to death of the, "if you want to do this you have to run Windows" crap."
I hear ya. It pisses me off I can't play Dreamcast games on my Playstation 2.
Media files are OS-independent, much like TVs are input-device-independent (if they follow the standards, everything works fine). Thus a more appropriate analogy would be that you can't hook up your Dreamcast to a regular TV. Instead it turns out that Dreamcase games will only display on a special kind of TV...
thus, Searle's example of a Chinese room built on a stimulus-response system would be impossible to construct, making his thought experiment incoherent.
Although I basically agree with you, let's remember that the original thrust of the Chinese Room argument was to prove that _programs_ could not emulate humans (the Turing Test being just one example of what "emulate" meant). I seem to recall that the Chinese Room was supposed to have computational power equivalent to a Turing Machine. If it does, then the whole Chomsky argument actually works in Searle's favor. It says that no computer program, at least one running on a serial computer, could possibly emulate human language perfectly, because it could achieve no better than Stimulus-Reponse behavior. I vaguely recall there being some issues about where the memory (the equivalent of a Turning Machine's tape) is in the Chinese Room, so I'm not sure how this argument turns out.
That's part of my problem with Searle's Chinese Room thought experiment. He's saying that an automaton responding to Chinese following rules would not "understand" Chinese in the way a human who speaks the language would. But this is presupposing that the way a human who "understands" Chinese does so is not through just a very long list of rules coded in neurons, which I consider to be a rather controversial assumption.
Actually, I would suggest to you that the situation is exactly the opposite. Searle's Chinese Room encodes Chinese as a very long list of rules. This system is described in a manner identical to a behavorist-style stimulus-response system. Stimulus X is sent in, response Y is generated (I think behaviourists allowed for complex S-R chains, but ultimately both systems just follow a list of rules). Didn't Chomsky make his name by proving (in his book "Syntactic Structures") that language had to be encoded in a manner which is more complex than a stimulus-response system? Does that not imply that the Chinese Room's processing of Chinese is by definition deficient compared to a human's?
Note that this does not, by itself, rule out the possibility that the Chinese Room could pass the Turing Test, since that only requires fooling a human, not being mathematically equivalent to one. My point is that humans are almost certainly more complex than the Chinese Room or any other "list of rules" based system. I'm pretty sure Chomsky proved that.
As a side note, any psychologist who studies language can give you even better reasons why language is not just a list of rules (e.g. fuzzy categories).
You write: "On the other hand, if we had a really parallel computer in which a bit could be written to twice (or more times) simultaneously, could the results be simulated on a serial machine?"
But this isn't the technical definition of a parallel computer (pick up a textbook on the theory of computation that describes Turing Machines, DFAs etc. Try the one by Sipser).
Sure, why not.:)
Any computation that can be performed by a parallel computer can be performed by a serial one.
I was hoping to get some clarification on this, not just a restatement, but okay. Maybe I'll have to read that book.
What you describe sounds more like a quantum computer, which is a qualitatively different type of beast. (although I'm not sure I understand precisely what you mean by "written to twice simultaneously).
Could be.
The reason why I this "written to twice" computer is that the term "parallel computer" seems to be often used to refer to systems which are equivalent to multiple serial computers running at the same time, like a cluster of x86 computers. Sure, I agree that this can be emulated by a serial computer. This type of parallel computer operates in discrete steps, so presumably it can be emulated by running through all those steps on a serial computer.
But what about a system where data in the system can be written to twice simultaneously? In other words, one where multiple read/write heads are changing the symbol on the same spot on the tape at the same time. This may sound like a weird beast. But isn't a neuron an example of such a thing?
Certainly a neuron is performing calculations. If the voltage differential across a calcium channel hits a certain threshold, the channel will open (a "bit" of data": 1=open, 0=closed). If it opens in the right place and at the right time, the neuron will fire. Neurons are floating in a sea of ions, each with its own electric field, and controlling the flow of these ions is how neurons influence one another. The neuron "sums" voltage differences at several points and determines whether they exceed certain thresholds. Is this not a case of multiple read/write heads altering the same bit of tape at the same time?
Now, can a serial computer emulate multiple read/write heads altering the tape at the same time? I think this is the same question as "How would you emulate this neuron?" So how do you emulate this neuron on a serial computer? You can certainly add up the electric fields of all of the ions at any given time and determine the state of the neuron at one point in time. Serial computers can do that just fine. But to properly emulate the neuron, you need to do this with infinite temporal resolution. The _exact_ instant when that calcium channel opens could affect the results of the neuron's calculations (e.g. whether or not one neuron causes the next one to fire could potentially depend on infinitely small time differences). So perfect precision is required. I don't think a serial computer can do that.
A Parallel computer can't actually do anything that a serial computer can't do, other then doing things more efficiently.
I'd like to know what you mean exactly. I think there may be some limitations to that assertion.
I agree that an SMP computer or a Beowulf cluster can't do anything that a serial computer can do. The main reason being that in some ways (e.g. memory access) these devices are somewhat serial. Any given bit of data can only be accessed (e.g. written to) by one process at a time.
On the other hand, if we had a really parallel computer in which a bit could be written to twice (or more times) simultaneously, could the results be simulated on a serial machine?
In a related vein, it is taken for granted in certain fields that the brain is a massively parallel computer, and that it could never be emulated on a serial computer. This platitude is not just based on the Chinese Room argument either (which I think is probably subtly flawed). Rather, it has to do with the complexity of the "processors" (neurons) in the brain.
Basically, neurons are acted upon my many events simultaneously, and those events are analog (e.g. multiple ion channels are letting several types of ions flow in and out of the neuron at variable rates and in various locations). Basically, you would need to update a model of the state of this neuron at infinitely small intervals. Unless time is quantized, it can't be done.
The *nix community needs to get some serious developer support before companies will really start to look at it seriously. Also the whole RTFM attitude is doing tons of harm to the movement as well. When the only support you can get for an OS is found on the web, and half the responses are along the lines of "RTFM 1d10t, y0ur a 1user, and 1m 37337" this does not instill confidence in that OS.
Now now, you're comparing apples and oranges here.
If you go direct to a Linux vendor like RedHat or a company like The Kompany for product support, you'll get a reasonable response. If you go to usenet for support, you may well get the sort of garbage you were talking about.:) I suspect this is true of any software product or any OS (are the Microsoft/Mac/Whatever groups so much more polite and helpful?)
In general, if you want commercial-level support, go with a commercial product. That should be obvious. Buy StarOffice instead of using OpenOffice. The question, then, is whether or not enough *nix software has that level of support, and how many users really demand that level of support. That is a complicated question. I suspect that the correct answer is "It depends on the particular app you're talking about".
Of course, it never hurts to have the _option_ to look at the source code, go to Usenet for support, etc. as well.
This only furthers the evidence that global warming has not been caused by human action.
Actually, it doesn't, due to the near total lack of any evidence in the article. It is at most a vague summary of what was concluded.
More to the point, this is just a summary. The study could have massive flaws, which is why we should wait until scientists have had time to critique it before drawing any conclusions, especially given that there are conflicting studies (at least one is mentioned at the start of the article) and reasons to disagree with the article's conclusions (see Dr. Brown's comments at the end of the article). This peer review is important. Sometimes evidence, such as certain well-known satellite results, can turn out to be wrong.
The expectation that Linux will fulfill the hardware driver installation off the distribution CD, when you admit that you may have to replace the entire hardware for XP, is inherently unfair and beyond what can be reasonably expected in any operating system. XP is not without it's major problems when it comes to older hardware ( especially scanners ) support and driver conflict problems.
Personally, I think you've got the right idea, but I'd put it a bit differently.
This writer left out one important thing: How long did it take to set up all that hardware on Windows 95? She seems to expect Linux to have the ease of install of a Windows XP, and when it Linux doesn't install that well on her hardware, she blames Linux. Yet it's not clear than any OS could install that easily. XP is one thing, but how much effort did it take to get multiple SCSI scaners and multiple parallel-port devices working on Win95? Did it take some work?
For a proper comparison, this review should have thoroughly discusses Linux vs. Win95, and probably vs. XP as well. Here's what we got instead:
* WinXP supposedly wouldn't work, though it wasn't tested.
* Linux (various distros) was a major pain.
* Win95 works, but we don't know if it was a pain or not.
Unfortunately, that's not a very informative comparison. It's comparing seven years of experience tweaking Win95 to maybe 4-6 hours per distro with Linux to 0 hours with XP or another modern MS operating system.
The poll I saw showed that ~65% supported the war on Iraq, but ~20% believed that all the Sept. 11 hijackers were Iraqi
I've seen a poll which claimed that 53% of the population believes that Hussein was personally responsible for the terrorist attacks on 9/11. A recent Tom Tomorrow cartoon claimed that it was 45%.
So yes, you're right. There are a lot of clueless types (or at least people who have a lot of facts wrong) supporting the war.
Well, no...I guess the picture you get depends upon the media you get your news from, but a couple of weeks before kick off when opposition to the war was at its strongest here in the UK, it was - disappointingly - very much under-reported even on domestic BBC news broadcasts. CNN etc. seemed to be showing even more anti war activity in the US so assuming the inevitable biased reporting I'd have to guess that opposition to the war in the US was at the least very significant. At the same time also, the increasing pro-war "support" seemed to consist, for most people, of a kind of resigned and grudging acceptance (pretty much as it has been in the UK).
Although there has been lots of anti-war protesting in the U.S., the general public is for the war. The last poll I saw showed 68% of the public supporting the war. That's a pretty solid majority.
The Second Superpower has no teeth. As witness the way support for the anti-war movement just melted away in the last days leading up to the war - right about the time it became really clear that Bush didn't give a flying fuck about public opinion. The vast majority of people just gave up. I don't wish to be hypocritical; I count myself among them.
I assume you mean _International_ Public Opinion (from George Bush's point of view, that is). With as popular as this war is in the U.S., and with Bush's otherwise lackluster record (what else did he do these last two years except wars and tax cuts?), I would recommend viewing this war as a lead in to the next election season. Remember, the Primaries are only 9 months away, and the campaigning will doubtless start well before that...
There's a small joke that goes around in the academic world
Biologists like to think they are chemists. Chemists like to think they are physicists. Physicists like to think they are mathematicians. And mathematicians like to think they are god.
FYI, there's another version of that joke which ends on "philosophers".
Slashdot Mirror
There are a fair number of posts here that say something like:
:). And of course there will probably always be issues with weak/null default passwords (because programmers "screw up" too...)
:)
This will always be a problem because people are just stupid.
At this point don't you think the "You are an idiot, I'm going to educate you," "awareness raising" security efforts by IT (and HR) people have basically failed? An irritatingly intrusive security approach combined with condescension to the users -- that should work, right? So let's force them to change passwords every month, but then chide them about writing down their passwords anywhere. Good idea. Makes things less secure, but as long as they're more secure in theory...
I'm not sure that makes sense. It's easy to be down on those "arrogant IT people", but what evidence do you have that arrogant IT people are hurting security? Maybe being pushy and mean actually does help people remember to do an unpleasant job, namely paying attention to security. For that matter, what evidence do you have that the basic point about people being stupid is wrong? Without saying it explicitly, you are arguing that passwords could be secure if people were just trained better. I disagree.
See, I think the conclusion which should be drawn from the whole "people are stupid" thing is simply that passwords aren't a security system which works very well. I wouldn't use the word "stupid" to describe people, but I would argue that people will make mistakes. I think that sooner or later someone is bound to use a too-simple password or accidentally give out their password no matter how well they've been trained. Furthermore, the types of passwords which are secure are hard for humans to remember, which will lead to the use of post-it notes wherever you have passwords (a _major_ problem with passwords). Unlike some other security methods, dictionary-type attacks tend to work on passwords (even good ones are easier to brute-force that a 2048 bit DSA key, or a retinal map
Really, this whole "people are stupid" thing is just an argument for some other security method. I've argued that passwords just aren't a terribly secure security method. They're a _convenient_ security method. They're cheap and easy to implement, and can be layered on top of other methods (e.g. physical tokens -- think "ATM card"). That's why they're so commonplace. But at the level of theory, they have many flaws, the main one being that they force everyone in an organization to become at least a small part of the security team. Compare this to, say, fingerprint identification, which require only that members of the organization keep their fingers intact.
...by simply ignoring the horrific examples of abusive governments throughout history. How can a responsible thinker so easily shrug off the need to protect oneself from the unknown abuses of the future just because one may think things are relatively agreeable at present?"
This is a perfect example of why gun ownership is a good idea and why our forefathers thought the Second Ammendment was a good idea.
Ummm... I suspect that a "responsible" thinker would pick the appropriate weapon, rather than recommending violence as the solution for every problem. Consider this: which of these items will be more useful in helping you prevent violations of your personal privacy:
1) A gun
2) An ACLU membership card
When I do on-site work, I often have to ask people their passwords. I can't think of one time when anybody refused to tell me.
They probably figure that they're supposed to. I mean, really, does the average office worker want to tell the boss that their $150/hour consultant had to stay extra time while you figured out whether or not you were allowed to give them the information they requested? Sure, they're still making a mistake, but at least there's a reason for it.
I wonder if something similar happened in this UK study. It sounds like everyone in the company participated (up to and including the CEO). I wonder how this interview worked exactly. Did employees get a memo from the CEO saying that they would be interviewed by some mysterious individuals about some unknown topic (you can't _say_ that it's about security, or the workers' behavior will change completely) and that they were expected to participate? Think "Office Space" here, people. Who wouldn't want to cooperate under those circumstances?
The process of providing a service to individuals (or businesses) who's end result is not conveyed for 4 years (1460 Days).
:P
Boy, this guy's a genius. 4 years = 1460 days? I assume that most of us know that there's one leap year (almost) every 4 years.
Videogames are meant to be an escape, not a acurate depiction of life, that is boring.
Which is what I said about "The Sims", but nobody seems to believe me...
Well, considering 1GB CompactFlash cards are around $200, $100 for 1.5GB is pretty cheap for ultra-small storage.
Edit for accuracy:
Well, considering 1GB CompactFlash cards are presently around $200, $100 for 1.5GB will be pretty cheap for ultra-small storage in Fall, which is when they're expected to be available.
Now, I'm not sure that I agree with this statement.
Why does the USPS need to get it's act together? you cite that our already privatzed postal service is the envy of the world, but why say it needs to get its act together? They are efficient, statistically reliable (anecdotes about US mail getting lost are mere, well, anecdotes) and very cheap.
:)
Because it's losing money. They lost quite a bit last year.
Also, if you've ever waited in line at a post office, you know that some aspects of their service are not exactly the model of efficiency.
Try sending 2 oz letter 3500 miles for $0.36 in any other country in the world.
:)
I assume that joke here is that there are hardly any other countries in the world which have 3500 miles for a letter to go. Even if the mail in, say, Japan were free, you still couldn't send a letter 3500 miles. I guess nobody got it.
"I would be a lot less anti-Microsoft if they actually put forth any effort at all to be compatible and/or interoperate with other OSes. I too am sick to death of the, "if you want to do this you have to run Windows" crap."
I hear ya. It pisses me off I can't play Dreamcast games on my Playstation 2.
Media files are OS-independent, much like TVs are input-device-independent (if they follow the standards, everything works fine). Thus a more appropriate analogy would be that you can't hook up your Dreamcast to a regular TV. Instead it turns out that Dreamcase games will only display on a special kind of TV...
thus, Searle's example of a Chinese room built on a stimulus-response system would be impossible to construct, making his thought experiment incoherent.
Although I basically agree with you, let's remember that the original thrust of the Chinese Room argument was to prove that _programs_ could not emulate humans (the Turing Test being just one example of what "emulate" meant). I seem to recall that the Chinese Room was supposed to have computational power equivalent to a Turing Machine. If it does, then the whole Chomsky argument actually works in Searle's favor. It says that no computer program, at least one running on a serial computer, could possibly emulate human language perfectly, because it could achieve no better than Stimulus-Reponse behavior. I vaguely recall there being some issues about where the memory (the equivalent of a Turning Machine's tape) is in the Chinese Room, so I'm not sure how this argument turns out.
That's part of my problem with Searle's Chinese Room thought experiment. He's saying that an automaton responding to Chinese following rules would not "understand" Chinese in the way a human who speaks the language would. But this is presupposing that the way a human who "understands" Chinese does so is not through just a very long list of rules coded in neurons, which I consider to be a rather controversial assumption.
Actually, I would suggest to you that the situation is exactly the opposite. Searle's Chinese Room encodes Chinese as a very long list of rules. This system is described in a manner identical to a behavorist-style stimulus-response system. Stimulus X is sent in, response Y is generated (I think behaviourists allowed for complex S-R chains, but ultimately both systems just follow a list of rules). Didn't Chomsky make his name by proving (in his book "Syntactic Structures") that language had to be encoded in a manner which is more complex than a stimulus-response system? Does that not imply that the Chinese Room's processing of Chinese is by definition deficient compared to a human's?
Note that this does not, by itself, rule out the possibility that the Chinese Room could pass the Turing Test, since that only requires fooling a human, not being mathematically equivalent to one. My point is that humans are almost certainly more complex than the Chinese Room or any other "list of rules" based system. I'm pretty sure Chomsky proved that.
As a side note, any psychologist who studies language can give you even better reasons why language is not just a list of rules (e.g. fuzzy categories).
why not establish mirrors before you submit to slashdot, especially if it's a self referral?
Maybe he should have "first post!"ed the text of his site?
Someone modded "Offtopic" making a joke about a Samba developer posting to /. in a Samba article? Who moderated that one?
:)
Remember, this is SlashDot. The moderator probably just read the subject line before making up their mind.
You write: "On the other hand, if we had a really parallel computer in which a bit could be written to twice (or more times) simultaneously, could the results be simulated on a serial machine?"
:)
But this isn't the technical definition of a parallel computer (pick up a textbook on the theory of computation that describes Turing Machines, DFAs etc. Try the one by Sipser).
Sure, why not.
Any computation that can be performed by a parallel computer can be performed by a serial one.
I was hoping to get some clarification on this, not just a restatement, but okay. Maybe I'll have to read that book.
What you describe sounds more like a quantum computer, which is a qualitatively different type of beast. (although I'm not sure I understand precisely what you mean by "written to twice simultaneously).
Could be.
The reason why I this "written to twice" computer is that the term "parallel computer" seems to be often used to refer to systems which are equivalent to multiple serial computers running at the same time, like a cluster of x86 computers. Sure, I agree that this can be emulated by a serial computer. This type of parallel computer operates in discrete steps, so presumably it can be emulated by running through all those steps on a serial computer.
But what about a system where data in the system can be written to twice simultaneously? In other words, one where multiple read/write heads are changing the symbol on the same spot on the tape at the same time. This may sound like a weird beast. But isn't a neuron an example of such a thing?
Certainly a neuron is performing calculations. If the voltage differential across a calcium channel hits a certain threshold, the channel will open (a "bit" of data": 1=open, 0=closed). If it opens in the right place and at the right time, the neuron will fire. Neurons are floating in a sea of ions, each with its own electric field, and controlling the flow of these ions is how neurons influence one another. The neuron "sums" voltage differences at several points and determines whether they exceed certain thresholds. Is this not a case of multiple read/write heads altering the same bit of tape at the same time?
Now, can a serial computer emulate multiple read/write heads altering the tape at the same time? I think this is the same question as "How would you emulate this neuron?" So how do you emulate this neuron on a serial computer? You can certainly add up the electric fields of all of the ions at any given time and determine the state of the neuron at one point in time. Serial computers can do that just fine. But to properly emulate the neuron, you need to do this with infinite temporal resolution. The _exact_ instant when that calcium channel opens could affect the results of the neuron's calculations (e.g. whether or not one neuron causes the next one to fire could potentially depend on infinitely small time differences). So perfect precision is required. I don't think a serial computer can do that.
Anyway, that's all I was getting at.
A Parallel computer can't actually do anything that a serial computer can't do, other then doing things more efficiently.
I'd like to know what you mean exactly. I think there may be some limitations to that assertion.
I agree that an SMP computer or a Beowulf cluster can't do anything that a serial computer can do. The main reason being that in some ways (e.g. memory access) these devices are somewhat serial. Any given bit of data can only be accessed (e.g. written to) by one process at a time.
On the other hand, if we had a really parallel computer in which a bit could be written to twice (or more times) simultaneously, could the results be simulated on a serial machine?
In a related vein, it is taken for granted in certain fields that the brain is a massively parallel computer, and that it could never be emulated on a serial computer. This platitude is not just based on the Chinese Room argument either (which I think is probably subtly flawed). Rather, it has to do with the complexity of the "processors" (neurons) in the brain.
Basically, neurons are acted upon my many events simultaneously, and those events are analog (e.g. multiple ion channels are letting several types of ions flow in and out of the neuron at variable rates and in various locations). Basically, you would need to update a model of the state of this neuron at infinitely small intervals. Unless time is quantized, it can't be done.
The *nix community needs to get some serious developer support before companies will really start to look at it seriously. Also the whole RTFM attitude is doing tons of harm to the movement as well. When the only support you can get for an OS is found on the web, and half the responses are along the lines of "RTFM 1d10t, y0ur a 1user, and 1m 37337" this does not instill confidence in that OS.
:) I suspect this is true of any software product or any OS (are the Microsoft/Mac/Whatever groups so much more polite and helpful?)
Now now, you're comparing apples and oranges here.
If you go direct to a Linux vendor like RedHat or a company like The Kompany for product support, you'll get a reasonable response. If you go to usenet for support, you may well get the sort of garbage you were talking about.
In general, if you want commercial-level support, go with a commercial product. That should be obvious. Buy StarOffice instead of using OpenOffice. The question, then, is whether or not enough *nix software has that level of support, and how many users really demand that level of support. That is a complicated question. I suspect that the correct answer is "It depends on the particular app you're talking about".
Of course, it never hurts to have the _option_ to look at the source code, go to Usenet for support, etc. as well.
...generating huge cryptographically strong random numbers.
;)
Good idea. It's getting late, and counting sheep doesn't work too well.
I'm starting a "guess how much karma Jeremy Allison will get today" pool. Anyone want to enter?
This only furthers the evidence that global warming has not been caused by human action.
Actually, it doesn't, due to the near total lack of any evidence in the article. It is at most a vague summary of what was concluded.
More to the point, this is just a summary. The study could have massive flaws, which is why we should wait until scientists have had time to critique it before drawing any conclusions, especially given that there are conflicting studies (at least one is mentioned at the start of the article) and reasons to disagree with the article's conclusions (see Dr. Brown's comments at the end of the article). This peer review is important. Sometimes evidence, such as certain well-known satellite results, can turn out to be wrong.
The expectation that Linux will fulfill the hardware driver installation off the distribution CD, when you admit that you may have to replace the entire hardware for XP, is inherently unfair and beyond what can be reasonably expected in any operating system. XP is not without it's major problems when it comes to older hardware ( especially scanners ) support and driver conflict problems.
Personally, I think you've got the right idea, but I'd put it a bit differently.
This writer left out one important thing: How long did it take to set up all that hardware on Windows 95? She seems to expect Linux to have the ease of install of a Windows XP, and when it Linux doesn't install that well on her hardware, she blames Linux. Yet it's not clear than any OS could install that easily. XP is one thing, but how much effort did it take to get multiple SCSI scaners and multiple parallel-port devices working on Win95? Did it take some work?
For a proper comparison, this review should have thoroughly discusses Linux vs. Win95, and probably vs. XP as well. Here's what we got instead:
* WinXP supposedly wouldn't work, though it wasn't tested.
* Linux (various distros) was a major pain.
* Win95 works, but we don't know if it was a pain or not.
Unfortunately, that's not a very informative comparison. It's comparing seven years of experience tweaking Win95 to maybe 4-6 hours per distro with Linux to 0 hours with XP or another modern MS operating system.
The poll I saw showed that ~65% supported the war on Iraq, but ~20% believed that all the Sept. 11 hijackers were Iraqi
I've seen a poll which claimed that 53% of the population believes that Hussein was personally responsible for the terrorist attacks on 9/11. A recent Tom Tomorrow cartoon claimed that it was 45%.
So yes, you're right. There are a lot of clueless types (or at least people who have a lot of facts wrong) supporting the war.
I've always called it the 90-10 rule. 90% of the work takes 10% of the time, 10% of the work takes 90% of the time.
The really scary part is that this rule is recursive.
Well, no...I guess the picture you get depends upon the media you get your news from, but a couple of weeks before kick off when opposition to the war was at its strongest here in the UK, it was - disappointingly - very much under-reported even on domestic BBC news broadcasts. CNN etc. seemed to be showing even more anti war activity in the US so assuming the inevitable biased reporting I'd have to guess that opposition to the war in the US was at the least very significant. At the same time also, the increasing pro-war "support" seemed to consist, for most people, of a kind of resigned and grudging acceptance (pretty much as it has been in the UK).
Although there has been lots of anti-war protesting in the U.S., the general public is for the war. The last poll I saw showed 68% of the public supporting the war. That's a pretty solid majority.
The Second Superpower has no teeth. As witness the way support for the anti-war movement just melted away in the last days leading up to the war - right about the time it became really clear that Bush didn't give a flying fuck about public opinion. The vast majority of people just gave up. I don't wish to be hypocritical; I count myself among them.
I assume you mean _International_ Public Opinion (from George Bush's point of view, that is). With as popular as this war is in the U.S., and with Bush's otherwise lackluster record (what else did he do these last two years except wars and tax cuts?), I would recommend viewing this war as a lead in to the next election season. Remember, the Primaries are only 9 months away, and the campaigning will doubtless start well before that...
There's a small joke that goes around in the academic world
Biologists like to think they are chemists. Chemists like to think they are physicists. Physicists like to think they are mathematicians. And mathematicians like to think they are god.
FYI, there's another version of that joke which ends on "philosophers".