Uhmm, yes. You are slowly killing yourself by doing this. Iodine tablets are intended as a temporary solution, such as during a disaster, camping trip, etc... It's not considered healthy to use them all your life. One producer, potable aqua, recommends that "extended daily use should not exceed six weeks". This limit is of course totally arbitrary and highly conservative, and most likely safe to exceed, but I assume they put it in their brochures for a reason. Research is scarce, and it may turn out that iodine tablets is safe after all, but personally, I wouldn't take the risk. If I were you, and I lived somewhere where I needed to filter my water, I would probably search for a healthier long-term alternative than iodine tablets.
Wonder how that could be applied to computer memory...
I believe that if we create an AI that has a memory which is working in exactly the same way as ours, with exactly the same mechanism for remembering and/or forgetting things, that we could teach the AI these mnemotic tecniques, just like we can learn them ourselves. Apart from that, it's not applicable.
the mathematical units have grown faster and massively parallel in nature!
This is not english. Even if I try, I can't guess what you are trying to tell. It makes no sense whatsoever...
We haven't done much past anything turing computable anyway
Please understand what "turing computable" means. All computers are devices capable of emulating a turing machine (at least if it fits within RAM, etc). And computers is something you can emulate on a turing machine. Your criticism is like someone complaining that "In the field of cars, we haven't done much past transportation anyway".
It's crazy, but they keep shuffling things over the years, people aren't bored of it anyway, everybody buys stuff and everybody wins.
It's not crazy. It's common sense. If someone comes up with a better part for a computer, someone is going to buy it. If someone comes up with a better part for a lawn-mower, someone is going to buy it. Despite what you may think when you see todays youth, I believe common sense is here to stay.
Intel wants more GPU power in their CPU's. NVidia is using features of their GPU to do problem solving. Which one will win out?
The answer is obvious. Untill GPUs run Windows, and motherboards have slots for GPUs, but you add the CPU on a special "processing board" connected over a fast PCI Express slot, Intel wins.
God forbid somebody should be allowed indefinite free tech support from their children, when they are like the average computer user: completely clueless about even the simplest things, will get a virus/spyware/whatever in less than 20 seconds if unsupervised, does not remember instructions, and is not interested in learning why things go wrong, since you are always there to fix them.
Of course, a much better solution is what I do. Don't give free tech support. I think most people will find that easier to understand, then that you are somehow forcing them to use some strange risky internet thing on their computer, even though they already have paid for and own a legal copy of the worlds most famous operating system: windows!
No, an "xxx algorithm" really means an algorithm for doing xxx. Try "sorting algorithm". Is it an algorithm that uses sorting to e.g. analyze server logs, or is it an algorithm you use for sorting?
In a large program, it isn't always transparent where a value comes from. You can't just put a check in front of the division operator. What are you going to do in the case of zero? Quit the program? That's no better than the OS killing your program because of divide-by-zero error. In a large program, only one person (if you are that lucky!), the architect, designs the program according to the customers (constantly changing) specifications. Different parts of the program are written by different programmers working on different teams, where each programmer only has a limited understanding of the part his team works on. Nobody has a full understanding of the whole system. Parts that are already written can be affected by changes to the spec, without anyone realizing. Large systems are difficult. Sarcasm doesn't help understanding the issues involved.
As for your last point. Maybe you are right. Perhaps I like being called an asshole.
In english "division algorithm" means a step-by-step description of how one does division . It does not mean a step-by-step description of how one does something else that is unspecified, that at some step involves a division. (Well, technically this isn't true either. The division algorithm is a mathematical theorem)
And yes, anywhere you use division in a program, you should make sure that the denominator is non-zero. However, if you have participated in writing or maintained a real-world program with more than 100000 lines of code, you will realize that errors happen, even though the programmers who have been working on it are highly talented people. That's the difference between theory and practice.
And no, that part of the package was certainly experimental. If it wasn't, it would not exhibit this embarassing error. If you wish, you can continue to insist that the subroutine that failed so spectacularly in this experimental software, was a reliable well-tested and well-understood piece of code. But I doubt you will convince many.
Entering a zero into a field causes the ship's propulsion to die because some programmer, and all his reviewers, couldn't be bothered to check for zero in a division algorithm.
Well, that's probably because the programmer didn't write the division algorithm himself. I may be going out on a limb here, but I believe the programmer may have used a built-in operator from the programming language he was using, the operator being called "/".
But seriously, these sort of things happen. And in fact, at the time of the incident, according to the article, Yorktown was "used as the test bed for the Navy's Smart Ship program". Do you expect the programmers to come up with perfect software even before it's being tested on a ship? Don't you realize that finding these sort of glitches is exactly the reason why they called the ship a "test bed"? Sure, it's embarassing, but it would be more worrying if this was found in production software. And apart from this bad experience, I'm shure they had other interesting experiences with this "test bed", some successes, and some failures.
You are confusing the issues. First, the finished code must be thrustworthy. This is done by having skilled programmers, skilled managers, using the right tools, the right development methods, as well as the right testing methods. And of course you also need to be able to trust those programmers, so they don't put in backdoors, deliberate bugs, etc. Although the risk is probably pretty low.
Secondly, you need everyone involved in the project to be thrustworthy. Having the best team in the world develop the greatest defence software in the world, matters little if your enemies are able to bribe the team-members later, to get all the nasty details about the software (and maybe even a copy of it, straight out of CVS). The bribing could happen weeks, months, or years after the project is over, and the team members unemployed and starving.
Having people you can trust, is easier in your own country than when hiring a foreigner to work from a foreign country. And controlling the security and compartmentalization (nobody needs to know more than their part), is also easier when it's done within your own borders. That's why, when you work for a defence contractor, you usually need a security clearance in addition to being qualified for the job. Obviously, you can't legally conduct extensive security checks abroad, and one such check would probably cost a lot more than just hiring a domestic developer.
Of course, compartmentalization leads to only one or very few people getting the whole picture, which again leads to a loss of code thrustworthyness. These are all tradeoffs.
Look, I don't need your permission to complain about something. Firefox has a problem with memory usage. You choose to deny this based on personal experience and ideological reasons (i.e. I need to contribute in order to have a right to complain, what kind of twisted mind would come up with something like that?) I don't care about your denial other than to try to get the facts right. And I don't care much for your ideology either.
While it's obviously true that the firefox developers "have been hunting for memory leaks and fixing them", it doesn't mean that they are finished. In fact, these kinds of tasks usually never get finished. Whenever you introduce a new feature, fix a bug, rewrite or refactor code, or just do some change, there is a risk of introducing a new memory error. That is the nature of programming, and especially the nature of programming in C++. And with a program with gazillions of lines of source code, nobody will be able to finish finding the combinatorial explosion of memory errors that already exist, in a reasonable time-frame, regardless or resources at disposal. It's like telling someone that you cleaned your kitchen 10 years ago.
Then you're obviously doing something unusual. Been browsing multiple tabs in Firefox 2 for a few hours now. Only 47 MB used. Or are you running 80 tabs each with a stupid Java applet and 60 tabs with a complex nested table test in each?
No, I'm not doing anything unusual. Yes, I'm running 80 tabs each with a stupid Java applet and 60 tabs with a complex nested table in each. Java applets and complex nested tables are out there for a while, actually since before firefox development was started. I expect my browser not to leak memory just because it encounters it.
Also, I expect my browser to use sane amounts of memory for each tab. If opening another tab uses umpteen megabytes of memory, something is wrong. A tab should take no more memory than a few pointers, as well as some resources at the display server. Firefox already keeps closed tabs in the cache. I expect it to limit memory cache for open tabs as well!
OK. I've been away for a day. But now, I've come back, and surfed some more. Memory use is now 490 MB. Now, I'm on a machine with enough ram to spare, but this is getting ridiculous.
If you see memory use go up to 300 MB and want the problem fixed, you should describe what you're doing to cause that so the bug can be investigated. Until then, all we can say is that we simply don't see the problem you're referring to.
I'm not going to start debugging memory leaks in firefox. This is something that is better left to firefox developers and testers. I'm not a tester, I'm a user. I use the version of firefox that I installed with apt-get in ubuntu. There might be programs I actively use and care about that I would bother submitting detailed bug-reports about. Firefox is not one of them.
However, 490 MB is completely ridiculous. If the developers feel that going hunting for memory leaks is not of any importance, then that is their decision. I'm not paying for firefox, and therefore I can't vote with my money. But if they care even a tiny little bit about putting out a halfway decent product, then plugging memory leaks should be near the top of their priority list. If I remember correctly, this was one of the main reasons netscape scrapped the development of their previous browser. History repeats itself.
Maybe it's because you aren't surfing. If you actually use firefox, instead of simply letting it remain minimized at about:blank for a month, it will quickly use ridiculous amounts of memory. Here it's at 305MB and counting... Oops, just checked, it's now 306MB. And I haven't really done anything apart from writing this in the meantime.
...I am willing to assume that correlation equals causation. It is now proven: Men with low sperm-quality talk more on the phone than men with good sperm-quality. Women (who doesn't even have sperm), talk even more!
Well, the first public version of netscape navigator was pretty much exactly similar to ncsa mosaic, apart from some cosmetic changes. Netscape navigator was never "inspired" by mosaic. A more fitting description would be a name-change, or a fork. And it wasn't any more user-friendly than mosaic either. Probably because it was more or less written and maintained by the same people, working on the same code-base, but now employed by Netscape instead of NCSA. A company that was formed for the single reason that everyone that tried mosaic, instinctly knew the web was going to be the big next thing. The fact that the article mentions NCSA Mosaic is completely irrelevant, as long as it completely misrepresents it.
This man uses several pages to talk about the origins of the web and how revolutionary netscape navigator was, but he doesn't even remember it's immediate predecessor NCSA Mosaic, or the predecessor of the web: gopher? And you expect me to think this person is more qualified to predict the future of the web, than someone else, such as my grandmother?
Well obviously, it's not your job to handle assholes. It is my job. There are two ways to handle assholes, either you retract, or you show them who's the boss. If you do the latter, sooner or later, one of them will attack you. It's just human nature and statistics. Sure, with 99.9% of people, it is all about your ability to deal with human psychology, but the last 0.1% is just statistics. Some people are so fucked up in their heads that no amount of psychology or understanding from your side can make them act rationally. In that case it's my job to make sure nobody's hurt, including them.
So, sure, you can stay prepared. But does staying prepared help you to de-escalate a situation? If you are good at talking, you can do much better. You can at the same time as you encounter a dangerous individual (or group), calm them down, get them to act the way you want (and/or leave the area), and befriend them, so they are at your side if they ever witness another confrontation you have with some asshole. Now, that's psychology!
As for civilian life, I've never had a physical altercation (well, one when I was 11 years old). And I don't need a gun, but then again, I don't live a place where other people feel they need a gun either.
As for fingerprinting, I'm quite sure most of the people who are against it, also pay by credit card. I fail to see much difference. If you want to stay anonymous, don't go to public places where they demand your fingerprint, or ID, or whatever...
Learn to defend yourself you little pussy and they government won't have to do it for you.
Let me put it this way: I've probably handled more assaults by assholes than you have. And I've probably been in far more hostile situations related to gang-crime than you ever will. If you want to defend yourself, all you need to learn is to talk with people. If you think you can somehow magically pop the steroid pill, study martial arts, and beat any thug who gets into your way, you will realize the problem you've got when his 15 other buddies show up. That's why most people opt for the police, the alternative is the mafia. But if you absolutely want to pick a fight, you've probably found the right person, as I will not hurt you unless I have to. Other people you meet aren't necessarily nice people like me!
But seriously, the bars that I occasionally visit will ban troublemakers. How do they manage? Because when you go into a place they get to see your face and know what you look like.
So, exactly what's the difference of having an even more efficient system? Has it ever happened to you that you fail to recognize some person you have only seen one time? Has it ever happened to you that you fail to recognize some person you've only heard described by a person who has only seen him once? Are you usually working at a place where there are hundreds of people walking in and out every night, and not the same people either?
So let me get this straight. What you are basically saying is "Don't judge them by what they do, judge them by what they say!".
It seems to me that you need a good beating with a clue-stick if this is your position in everyday life. Now, it might just happen to be true in this specific case, with this specific company, in this specific time-period. And you may have other data to back this assumption up with. But from the data presented so far, I would consider the closed DRM-enabled platform with more skepticism then a competing open platform.
Slashdot Mirror
Uhmm, yes. You are slowly killing yourself by doing this. Iodine tablets are intended as a temporary solution, such as during a disaster, camping trip, etc... It's not considered healthy to use them all your life. One producer, potable aqua, recommends that "extended daily use should not exceed six weeks". This limit is of course totally arbitrary and highly conservative, and most likely safe to exceed, but I assume they put it in their brochures for a reason. Research is scarce, and it may turn out that iodine tablets is safe after all, but personally, I wouldn't take the risk. If I were you, and I lived somewhere where I needed to filter my water, I would probably search for a healthier long-term alternative than iodine tablets.
That's simply because "devolvation" is not dictionarified yet. But you comprehense it anyway.
I believe that if we create an AI that has a memory which is working in exactly the same way as ours, with exactly the same mechanism for remembering and/or forgetting things, that we could teach the AI these mnemotic tecniques, just like we can learn them ourselves. Apart from that, it's not applicable.
This is not english. Even if I try, I can't guess what you are trying to tell. It makes no sense whatsoever...
Please understand what "turing computable" means. All computers are devices capable of emulating a turing machine (at least if it fits within RAM, etc). And computers is something you can emulate on a turing machine. Your criticism is like someone complaining that "In the field of cars, we haven't done much past transportation anyway".
It's not crazy. It's common sense. If someone comes up with a better part for a computer, someone is going to buy it. If someone comes up with a better part for a lawn-mower, someone is going to buy it. Despite what you may think when you see todays youth, I believe common sense is here to stay.
It can run certain computational tasks faster.
You seriously need to back up this statement with some hard data, otherwise most people will think you are bullshitting us (and so do I). Link please?
Yeah, that sounds like a rational, well thought out plan to get rid of spam. Why don't you start doing it?
God forbid somebody should be allowed indefinite free tech support from their children, when they are like the average computer user: completely clueless about even the simplest things, will get a virus/spyware/whatever in less than 20 seconds if unsupervised, does not remember instructions, and is not interested in learning why things go wrong, since you are always there to fix them.
Of course, a much better solution is what I do. Don't give free tech support. I think most people will find that easier to understand, then that you are somehow forcing them to use some strange risky internet thing on their computer, even though they already have paid for and own a legal copy of the worlds most famous operating system: windows!
No, an "xxx algorithm" really means an algorithm for doing xxx. Try "sorting algorithm". Is it an algorithm that uses sorting to e.g. analyze server logs, or is it an algorithm you use for sorting?
In a large program, it isn't always transparent where a value comes from. You can't just put a check in front of the division operator. What are you going to do in the case of zero? Quit the program? That's no better than the OS killing your program because of divide-by-zero error. In a large program, only one person (if you are that lucky!), the architect, designs the program according to the customers (constantly changing) specifications. Different parts of the program are written by different programmers working on different teams, where each programmer only has a limited understanding of the part his team works on. Nobody has a full understanding of the whole system. Parts that are already written can be affected by changes to the spec, without anyone realizing. Large systems are difficult. Sarcasm doesn't help understanding the issues involved.
As for your last point. Maybe you are right. Perhaps I like being called an asshole.
Why do you think that we would want you here. Please stay in America.
In english "division algorithm" means a step-by-step description of how one does division . It does not mean a step-by-step description of how one does something else that is unspecified, that at some step involves a division. (Well, technically this isn't true either. The division algorithm is a mathematical theorem)
And yes, anywhere you use division in a program, you should make sure that the denominator is non-zero. However, if you have participated in writing or maintained a real-world program with more than 100000 lines of code, you will realize that errors happen, even though the programmers who have been working on it are highly talented people. That's the difference between theory and practice.
And no, that part of the package was certainly experimental. If it wasn't, it would not exhibit this embarassing error. If you wish, you can continue to insist that the subroutine that failed so spectacularly in this experimental software, was a reliable well-tested and well-understood piece of code. But I doubt you will convince many.
By the way, I don't like being called an asshole.
Entering a zero into a field causes the ship's propulsion to die because some programmer, and all his reviewers, couldn't be bothered to check for zero in a division algorithm.
Well, that's probably because the programmer didn't write the division algorithm himself. I may be going out on a limb here, but I believe the programmer may have used a built-in operator from the programming language he was using, the operator being called "/".
But seriously, these sort of things happen. And in fact, at the time of the incident, according to the article, Yorktown was "used as the test bed for the Navy's Smart Ship program". Do you expect the programmers to come up with perfect software even before it's being tested on a ship? Don't you realize that finding these sort of glitches is exactly the reason why they called the ship a "test bed"? Sure, it's embarassing, but it would be more worrying if this was found in production software. And apart from this bad experience, I'm shure they had other interesting experiences with this "test bed", some successes, and some failures.
You are confusing the issues. First, the finished code must be thrustworthy. This is done by having skilled programmers, skilled managers, using the right tools, the right development methods, as well as the right testing methods. And of course you also need to be able to trust those programmers, so they don't put in backdoors, deliberate bugs, etc. Although the risk is probably pretty low.
Secondly, you need everyone involved in the project to be thrustworthy. Having the best team in the world develop the greatest defence software in the world, matters little if your enemies are able to bribe the team-members later, to get all the nasty details about the software (and maybe even a copy of it, straight out of CVS). The bribing could happen weeks, months, or years after the project is over, and the team members unemployed and starving.
Having people you can trust, is easier in your own country than when hiring a foreigner to work from a foreign country. And controlling the security and compartmentalization (nobody needs to know more than their part), is also easier when it's done within your own borders. That's why, when you work for a defence contractor, you usually need a security clearance in addition to being qualified for the job. Obviously, you can't legally conduct extensive security checks abroad, and one such check would probably cost a lot more than just hiring a domestic developer.
Of course, compartmentalization leads to only one or very few people getting the whole picture, which again leads to a loss of code thrustworthyness. These are all tradeoffs.
Look, I don't need your permission to complain about something. Firefox has a problem with memory usage. You choose to deny this based on personal experience and ideological reasons (i.e. I need to contribute in order to have a right to complain, what kind of twisted mind would come up with something like that?) I don't care about your denial other than to try to get the facts right. And I don't care much for your ideology either.
While it's obviously true that the firefox developers "have been hunting for memory leaks and fixing them", it doesn't mean that they are finished. In fact, these kinds of tasks usually never get finished. Whenever you introduce a new feature, fix a bug, rewrite or refactor code, or just do some change, there is a risk of introducing a new memory error. That is the nature of programming, and especially the nature of programming in C++. And with a program with gazillions of lines of source code, nobody will be able to finish finding the combinatorial explosion of memory errors that already exist, in a reasonable time-frame, regardless or resources at disposal. It's like telling someone that you cleaned your kitchen 10 years ago.
No, I'm not doing anything unusual. Yes, I'm running 80 tabs each with a stupid Java applet and 60 tabs with a complex nested table in each. Java applets and complex nested tables are out there for a while, actually since before firefox development was started. I expect my browser not to leak memory just because it encounters it.
Also, I expect my browser to use sane amounts of memory for each tab. If opening another tab uses umpteen megabytes of memory, something is wrong. A tab should take no more memory than a few pointers, as well as some resources at the display server. Firefox already keeps closed tabs in the cache. I expect it to limit memory cache for open tabs as well!
I'm not going to start debugging memory leaks in firefox. This is something that is better left to firefox developers and testers. I'm not a tester, I'm a user. I use the version of firefox that I installed with apt-get in ubuntu. There might be programs I actively use and care about that I would bother submitting detailed bug-reports about. Firefox is not one of them.
However, 490 MB is completely ridiculous. If the developers feel that going hunting for memory leaks is not of any importance, then that is their decision. I'm not paying for firefox, and therefore I can't vote with my money. But if they care even a tiny little bit about putting out a halfway decent product, then plugging memory leaks should be near the top of their priority list. If I remember correctly, this was one of the main reasons netscape scrapped the development of their previous browser. History repeats itself.
Maybe it's because you aren't surfing. If you actually use firefox, instead of simply letting it remain minimized at about:blank for a month, it will quickly use ridiculous amounts of memory. Here it's at 305MB and counting... Oops, just checked, it's now 306MB. And I haven't really done anything apart from writing this in the meantime.
Of course the most annoying thing is people who decide to write C instead of c.
...I am willing to assume that correlation equals causation. It is now proven: Men with low sperm-quality talk more on the phone than men with good sperm-quality. Women (who doesn't even have sperm), talk even more!
Well, the first public version of netscape navigator was pretty much exactly similar to ncsa mosaic, apart from some cosmetic changes. Netscape navigator was never "inspired" by mosaic. A more fitting description would be a name-change, or a fork. And it wasn't any more user-friendly than mosaic either. Probably because it was more or less written and maintained by the same people, working on the same code-base, but now employed by Netscape instead of NCSA. A company that was formed for the single reason that everyone that tried mosaic, instinctly knew the web was going to be the big next thing. The fact that the article mentions NCSA Mosaic is completely irrelevant, as long as it completely misrepresents it.
Sounds reasonable. After all, consoles got colour graphics before business computers too.
This man uses several pages to talk about the origins of the web and how revolutionary netscape navigator was, but he doesn't even remember it's immediate predecessor NCSA Mosaic, or the predecessor of the web: gopher? And you expect me to think this person is more qualified to predict the future of the web, than someone else, such as my grandmother?
Well obviously, it's not your job to handle assholes. It is my job. There are two ways to handle assholes, either you retract, or you show them who's the boss. If you do the latter, sooner or later, one of them will attack you. It's just human nature and statistics. Sure, with 99.9% of people, it is all about your ability to deal with human psychology, but the last 0.1% is just statistics. Some people are so fucked up in their heads that no amount of psychology or understanding from your side can make them act rationally. In that case it's my job to make sure nobody's hurt, including them.
So, sure, you can stay prepared. But does staying prepared help you to de-escalate a situation? If you are good at talking, you can do much better. You can at the same time as you encounter a dangerous individual (or group), calm them down, get them to act the way you want (and/or leave the area), and befriend them, so they are at your side if they ever witness another confrontation you have with some asshole. Now, that's psychology!
As for civilian life, I've never had a physical altercation (well, one when I was 11 years old). And I don't need a gun, but then again, I don't live a place where other people feel they need a gun either.
As for fingerprinting, I'm quite sure most of the people who are against it, also pay by credit card. I fail to see much difference. If you want to stay anonymous, don't go to public places where they demand your fingerprint, or ID, or whatever...
Let me put it this way: I've probably handled more assaults by assholes than you have. And I've probably been in far more hostile situations related to gang-crime than you ever will. If you want to defend yourself, all you need to learn is to talk with people. If you think you can somehow magically pop the steroid pill, study martial arts, and beat any thug who gets into your way, you will realize the problem you've got when his 15 other buddies show up. That's why most people opt for the police, the alternative is the mafia. But if you absolutely want to pick a fight, you've probably found the right person, as I will not hurt you unless I have to. Other people you meet aren't necessarily nice people like me!
So, exactly what's the difference of having an even more efficient system? Has it ever happened to you that you fail to recognize some person you have only seen one time? Has it ever happened to you that you fail to recognize some person you've only heard described by a person who has only seen him once? Are you usually working at a place where there are hundreds of people walking in and out every night, and not the same people either?
So let me get this straight. What you are basically saying is "Don't judge them by what they do, judge them by what they say!".
It seems to me that you need a good beating with a clue-stick if this is your position in everyday life. Now, it might just happen to be true in this specific case, with this specific company, in this specific time-period. And you may have other data to back this assumption up with. But from the data presented so far, I would consider the closed DRM-enabled platform with more skepticism then a competing open platform.