Generic answer to any justification for amissing feature that uses that phrase: Bite me; I'm the fracking user here, I should be allowed to do whatever I need.
Specifically, my SVN repository has 30+ top level directories. One is needed by all projects, a couple are needed by a handful of projects, and most are needed by just one project. Updating every project I never have and never will work on every time would be stupid.
I've no doubt Git blows the doors off SVN for managing the Linux kernel tree. But for my purposes, the issues KiloByte lists are big deals, and the distributed advantages of Git are pointless.
OK, if you want to intentionally keep one copy of data only, and keep it on magnetic tape, and have it last forever, you should do a refresh every 6 months or so (or just every time you get back from having your head examined).
But the chances that a particular tape (or 3" floppy) are still good after years of storage are still pretty good.
"Think of all the money they could have saved if they sent a few robots up there."
They could have sent thousands of robots.
We've got two rovers operating on Mars for years for a fraction of the price it costs to send one human to the IIS in low Earth orbit. There's no question the robots get you more science for your buck, all the humans cling to is that they are better PR, but I wonder if that's true anymore? Here's a test: Without looking it up, think of the names of those rovers on Mars. Now think of the names of the current ISS inhabitants. You're paying hundreds of times as much for every day the ISS inhabitant is there.
"They are not as bad as tapes where you need to rewrite the whole tape once every 6 months or your lose your data"
What tapes are you talking about? I've dealt with, off the top of my head, 5 types of media that might be called "tapes", and this wasn't close to true of any of them.
"Is it unreasonable to have some sort of secondary screening if you set off some set of flags? So you are detained a few minutes. That makes sense."
The point being made here is that no, it does not make sense. It is a balancing act, and there is at least some cost (time, money, privacy, convenience), and thanks to the base rate fallacy (one name for what we're discussing) there isn't any security being gained by trying to detect if I'm a terrorist at the airport. There just aren't enough terrorists to possibly detect them by the means available at that point. Trying to detect if I'm carrying a bunch of explosives is another matter; that's potentially doable, and we should. (not by scanning shoes, but I digress...)
"If the screening does not work, point to a story where a real terrorist made it on the plane with weapons or contraband."
Perhaps you missed my point with the red shirt, let me spell it out: I've got this lucky red shirt. If I wear it when I go to the doctor, I don't get cancer. If the shirt doesn't prevent cancer, show me the tumor. Just because a terrorist hasn't gotten on a (US) plane doesn't mean the screening works. It could also mean they haven't tried, or that they've been prevented by other means. Maybe theoretically possible ones, for example.
"they need some time to trickle your name through their information algorithms to see if something hits. Seems reasonable to me, and it is not much of an imposition on you. If you don't like it, you can always drive or walk "
It's the trickling names through information algorithms that is demonstrably, obviously useless. And while I resent the pointless imposition; the waste of my tax dollars is not saved even if I drive or walk.
"Do you automatically assume they are bad, or do you say we put them through a tougher form of screening."
Well, obviously you can't assume they're bad, but it might not even make sense to put them through tougher screening. How much does the extra screening cost, in dollars and inconvenience to innocent people? What is this second-round screenings false-positive rate?
If you've got a 99% accurate test, that gets you down from a million to 10,000. Now you put those people through your more expensive and inconvenient second round test, which is also 99% accurate, and you're down to 100 wrong people. If you're got two completely independent 99% accurate tests on a population of a million people, it's still almost certain that someone you identify as a potential terrorist is perfectly innocent.
Also note that 99% accuracy is a fantasy: it's never anywhere close to that good. And the populations you're searching are all much bigger than 1 million people.
Algorithms for NP complete problems is an odd analogy, but I guess it works: We attack these by accepting very good solutions rather than insisting on provably optimal ones. If you don't care about finding the actual terrorist, but just want any one of the 100 people most likely to be a terrorist, massive screening is great. If you want to find the provably optimal solution in reasonable time, or the actual terrorist at reasonable expense... sorry.
"So it does not work perfectly, I believe your math. What should we do? Stick our heads in the sand and ignore the threat? Rationalize that you are more likely to die in a car accident, so take no action?"
There are more options for what to do than "anything" and "nothing". We should do things that make sense, and that work. If someone points out that one thing we could do doesn't work, it does not make sense to say "Oh well, we gotta do something". We shouldn't do things that don't work, not even if we can't think of anything that does.
"I think people that pay cash for a one-way airline ticket need extra scrutiny."
How much extra scrutiny, at how much extra cost? It depends on how likely they are to be bad guys, doesn't it? People who buy one way tickets with cash are almost guaranteed to not be terrorists, because a lot of people do that every day for perfectly reasonable reasons, and there aren't very many terrorists. That's not even considering that actual terrorists can trivially adapt to your test and avoid scrutiny by not doing that. Spending any resources looking at last-minute one-way ticket buyers is a waste.
"If the DHS is set up to fail, they appear to have not had any failures in the last few years. May not be perfect, but maybe it is working?"
I wore my lucky red shirt to the doctors office again, and again I didn't have cancer...
DHS/TSA, for all I know, may be doing various effective, but less visible things. The specific, visible task of identifying terrorists at airport security checkpoints is basically impossible.
Why does someone use so many big, BS-laden words and not log in? You don't know what you're talking about, as I'd be happy to explain, but I've sworn off making substantive replies to ACs. Sigh. The idiocy level is tempting, but no, I must be strong.
"You can take a public domain work an release it under an restricted license. "
Yes! Yes! I emphatically agree! Please stop telling me this as it is beside the point.
"This are all examples not for 'make a proprietary work based on public domain work' but for turning a public domain work into a proprietary work..."
No! No! You are on drugs! Please stop telling me this as it is not true.
The public domain work remains public domain. If I release a work based on the public domain work under a proprietary license, then take the only copy of the public domain work, set it on fire and shoot it into outer space, I have still not made the public domain work proprietary. "For all practical purposes"? Yes. "Really Truly"? No.
"Therefor we need copyleft,... to make sure that the original will be free as long as it exists."
Your example of the nasty thing one might do with a public domain work is predicated on the original no longer existing. As long as the original exists, it will be free. One should not simplify ideas to the point of falsehood, because reasoning from falsehoods will lead you to more falsehoods, like this. Copyleft ensures arguably wonderful things about derivative works. If I've got the original, public domain is just as good, or better (for me).
You're correct of course. I'd forgotten wall current would be a sine wave. The one time I drove LEDs off an alternating source, it was a bicycle-hub generator produced a square wave, so the rectifier was sufficient.
The original public domain work remains public domain. Necessarily. Always. Forever. There is no extant legal mechanism that can make a public domain work proprietary.
Everything you are describing are examples of using public domain works to make proprietary works. You may argue that doing so is a bad thing if you like, but it is not the same as making the public domain work itself proprietary. Because that is impossible.
The statement "Public domain works can be used to make proprietary works" may be the start of an excellent argument in favor of copyleft. The statement "Public domain works can be made proprietary" is simply false.
You certainly should put resistors in series as necessary to limit the current to the right level.
But that doesn't prevent the cascading failure problem I described with putting them in parrallel. Between putting two leds in parallel, each with their own resistor, and putting the two of them in series with one smaller resistor, series is better if you've got the voltage. One LED failing will cause the other to go out either way, but it won't destroy it if it's in series.
Some people want to use a network connection in one sort of way. Other people want to use a network connection in a different sort of way. Pricing models may have to adapt to let everyone pay for what they want in a reasonable way.
Because someone wants a different level of internet service than you, that doesn't make them a bad person.
One often wires LEDs in series. If you put them in parallel, when one becomes damaged the excess current will flow through the others, and LEDs are easily destroyed by excess current. So you're not gaining anything over series, they still all go out when one does, but parallel will make sure all your LEDs are destroyed.
"I don't understand why threads are useful, except if you have multiple processors"
OK, but are you suggesting multiple core processors are, at this point, rare? Yes, threads are only really useful if you have the sort of processor most new computers have.
Or rather, each hit is handed off to a separate process that handles it, but assuming there are no errors, that process will be reused for a future hit. Thus, we can still cache stuff and skip startup costs most of the time, but still use code that may not be thread safe. I entirely agree with the articles author; third party code that claims to be thread safe often isn't. With process isolation, I don't have to care.
Note that I'm running fairly low volume but extremely processor intensive web apps; not high numbers of hits to simple html pages. YMMV.
"As someone involved in the Web at the time, this really was something of a non-event"
Certainly. I had my first job out of college, and my boss tasked me with setting up a fairly complex gopher site, as well as investigating alternatives. (I believe the phrase "this http thing" was used.) The webs technological superiority was so blindingly obvious, the licensing was not even a factor.
Um, no, actually it can't. Proprietary works can be made based on public domain works, but the public domain works are still public domain. It does not help the FSF argument that they are fond of simplifying concepts like this to the point of falsehood.
Shocking, I know, but the Slashdot summary is somewhat inaccurate on this point.
It isn't the 15th anniversary of the Internet, obviously. Nor is it the 15th anniversary of the Web, though that's closer. It's the 15th anniversary of the day when CERN put their code for the first web server and browser into the public domain.
We're still a couple months short of the day I first heard of it, which I assume all will agree is the really significant milestone.
"unless you belive the NSA has a secret lab of mathmaticians that are years ahead of the rest of them"
Well, they do send recruiting letters to every Math grad from every fairly good college. I didn't respond to the one they sent me, so I can't tell you if they're putting them all to work flipping burgers in the cafeteria, but that's not my guess.
"Shouldn't maximizing their productivity be the priority here?"
Productivity at what? They're not up there because there was anything particular we wanted to do where it seemed like putting humans in orbit was the best way to go about it. Putting humans in orbit was the whole premise, so without asking why, they spent enormous piles of money to do it. Yes, this is yet further evidence that they still haven't come up with any thing useful for ISS residents to do. Unless you count building the ISS in hopes of getting it completed before we drop it in the ocean because we don't need it for anything.
In the one place I'm aware of such a policy, removing the batteries is not a security measure; it's a convenience measure. When RF transmissions are detected, it is inconvenient to have alarms go off, and to send armed guards to search everyone just because someone forgot where they were and turned their cell phone on. Minimizing false positives is an important part of many security measures.
Slashdot Mirror
"...you shouldn't be allowed to..."
Generic answer to any justification for amissing feature that uses that phrase: Bite me; I'm the fracking user here, I should be allowed to do whatever I need.
Specifically, my SVN repository has 30+ top level directories. One is needed by all projects, a couple are needed by a handful of projects, and most are needed by just one project. Updating every project I never have and never will work on every time would be stupid.
I've no doubt Git blows the doors off SVN for managing the Linux kernel tree. But for my purposes, the issues KiloByte lists are big deals, and the distributed advantages of Git are pointless.
OK, if you want to intentionally keep one copy of data only, and keep it on magnetic tape, and have it last forever, you should do a refresh every 6 months or so (or just every time you get back from having your head examined).
But the chances that a particular tape (or 3" floppy) are still good after years of storage are still pretty good.
"Think of all the money they could have saved if they sent a few robots up there."
They could have sent thousands of robots.
We've got two rovers operating on Mars for years for a fraction of the price it costs to send one human to the IIS in low Earth orbit. There's no question the robots get you more science for your buck, all the humans cling to is that they are better PR, but I wonder if that's true anymore? Here's a test: Without looking it up, think of the names of those rovers on Mars. Now think of the names of the current ISS inhabitants. You're paying hundreds of times as much for every day the ISS inhabitant is there.
"The asteroids are a likely resource for Earth."
A resource for what? What is there that can be more efficiently gotten from asteroids than from somewhere on Earth?
"They are not as bad as tapes where you need to rewrite the whole tape once every 6 months or your lose your data"
What tapes are you talking about? I've dealt with, off the top of my head, 5 types of media that might be called "tapes", and this wasn't close to true of any of them.
"Is it unreasonable to have some sort of secondary screening if you set off some set of flags? So you are detained a few minutes. That makes sense."
The point being made here is that no, it does not make sense. It is a balancing act, and there is at least some cost (time, money, privacy, convenience), and thanks to the base rate fallacy (one name for what we're discussing) there isn't any security being gained by trying to detect if I'm a terrorist at the airport. There just aren't enough terrorists to possibly detect them by the means available at that point. Trying to detect if I'm carrying a bunch of explosives is another matter; that's potentially doable, and we should. (not by scanning shoes, but I digress...)
"If the screening does not work, point to a story where a real terrorist made it on the plane with weapons or contraband."
Perhaps you missed my point with the red shirt, let me spell it out: I've got this lucky red shirt. If I wear it when I go to the doctor, I don't get cancer. If the shirt doesn't prevent cancer, show me the tumor. Just because a terrorist hasn't gotten on a (US) plane doesn't mean the screening works. It could also mean they haven't tried, or that they've been prevented by other means. Maybe theoretically possible ones, for example.
"they need some time to trickle your name through their information algorithms to see if something hits. Seems reasonable to me, and it is not much of an imposition on you. If you don't like it, you can always drive or walk "
It's the trickling names through information algorithms that is demonstrably, obviously useless. And while I resent the pointless imposition; the waste of my tax dollars is not saved even if I drive or walk.
"Do you automatically assume they are bad, or do you say we put them through a tougher form of screening."
Well, obviously you can't assume they're bad, but it might not even make sense to put them through tougher screening. How much does the extra screening cost, in dollars and inconvenience to innocent people? What is this second-round screenings false-positive rate?
If you've got a 99% accurate test, that gets you down from a million to 10,000. Now you put those people through your more expensive and inconvenient second round test, which is also 99% accurate, and you're down to 100 wrong people. If you're got two completely independent 99% accurate tests on a population of a million people, it's still almost certain that someone you identify as a potential terrorist is perfectly innocent.
Also note that 99% accuracy is a fantasy: it's never anywhere close to that good. And the populations you're searching are all much bigger than 1 million people.
Algorithms for NP complete problems is an odd analogy, but I guess it works: We attack these by accepting very good solutions rather than insisting on provably optimal ones. If you don't care about finding the actual terrorist, but just want any one of the 100 people most likely to be a terrorist, massive screening is great. If you want to find the provably optimal solution in reasonable time, or the actual terrorist at reasonable expense... sorry.
"So it does not work perfectly, I believe your math. What should we do? Stick our heads in the sand and ignore the threat? Rationalize that you are more likely to die in a car accident, so take no action?"
There are more options for what to do than "anything" and "nothing". We should do things that make sense, and that work. If someone points out that one thing we could do doesn't work, it does not make sense to say "Oh well, we gotta do something". We shouldn't do things that don't work, not even if we can't think of anything that does.
"I think people that pay cash for a one-way airline ticket need extra scrutiny."
How much extra scrutiny, at how much extra cost? It depends on how likely they are to be bad guys, doesn't it? People who buy one way tickets with cash are almost guaranteed to not be terrorists, because a lot of people do that every day for perfectly reasonable reasons, and there aren't very many terrorists. That's not even considering that actual terrorists can trivially adapt to your test and avoid scrutiny by not doing that. Spending any resources looking at last-minute one-way ticket buyers is a waste.
"If the DHS is set up to fail, they appear to have not had any failures in the last few years. May not be perfect, but maybe it is working?"
I wore my lucky red shirt to the doctors office again, and again I didn't have cancer...
DHS/TSA, for all I know, may be doing various effective, but less visible things. The specific, visible task of identifying terrorists at airport security checkpoints is basically impossible.
Why does someone use so many big, BS-laden words and not log in? You don't know what you're talking about, as I'd be happy to explain, but I've sworn off making substantive replies to ACs. Sigh. The idiocy level is tempting, but no, I must be strong.
"You can take a public domain work an release it under an restricted license. "
Yes! Yes! I emphatically agree! Please stop telling me this as it is beside the point.
"This are all examples not for 'make a proprietary work based on public domain work' but for turning a public domain work into a proprietary work..."
No! No! You are on drugs! Please stop telling me this as it is not true.
The public domain work remains public domain. If I release a work based on the public domain work under a proprietary license, then take the only copy of the public domain work, set it on fire and shoot it into outer space, I have still not made the public domain work proprietary. "For all practical purposes"? Yes. "Really Truly"? No.
"Therefor we need copyleft,
Your example of the nasty thing one might do with a public domain work is predicated on the original no longer existing. As long as the original exists, it will be free. One should not simplify ideas to the point of falsehood, because reasoning from falsehoods will lead you to more falsehoods, like this. Copyleft ensures arguably wonderful things about derivative works. If I've got the original, public domain is just as good, or better (for me).
You're correct of course. I'd forgotten wall current would be a sine wave. The one time I drove LEDs off an alternating source, it was a bicycle-hub generator produced a square wave, so the rectifier was sufficient.
The original public domain work remains public domain. Necessarily. Always. Forever. There is no extant legal mechanism that can make a public domain work proprietary.
Everything you are describing are examples of using public domain works to make proprietary works. You may argue that doing so is a bad thing if you like, but it is not the same as making the public domain work itself proprietary. Because that is impossible.
The statement "Public domain works can be used to make proprietary works" may be the start of an excellent argument in favor of copyleft. The statement "Public domain works can be made proprietary" is simply false.
You certainly should put resistors in series as necessary to limit the current to the right level.
But that doesn't prevent the cascading failure problem I described with putting them in parrallel. Between putting two leds in parallel, each with their own resistor, and putting the two of them in series with one smaller resistor, series is better if you've got the voltage. One LED failing will cause the other to go out either way, but it won't destroy it if it's in series.
Some people want to use a network connection in one sort of way. Other people want to use a network connection in a different sort of way. Pricing models may have to adapt to let everyone pay for what they want in a reasonable way.
Because someone wants a different level of internet service than you, that doesn't make them a bad person.
One often wires LEDs in series. If you put them in parallel, when one becomes damaged the excess current will flow through the others, and LEDs are easily destroyed by excess current. So you're not gaining anything over series, they still all go out when one does, but parallel will make sure all your LEDs are destroyed.
No, sadly hooking up a bridge recifier before the LEDs won't make them flicker faster. It will make them not flicker at all.
"I don't understand why threads are useful, except if you have multiple processors"
OK, but are you suggesting multiple core processors are, at this point, rare? Yes, threads are only really useful if you have the sort of processor most new computers have.
Mine does.
Or rather, each hit is handed off to a separate process that handles it, but assuming there are no errors, that process will be reused for a future hit. Thus, we can still cache stuff and skip startup costs most of the time, but still use code that may not be thread safe. I entirely agree with the articles author; third party code that claims to be thread safe often isn't. With process isolation, I don't have to care.
Note that I'm running fairly low volume but extremely processor intensive web apps; not high numbers of hits to simple html pages. YMMV.
"As someone involved in the Web at the time, this really was something of a non-event"
Certainly. I had my first job out of college, and my boss tasked me with setting up a fairly complex gopher site, as well as investigating alternatives. (I believe the phrase "this http thing" was used.) The webs technological superiority was so blindingly obvious, the licensing was not even a factor.
"Public domain software can be made proprietary."
Um, no, actually it can't. Proprietary works can be made based on public domain works, but the public domain works are still public domain. It does not help the FSF argument that they are fond of simplifying concepts like this to the point of falsehood.
I used to think that incompetent malice was better than competent malice. I stand corrected.
Shocking, I know, but the Slashdot summary is somewhat inaccurate on this point.
It isn't the 15th anniversary of the Internet, obviously. Nor is it the 15th anniversary of the Web, though that's closer. It's the 15th anniversary of the day when CERN put their code for the first web server and browser into the public domain.
We're still a couple months short of the day I first heard of it, which I assume all will agree is the really significant milestone.
"unless you belive the NSA has a secret lab of mathmaticians that are years ahead of the rest of them"
Well, they do send recruiting letters to every Math grad from every fairly good college. I didn't respond to the one they sent me, so I can't tell you if they're putting them all to work flipping burgers in the cafeteria, but that's not my guess.
"Shouldn't maximizing their productivity be the priority here?"
Productivity at what? They're not up there because there was anything particular we wanted to do where it seemed like putting humans in orbit was the best way to go about it. Putting humans in orbit was the whole premise, so without asking why, they spent enormous piles of money to do it. Yes, this is yet further evidence that they still haven't come up with any thing useful for ISS residents to do. Unless you count building the ISS in hopes of getting it completed before we drop it in the ocean because we don't need it for anything.
In the one place I'm aware of such a policy, removing the batteries is not a security measure; it's a convenience measure. When RF transmissions are detected, it is inconvenient to have alarms go off, and to send armed guards to search everyone just because someone forgot where they were and turned their cell phone on. Minimizing false positives is an important part of many security measures.