This is an excellent polemic. I would add it's only part of a larger phenomenon: Concentrated interests dominate diffuse interests. Irrespective of merit, right, numbers, or aggregate welfare.
Media companies represent one concentrated interest. They exist mostly as a middleman between the artist/creator and their public. Their entire livelihood is threatened by new communications/electronic technology. (They mustn't be adding value anywhere else to feel so threatened!) So they fight tooth-and-nail in the Courts and Legislatures, spending money lavishly. No surprise.
Consumers are the diffuse interest. Sure, they like the music/TV/whatever, but they won't die without it. What most are willing to do to protect their rights is very limited. Even if they are very numerous, their interest is difficult to concentrate. So they lose.
Ultimately, this is a very bad thing for society. It harkens a return to feulalism.
An interesting German ruling, and the consequent market response.
M$ has insisted long and hard in US Courts that it is _licencing_ it's products thru shrinkwraps hidden inside. For OEMs and big corps, it sure does get signed licence agreements. But at retail, it behaves as if it were simply selling copyrighted works (books, music, videos).
M$ doesn't do even simple practical things like have purchasers call for a key, insist on registration, or tear-off registration to protect their licencing status. Apparently their marketing department has vetoed these things as expensive or frightening to customers.
So why should the Courts grant them licencing status when M$ has not done what they could for that status?
I wrote an ASM `life` 15 years ago with DEBUG. On a 4.77 MHz 8088 it ran 80x25 at about 1.5 fps. And it was very primitive: read the display memory and calc'd all the cells.
Sure, I could have used a regen buffer and skipped the empty cells. But then it would run like it does on my 500 MHz machine -- too fast, just a blur. Lame excuse:)
What does Honeypot want? Cheap forensic analysis on a cracked box?
Well if you want to try, have a read of the Nov & Dec Dr.Dobbs. It has a pair of articles about recovering deleted data and has pointers to useful tools.
The security comparison URL is here. Caveat lector
Agreed that most of the vulnerabilities are associated with the programs themselves. Newer versions fix old holes, but may make new ones.
The distro can influence the final result by/etc/passwd, permissions, crontabs and the other configurations it makes or simply not installing some questionable pgms by default.
I too am a Slackware afficionado, but I wouldn't call Slackware the "best" without qualifying it. Nor would I call RedHat or Debian or... the unqualified best.
I certainly wouldn't call Slackware the best distro for newbies accustomed to MS-Windows and uninterested in learning the guts of Unix. That would be RedHat (or maybe Corel).
But Slackware has some unique features that probably make it the best distro for someone coming from BSD (or SunOS) or who wants to learn the guts of Unix.
Slackware runs a little behind the times in terms of program updates, but that also means that it has the fewest security holes (URL forgotten).
AFAIK, it is the only distro with a BSD-style/etc/rc.d rather than the mess'o'symlinks SysV-style. That makes administration much easier to learn without depending on a tool like `linuxconf`.
Interesting -- a mechanical computer. Presumably analog. Well they had to use something, and it was probably good to a minute-of-arc. I'd be interested in hearing more about the operating principles (tuned pendulum on a gyro?).
But they're going to need something better when the USN strips the rifling out of those bores and switchs to saboted submunitions with 50+ mile range. Or maybe the shot'll be terminally guided if it's not too long for the loading cradle inside that cramped turret.
Aircraft and ships are constantly being tossed around by the fluid through which they travel.
Fixed-trajectory munitions will miss unless the roll, pitch and yaw are compensated for. Hitting a 10 ft target at 100 miles isn't easy.
What will probably happen is a target is locked into the fire control computer, the operator presses the fire button, and the computer waits a few milliseconds until the weapon is on target (after compensating for aiming offset, refractive index gradients, target movement, etc).
I'm sure the big 16" guns on US Battleships have very sophisticated firing [delay] systems. Just so they can hit the broadside of a barn at 15 miles. Otherwise, they could miss the broadside of a mountain!
Yes, I've seen HDTV from digital source. My photographers mind boggles at the improved details. You can see individual hair and pores. It _is_ very nice.
Yet you don't watch TV to admire the resolution. You watch it for the story and/or the imagery. Most often you forget about the resolution or lack thereof. How often have you watched TV through interference? Your mind fills in much of the missing resolution anyways.
Success of this scheme will depend on whether the mass of consumers is wants and is willing to pay for higher resolution TV. Evidence so far is no. HDTV is not selling well.
NTSC/PAL may well suck, but its enough to get the story across and that's what most people care for. The impact of special-effects and sweeping vistas is mostly related to image viewer angle, not image resolution. Big screen TVs (which are popular) take care of this.
On the question of tape versus disk, the big advantage of disk is random access. It's of little value in video entertainment which is mostly watched serially with very few jumps.
It really isn't much. The Supremes just declined to hear the case. They decline cases all the time because of the large number that reach them.
More importantly, the Virginia case was brought by state employees, not state citizens. Certainly governments can control their employees with means they cannot use on the general population. They give some pay raises, others not.
It would be more interesting if this were a case of citizens who wanted to surf/post pr0n at their local library. In my mind, that would be a clearer case of 1st Amendment free-speech rights.
Don't get too hung up in the computer biz. Technology is useful for more than just pushing information around. And not all information needs to be treated by technology.
It doesn't take technology to decide who sees what information (one of the most important decisions).
Or think of a biotech company: There's alot of technology on the bio-side.
We clearly disagree. I take your vitriol as a sign you feel insecure in your position.
You should understand my application: modelling multicomponent sequential chemical reactions and predicting the yields. NNs do very poorly at this. The diffeq's work great because we do have some fundamental understanding of the underlying elementary processes.
With this, we can extrapolate with surprising success. And interpolate, both with surprisingly few parameters (a few dozen for 150 components). I was glad to see your admission the NNs cannot extrapolate. Extrapolation is very important to us, and we do it well! Dumber number crunching methods may well be incapable, so we should avoid them.
You may be right when you have no idea about the relationships in the data, and just want the NN to "work". But a big correlation is not my idea of "smart".
When I fit a curve to some data, it's because I have a certain fundamental understanding about that data and how it should behave. Only an idiot would try to fit a parabola to an Arrhenius curve.
AFAIK, no NN gives this insight. It is therefore "dumb" in both senses of the word.
I work with statisticians and modellers, and they have an extremely low opinion of neural nets.
As described to me, neural nets are _huge_ (every datapoint is in) underconstrained matricies with an infinite number of equally valid solutions. "Training" [programming] them is an exercise to find a strategy for the "best" solution.
Practically, when NNs are well done they will give you back the data you fed into them! When exceptionally good, they will give reasonable interpolations on the data. But forget about correct extrapolations.
AFAIK it is the only distro that uses a BSD-style/etc/rc.d rather than the SysV-style mess'o'symlinks. Easy to reconfigure with `vi`, and very similar to Slowlaris and FreeBSD boxen. No need for linuxconf.
A very nice historical piece, but hardly very controversial. All the items mentioned were good or exceptional in their day, but for one reason or other were superceded.
One excellent attribute does NOT make a successful product. It takes a combination, and technical excellence is only one of them. Cost matters a great deal, too!
An interesting an important article. One key point is to invalidate keyword filtering by massive redundant use of encryption.
So when is SlashDot going to do it's bit and make everything HTTPS? Almost all browsers have it, and it's a simple and transparent way to increase encrypted traffic to nullify keyword filters.
How much of this alleged increased productivity was simply due to the Hawthorne Effect?
Researchers many years ago at a GE plant in Hawthorne, England wanted to demonstrate the effect of improved lighting. So they increased lighting levels, and lo, productivity went up.
The problem came during the check-back when they lowered lighting levels to the original lux. Productivity went up even further!
It turns out the Heizenberg's uncertainty principle applies to people as well: If you measure and watch something, people react to the closer attention.
Oh, boo hoo hoo: Big Bad Blue has stolen BSDs partition number. That's really ignorant of them, and they deserve to be punished by losing sales. But for anyone who want to develop a fix, here are some ideas:
1) Patch the fsck'ing BIOS. [Dangerous on a laptop]
2) Modify the bootloader so that it re-writes the partition table on every boot using some other type. Writeback at shutdown.
3) recompile your BSD kernel to use a different filesystem type. Or really use a different type for the filesystems [but then you look one of the biggest reasons for running BSD--softupdates].
This can be gotten around. IBM has just done something stupid. They probably ought to have stolen the XENIX fs number instead!
A simple case: Computer programs that are written often have mistakes in them [bugs]. To find and correct these mistakes, programmers use special powerful programs called "debuggers". But these same debuggers are one of the main tools used to break encryption.
So when the DMCA outlaws tools to circumvent encryption, does it then require all software to be buggy?
I agree with the posting that Linux code is unlikely to help MS-Windows. But I must have another thing coming, because I think many of the MS-Windows instabilities are due to flawed design decisions. Not just code bugs.
The Registry and it's access methods, for one. Including foreign drivers into the NT4 kernel for another. FAT in general. Idle busyloop in 95|98|ME [how the EPA let that one slip, I'll never know]. Low tick frequency.
Knowing these design flaws helps to run MS products stably. I run MS-Win95b too much, but haven't had a crash in 6 months. It does take alot of maintenance to keep them this way, though. My wife won't let me at her box until it crashes too often. Then I'll sweat for a couple of hours [no reinstall] and it'll be good for a couple of weeks. [double entendre intended]
First, I will pick a nit: Grandma's ring has great utility to her and to Grandpa who bought it for her. Utility is not exclusively a mechanical practicality measure. It is a subjective human measure.
Your scarity point is interesting, but all software is fundamentally equally scarce or available because it's all under copyright. The licence terms may be generous (GPL) or restrictive (MS-EULA) but it's all monopolies.
More to the point, market price is controlled by supply versus demand. MS controls the supply of it's products tightly (to those who will pay it's price or do other things), free software forbids control of supply. So MS sells for $$$, while free software is nearly costless.
This doesn't stop some unclueful tax offical from deciding that free software performs the same function and has the same intrinsic value as M$,
and hence ought to pay the same tax. He rules that the $2 charge for CheapBytes CDs is just an attempt at tax evasion. The taxpayer can then dispute these charges, and hopefully someone more clueful will decide that the price was real, not an attempt at tax evasion.
I _wish_ govt's could only tax when you earn or spend money. But property ownership taxes are everywhere. Almost always on real-estate, often on chattels [moveables] such as cars and other property.
Value is indeed a tricky proposition, but tax officals are very likely to value things higher to collect more taxes. If you pay $10 for Linux while MS-Windows costs $100, a tax offical might well view the $10 as a fraudulently low delcaration.
Well, I don't know anything about Polish Law. But I can see how this might happen: where there are VAT or other taxes on property/transfers, then some sort of fair market value must be used for the taxation. A common method of tax evasion is undervaluing the property.
The Polish tax official obviously thought that Linux and StarOffice were at least as good as MS-Windows and Office. As you going to dispute his good judgement? So they were worth as much, and should be taxed as much.
This is not an easy dilemma for free software to solve. The concept of something good for nothing sounds impossible to many people. Tax officials read that impossibility as tax evasion.
Slashdot Mirror
This is an excellent polemic. I would add it's only part of a larger phenomenon: Concentrated interests dominate diffuse interests. Irrespective of merit, right, numbers, or aggregate welfare.
Media companies represent one concentrated interest. They exist mostly as a middleman between the artist/creator and their public. Their entire livelihood is threatened by new communications/electronic technology. (They mustn't be adding value anywhere else to feel so threatened!) So they fight tooth-and-nail in the Courts and Legislatures, spending money lavishly. No surprise.
Consumers are the diffuse interest. Sure, they like the music/TV/whatever, but they won't die without it. What most are willing to do to protect their rights is very limited. Even if they are very numerous, their interest is difficult to concentrate. So they lose.
Ultimately, this is a very bad thing for society. It harkens a return to feulalism.
An interesting German ruling, and the consequent market response.
M$ has insisted long and hard in US Courts that it is _licencing_ it's products thru shrinkwraps hidden inside. For OEMs and big corps, it sure does get signed licence agreements. But at retail, it behaves as if it were simply selling copyrighted works (books, music, videos).
M$ doesn't do even simple practical things like have purchasers call for a key, insist on registration, or tear-off registration to protect their licencing status. Apparently their marketing department has vetoed these things as expensive or frightening to customers.
So why should the Courts grant them licencing status when M$ has not done what they could for that status?
I wrote an ASM `life` 15 years ago with DEBUG. On a 4.77 MHz 8088 it ran 80x25 at about 1.5 fps. And it was very primitive: read the display memory and calc'd all the cells.
:)
Sure, I could have used a regen buffer and skipped the empty cells. But then it would run like it does on my 500 MHz machine -- too fast, just a blur. Lame excuse
What does Honeypot want? Cheap forensic analysis on a cracked box?
Well if you want to try, have a read of the Nov & Dec Dr.Dobbs. It has a pair of articles about recovering deleted data and has pointers to useful tools.
The security comparison URL is here. Caveat lector
/etc/passwd, permissions, crontabs and the other configurations it makes or simply not installing some questionable pgms by default.
Agreed that most of the vulnerabilities are associated with the programs themselves. Newer versions fix old holes, but may make new ones.
The distro can influence the final result by
I too am a Slackware afficionado, but I wouldn't call Slackware the "best" without qualifying it. Nor would I call RedHat or Debian or ... the unqualified best.
/etc/rc.d rather than the mess'o'symlinks SysV-style. That makes administration much easier to learn without depending on a tool like `linuxconf`.
I certainly wouldn't call Slackware the best distro for newbies accustomed to MS-Windows and uninterested in learning the guts of Unix. That would be RedHat (or maybe Corel).
But Slackware has some unique features that probably make it the best distro for someone coming from BSD (or SunOS) or who wants to learn the guts of Unix.
Slackware runs a little behind the times in terms of program updates, but that also means that it has the fewest security holes (URL forgotten).
AFAIK, it is the only distro with a BSD-style
Interesting -- a mechanical computer. Presumably analog. Well they had to use something, and it was probably good to a minute-of-arc. I'd be interested in hearing more about the operating principles (tuned pendulum on a gyro?).
But they're going to need something better when the USN strips the rifling out of those bores and switchs to saboted submunitions with 50+ mile range. Or maybe the shot'll be terminally guided if it's not too long for the loading cradle inside that cramped turret.
Aircraft and ships are constantly being tossed around by the fluid through which they travel.
Fixed-trajectory munitions will miss unless the roll, pitch and yaw are compensated for. Hitting a 10 ft target at 100 miles isn't easy.
What will probably happen is a target is locked into the fire control computer, the operator presses the fire button, and the computer waits a few milliseconds until the weapon is on target (after compensating for aiming offset, refractive index gradients, target movement, etc).
I'm sure the big 16" guns on US Battleships have very sophisticated firing [delay] systems. Just so they can hit the broadside of a barn at 15 miles. Otherwise, they could miss the broadside of a mountain!
Yes, I've seen HDTV from digital source. My photographers mind boggles at the improved details. You can see individual hair and pores. It _is_ very nice.
Yet you don't watch TV to admire the resolution. You watch it for the story and/or the imagery. Most often you forget about the resolution or lack thereof. How often have you watched TV through interference? Your mind fills in much of the missing resolution anyways.
Success of this scheme will depend on whether the mass of consumers is wants and is willing to pay for higher resolution TV. Evidence so far is no. HDTV is not selling well.
NTSC/PAL may well suck, but its enough to get the story across and that's what most people care for. The impact of special-effects and sweeping vistas is mostly related to image viewer angle, not image resolution. Big screen TVs (which are popular) take care of this.
On the question of tape versus disk, the big advantage of disk is random access. It's of little value in video entertainment which is mostly watched serially with very few jumps.
It really isn't much. The Supremes just declined to hear the case. They decline cases all the time because of the large number that reach them.
More importantly, the Virginia case was brought by state employees, not state citizens. Certainly governments can control their employees with means they cannot use on the general population. They give some pay raises, others not.
It would be more interesting if this were a case of citizens who wanted to surf/post pr0n at their local library. In my mind, that would be a clearer case of 1st Amendment free-speech rights.
Don't get too hung up in the computer biz. Technology is useful for more than just pushing information around. And not all information needs to be treated by technology.
It doesn't take technology to decide who sees what information (one of the most important decisions).
Or think of a biotech company: There's alot of technology on the bio-side.
We clearly disagree. I take your vitriol as a sign you feel insecure in your position.
You should understand my application: modelling multicomponent sequential chemical reactions and predicting the yields. NNs do very poorly at this. The diffeq's work great because we do have some fundamental understanding of the underlying elementary processes.
With this, we can extrapolate with surprising success. And interpolate, both with surprisingly few parameters (a few dozen for 150 components). I was glad to see your admission the NNs cannot extrapolate. Extrapolation is very important to us, and we do it well! Dumber number crunching methods may well be incapable, so we should avoid them.
You may be right when you have no idea about the relationships in the data, and just want the NN to "work". But a big correlation is not my idea of "smart".
When I fit a curve to some data, it's because I have a certain fundamental understanding about that data and how it should behave. Only an idiot would try to fit a parabola to an Arrhenius curve.
AFAIK, no NN gives this insight. It is therefore "dumb" in both senses of the word.
I work with statisticians and modellers, and they have an extremely low opinion of neural nets.
As described to me, neural nets are _huge_ (every datapoint is in) underconstrained matricies with an infinite number of equally valid solutions. "Training" [programming] them is an exercise to find a strategy for the "best" solution.
Practically, when NNs are well done they will give you back the data you fed into them! When exceptionally good, they will give reasonable interpolations on the data. But forget about correct extrapolations.
I like Slackware. There, I admitted it :)
/etc/rc.d rather than the SysV-style mess'o'symlinks. Easy to reconfigure with `vi`, and very similar to Slowlaris and FreeBSD boxen. No need for linuxconf.
AFAIK it is the only distro that uses a BSD-style
A very nice historical piece, but hardly very controversial. All the items mentioned were good or exceptional in their day, but for one reason or other were superceded.
One excellent attribute does NOT make a successful product. It takes a combination, and technical excellence is only one of them. Cost matters a great deal, too!
An interesting an important article. One key point is to invalidate keyword filtering by massive redundant use of encryption.
So when is SlashDot going to do it's bit and make everything HTTPS? Almost all browsers have it, and it's a simple and transparent way to increase encrypted traffic to nullify keyword filters.
How much of this alleged increased productivity was simply due to the Hawthorne Effect?
Researchers many years ago at a GE plant in Hawthorne, England wanted to demonstrate the effect of improved lighting. So they increased lighting levels, and lo, productivity went up.
The problem came during the check-back when they lowered lighting levels to the original lux. Productivity went up even further!
It turns out the Heizenberg's uncertainty principle applies to people as well: If you measure and watch something, people react to the closer attention.
Oh, boo hoo hoo: Big Bad Blue has stolen BSDs partition number. That's really ignorant of them, and they deserve to be punished by losing sales. But for anyone who want to develop a fix, here are some ideas:
1) Patch the fsck'ing BIOS. [Dangerous on a laptop]
2) Modify the bootloader so that it re-writes the partition table on every boot using some other type. Writeback at shutdown.
3) recompile your BSD kernel to use a different filesystem type. Or really use a different type for the filesystems [but then you look one of the biggest reasons for running BSD--softupdates].
This can be gotten around. IBM has just done something stupid. They probably ought to have stolen the XENIX fs number instead!
A simple case: Computer programs that are written often have mistakes in them [bugs]. To find and correct these mistakes, programmers use special powerful programs called "debuggers". But these same debuggers are one of the main tools used to break encryption.
So when the DMCA outlaws tools to circumvent encryption, does it then require all software to be buggy?
I agree with the posting that Linux code is unlikely to help MS-Windows. But I must have another thing coming, because I think many of the MS-Windows instabilities are due to flawed design decisions. Not just code bugs.
The Registry and it's access methods, for one. Including foreign drivers into the NT4 kernel for another. FAT in general. Idle busyloop in 95|98|ME [how the EPA let that one slip, I'll never know]. Low tick frequency.
Knowing these design flaws helps to run MS products stably. I run MS-Win95b too much, but haven't had a crash in 6 months. It does take alot of maintenance to keep them this way, though. My wife won't let me at her box until it crashes too often. Then I'll sweat for a couple of hours [no reinstall] and it'll be good for a couple of weeks. [double entendre intended]
First, I will pick a nit: Grandma's ring has great utility to her and to Grandpa who bought it for her. Utility is not exclusively a mechanical practicality measure. It is a subjective human measure.
Your scarity point is interesting, but all software is fundamentally equally scarce or available because it's all under copyright. The licence terms may be generous (GPL) or restrictive (MS-EULA) but it's all monopolies.
More to the point, market price is controlled by supply versus demand. MS controls the supply of it's products tightly (to those who will pay it's price or do other things), free software forbids control of supply. So MS sells for $$$, while free software is nearly costless.
This doesn't stop some unclueful tax offical from deciding that free software performs the same function and has the same intrinsic value as M$,
and hence ought to pay the same tax. He rules that the $2 charge for CheapBytes CDs is just an attempt at tax evasion. The taxpayer can then dispute these charges, and hopefully someone more clueful will decide that the price was real, not an attempt at tax evasion.
I _wish_ govt's could only tax when you earn or spend money. But property ownership taxes are everywhere. Almost always on real-estate, often on chattels [moveables] such as cars and other property.
Value is indeed a tricky proposition, but tax officals are very likely to value things higher to collect more taxes. If you pay $10 for Linux while MS-Windows costs $100, a tax offical might well view the $10 as a fraudulently low delcaration.
Well, I don't know anything about Polish Law. But I can see how this might happen: where there are VAT or other taxes on property/transfers, then some sort of fair market value must be used for the taxation. A common method of tax evasion is undervaluing the property.
The Polish tax official obviously thought that Linux and StarOffice were at least as good as MS-Windows and Office. As you going to dispute his good judgement? So they were worth as much, and should be taxed as much.
This is not an easy dilemma for free software to solve. The concept of something good for nothing sounds impossible to many people. Tax officials read that impossibility as tax evasion.