Now that ISIS has stopped blocking the sites under
massive objection from free speech advocates, the
local government has released a press statement
in which they claim that ISIS gave in to racist pressure. Guess we're all nazis now because we
didn't want to allow our government to take the
easy route to complete content control.
The statement is here: Pressemitteilung 467/2001 der Bezirksregierung Düsseldorf vom 22.11.2001
The only problem I have now are legitimate mailing lists, like the PHP lists, which archive stuff to the web without obscuring addresses similarly.
This, asshole-ACs who post mailto links, "friends" who send electronic greeting cards and using email adresses in circumstances which don't allow for obfuscation are reasons why the problem needs to be solved on a different level. It is absolutely necessary that the simple knowledge of your email adress is worthless to spammers because that knowledge alone is either not enough to send you mail or because it is too costly to send unsolicited (or fake solicited) mail to you. These are the only attack vectors which work.
ISO8601 defines an international standard format for date/time which avoids some of the issues of other formats, like wrong sort order and ambiguity. It is summarized here: http://www.cl.cam.ac.uk/~mgk25/iso-time.html
You know, I would really LOVE to see someone do this. Even if it's only a short clip as a technology demonstration. I have these Monsters Inc. shots at 1536x922 resolution and they look just fantastic. Please, someone with access to high resolution film material, give us something to fill a 21" monitor at 1600x1200 resolution with unscaled video. I bet that seeing this ONCE would really make people want new tech...
Digital storage is not perfect. The medium on which digital data is stored decays, thus forcing us to copy the data over to a fresh medium. Most people think that this practice ensures perfect preservation, but due to several factors, this isn't the case. Storage and transmission systems for digital information all have an associated bit error rate, which is the number of bits expected to deviate from the intended value per number of total bits stored/transmitted. While most systems are designed to keep the bit error rate extremely low by using error correction codes (aka redundancy), that rate is never 0. That means that many copies will be perfect, but some are not and you can not detect such errors with absolute certainty. That wouldn't be much of a problem if we only needed to deal with errors caused by storage systems and processing units which are in good shape and thus exhibit only the specified error rates. Data passes many flawed systems and some of these flaws are discovered post mortem (for valuable data). The VIA chipset bug comes to mind, if you want a home user example. Also anyone who has ever experienced a "corrupted download" should be well aware of the imperfections of digital data handling. On a more professional scale, data archivers have started to realize that CD-ROMs won't make it to the estimated lifetime and therefore need to be copied much more often than expected. Some of the data is already lost, some will be lost because there isn't enough copy capacity (equipment and personnel) to handle the increased turnover rate. Storage space problems become bandwith problems. Home users don't realize this, because they very rarely have more than 1000 media per drive. But even then you'd have to start one year early if you do three copies every day of the year in order to safely transfer all data to fresh media. It is the high turnover rate compared to paper and microfilm which kicks digital media out of the long term conservation race. A faded document is better than the perfect digital copy which doesn't exist because it wasn't made for cost reasons.
Most people wouldn't even notice. The commercial part of the internet will always work hard to be well-connected to other countries. Normal people don't browse the net for linux kernel changelogs, decss or pdf-cracks. They buy linux distributions, DVDs and books, and will continue to do so on the partitioned net you propose. Business will adapt to the changing legal climate and these changes will be too slow to notice for most, until it is to hot to get out. Trying to make the internet snap is futile. The people who are behind the DMCA control what normal people perceive as "the internet". Only the "free" part of the internet would be affected in a radical way and that is exactly what RIAA, MPAA et al want.
I want one. Problem is, I am in Germany and TigerDirect will only take international orders by phone and they are very vague about additional handling and shipping charges, customs and taxes. Does anyone have advice about these topics, especially taxes and customs? I don't want to end up paying twice the price of the product in additional charges, just to get it to this side of the pond. That would kind of ruin the price/performance of the deal. Are there online services which handle international shippings for a reasonable fee and provide the knowhow upfront? Damn, I really want one.
More and more providers seem to like being a nanny. Port blockades, proxies, on-server mailviruscheckers, etc. wouldn't be so annoying if providers offered an easy and reliable way to turn these features off. There is a chance to circumvent the restrictions though: Make your traffic opaque, encrypt it with IPSec. After that, all that is left of your connections are target IP addresses and seemingly random data. No more ports and no more chances to interfere with your data through proxies. The provider can only allow all your connections or none.
The first day of the robot contest is over and Heise Online has posted a
report
on some of the most memorable events. Here's the translation of the linked page:
Today was the first day of the c't robot challenge at the Systems in Munich
and a rather simple construction based on Lego Mindstorms made for a genuine
surprise: The roughly soccer ball sized robot, which isn't equipped with
any kind of camera sensors und is thus called "Stevie Wonder" by its
creators, completed the "cleaning bot" challenge in only 6 minutes
and 24 seconds on its second attempt. On the first attempt it had
erroneously moved one of the garbage objects out of the target area and on
the third attempt it drove onto a wooden garbage object of the "cigarette
box" type, making the tracks lose ground contact. According to the contest
rules, the best - the second - attempt counts. "Stevie's" success was
above all astounding because, though being a relatively simple robot
compared to other constructions in this contest, it unerringly moved the
red cans and wooden cuboids, which served as garbage objects, into the blue
"recycling area". The construction team around Wolfgang Lang from Roetenbach
first anticipated problems connected to the cigarette box sized wooden
objects: The "Stevie" creators had expected real cigarette boxes and had
therefore built the robot with that kind of less sturdy lightweight objects
in mind. Then however, two hair ties, which they borrowed from a Heise
employee, were successfully used to tune "Stevie's" mechanics.
This day's other candidates had to face more or less serious
disappointments: Wolfgang Draxinger's Fishertechnik Computing based
"Mr. Proper" suffered from blasted power transistors, which forced
the student to do without the intelligently designed gathering
mechanics of his robot and instead redesign concept and software to
use only pushing. Since some contestants didn't show up, the referees
could grant him some time for reprogramming, but in the end, "Mr. Propper"
wouldn't move at all, or required interactive commands respectively,
so Draxinger had to give up with a heavy heart.
Team Metavox from Meerbusch had some more success: Their swivel flap
equipped "Metabot", which is based on a self-made platform, showed a
heart-warming tendency to hug walls, but at least mastered part of the
challenge. In the first two rounds, it didn't manage to leave the
starting room. In order to create a secondary ranking for those robots,
which couldn't find the target area on their own, the referees allowed
the camera equipped "Metabot" to start right in the doorway to the room
with the target area, after which it moved five of six garbage objects into
the recycling area after all. In order to find the direction of choice, the
robot turns in small steps and tries to estimate distances. Then it rushes
in the direction with the most free space at rather high speed. In doing so
it hit walls and corners several times. The design team around Christian
Winkgen shrugged off the orientation weaknesses of the rolling speed
merchant: "Metabot" had been built for fun, and maybe it simply had a
hard time correctly telling the colors of the contest arena apart.
The contest continues. Right now, "Stevie" is not only the audience's
favorite, it is also on top of the current rankings. But nothing is
decided yet: If all contestants who have been assigned a timeslot
show up, there will be 33 more performances. Today three of the
candidates where absent without notice, so it can't be predicted what will happen in the next days, but the contest will
certainly stay exciting.
The 5-4-3 rule is for unswitched ethernet. That rule treats a switch as an endpoint. And you don't have to worry about collisions either, because you'll probably go for full duplex. Of course fiber is still preferable because you get longer range without active components.
Congratulations, Bin Laden.
We ourselves destroy what is our most precious good: our freedom. It was based on trust and you destroyed not only the World Trade Center and the lives of thousands, with the help of so many journalists and politicians you also destroyed that trust which made our societies free. Now that everyone has become a suspect, when will Stasi personal be in the high ranks again, or are they already? Elected politicians complete what you started, and you made them do it.
The guide is for public LAN-parties, where some people may need guidance because it's their first time to such an event. For public events, a certain level of planning is necessary to avoid problems which just don't need to be considered when all attendants know eachother and the location and have met for a LAN before.
Besides, switches are cheap, the added cost compared to hubs is negligible. Switches avoid problems with network cards which are set to full duplex, perform much better in mixed 10mbit/100mbit situations and provide better overall bandwith. If new hardware needs to be bought anyway, don't go for hubs. If you got the hubs already, keep them.
It is worth noting though that other LAN guides recommend against DHCP because conflicting DHCP servers are hard to find unless you have manageable network equipment. And the more participants there are, the more likely someone will have forgotten that he's running such a server.
If I am misled, shown useless, false, or irrelevant content, then I will go away to a competing website, and likely never return.
Why would a content provider drive away its potential customers by acting like a jackass?
So what, if all the ad-blocking jackasses leave for the competitor's site? They'll cost the competitor's money for bandwith and give him nothing in return. A content provider who wants to reach as many people as possible, no matter what the costs are, wouldn't think about implementing ad-blocker-detection software in the first place.
The content provider might show reduced content to ad-blocking websurfers. At first, it looks like a stupid idea to present yourself in a worse way than you are capable of, but think about it: The content is what people want. Not giving the content away for free is what the content provider wants. By not delivering the full product to ad-blockers, these websurfers miss out on what they want, while the content provider still gives an overview of the content. This isn't a new scheme at all: Some games render themselves more and more useless the further you get into the game when they detect a copied cd. Some applications are known to produce false data if they reliably detect a copyright violation. This sort of selective product quality takes reliability away, and that is what many people will not accept "just to remove the ads".
The "one programmer, many users" assumption works only if the evolutionary steps don't occur too often and people have a fast and reliable way of detecting when it's time to upgrade. The arms race has just started and I think there are countless options for both sides, requiring considerable effort on the user side due to the necessary frequent upgrades. See "Punkbuster" for a similar situation. That, combined with the uncertainty that you might not be seeing what those see who don't block the ads, makes "normal" people shy away from joining the arms race. If you're religious about avoiding banner ads, that's a completely different story.
While the interested geeks develop anti-content-blocker schemes, MediaBeam realized what it's really about: The masses want "free" content. The effort, which is necessary to fool the webserver into thinking that ads are shown while in reality they are cleverly redirected to the bitbucket, is bigger than just downloading and suffering through them. Is websurfing really fun when you always have to question wether that page you see is really the one you wanted? Or is it just the heavily abbreviated version for ad-blockers? Why aren't people complaining about this slow server? Maybe it's fast for them cause they don't block the ads? Nah, that can't be it. My ad-blocker behaves just like the real browser, it just doesn't show the ads. Or are they using new methods of detection and my blocker doesn't fool them anymore? Guess I'll have to upgrade. Again.
Fileplanet was well known. It was good. Fileplanet changed. It became less "fast downloads, no hassle" and more "so so downloads, big hassle". People stop using Fileplanet. Some may have problems with letting go and try all sorts of stunts to get to their fileplanet files without giving their mail-address, but that is a temporary effect. Over time, people will (hopefully) realize that there are more efficient ways of data distribution than individual downloads from a few central servers.
Re:What's up with fileplanet requiring a login?...
on
Quake3 v1.30 Final Is Out
·
· Score: 3, Informative
The number of FilePlanet mirrors decreases, they are not faster than other sites (on the contrary, at least not for me), the pop-ups drive me mad, their download-queue scheme prevents me from comfortably using a download manager - in other words: they suck. Don't go there anymore.
Anyone mentioning the terrorist attacks on the World Trade Center, the Pentagon or the fourth crashed plane in an attempt to justify a change of law is not acting in the tradition of a free country. Using the terrorist attacks to finally get what Big Brothers always wanted should anger every free citizen.
And they can kiss that extra revenue goodbye when IPSec or IPv6 come into widespread use. No more ports, just opaque traffic. Of course, with things being like they are right now, IPSec is certain to become an outlaw's tool...
You're right, the ISP undersubscribes on bandwith. And yes, prices would be higher if they didn't do that. But no, dropping someone's packets to reduce the bandwith bill for the provider is not acceptable unless it's legitimized by the terms of service. Estimating the average bandwith requirements and deciding how much reserve you put on top of that is part of the ISP's job. If that estimate is proven wrong by a sudden increase in bandwith requirements due to worm proliferation or a new bandwith eating killer application, then that is the ISP's problem, not the customer's. Blocking ports of uninfected machines is an unacceptable measure. On the other hand, cutting off crackers in action (read: infected machines) is most likely covered by the ToS.
the bandwidth belongs to the ISP. They have to protect it.
Actually, no. The bandwith belongs to those who pay for it and that is the customer. Internet providers really have no business keeping packets of the net to save bandwith. They do however have the right to stop crackers and spammers if that is in their terms of service and I bet it is. If the service provider is nice, they can also try to protect their customers from crackers but as long as the actions are not covered by ToS, they should be prepared to stop nannying.
Slashdot Mirror
Now that ISIS has stopped blocking the sites under massive objection from free speech advocates, the local government has released a press statement in which they claim that ISIS gave in to racist pressure. Guess we're all nazis now because we didn't want to allow our government to take the easy route to complete content control.
The statement is here: Pressemitteilung 467/2001 der Bezirksregierung Düsseldorf vom 22.11.2001
The only problem I have now are legitimate mailing lists, like the PHP lists, which archive stuff to the web without obscuring addresses similarly.
This, asshole-ACs who post mailto links, "friends" who send electronic greeting cards and using email adresses in circumstances which don't allow for obfuscation are reasons why the problem needs to be solved on a different level. It is absolutely necessary that the simple knowledge of your email adress is worthless to spammers because that knowledge alone is either not enough to send you mail or because it is too costly to send unsolicited (or fake solicited) mail to you. These are the only attack vectors which work.
ISO8601 defines an international standard format for date/time which avoids some of the issues of other formats, like wrong sort order and ambiguity. It is summarized here: http://www.cl.cam.ac.uk/~mgk25/iso-time.html
You know, I would really LOVE to see someone do this. Even if it's only a short clip as a technology demonstration. I have these Monsters Inc. shots at 1536x922 resolution and they look just fantastic. Please, someone with access to high resolution film material, give us something to fill a 21" monitor at 1600x1200 resolution with unscaled video. I bet that seeing this ONCE would really make people want new tech...
Digital storage is not perfect. The medium on which digital data is stored decays, thus forcing us to copy the data over to a fresh medium. Most people think that this practice ensures perfect preservation, but due to several factors, this isn't the case. Storage and transmission systems for digital information all have an associated bit error rate, which is the number of bits expected to deviate from the intended value per number of total bits stored/transmitted. While most systems are designed to keep the bit error rate extremely low by using error correction codes (aka redundancy), that rate is never 0. That means that many copies will be perfect, but some are not and you can not detect such errors with absolute certainty. That wouldn't be much of a problem if we only needed to deal with errors caused by storage systems and processing units which are in good shape and thus exhibit only the specified error rates. Data passes many flawed systems and some of these flaws are discovered post mortem (for valuable data). The VIA chipset bug comes to mind, if you want a home user example. Also anyone who has ever experienced a "corrupted download" should be well aware of the imperfections of digital data handling. On a more professional scale, data archivers have started to realize that CD-ROMs won't make it to the estimated lifetime and therefore need to be copied much more often than expected. Some of the data is already lost, some will be lost because there isn't enough copy capacity (equipment and personnel) to handle the increased turnover rate. Storage space problems become bandwith problems. Home users don't realize this, because they very rarely have more than 1000 media per drive. But even then you'd have to start one year early if you do three copies every day of the year in order to safely transfer all data to fresh media. It is the high turnover rate compared to paper and microfilm which kicks digital media out of the long term conservation race. A faded document is better than the perfect digital copy which doesn't exist because it wasn't made for cost reasons.
Most people wouldn't even notice. The commercial part of the internet will always work hard to be well-connected to other countries. Normal people don't browse the net for linux kernel changelogs, decss or pdf-cracks. They buy linux distributions, DVDs and books, and will continue to do so on the partitioned net you propose. Business will adapt to the changing legal climate and these changes will be too slow to notice for most, until it is to hot to get out. Trying to make the internet snap is futile. The people who are behind the DMCA control what normal people perceive as "the internet". Only the "free" part of the internet would be affected in a radical way and that is exactly what RIAA, MPAA et al want.
I want one. Problem is, I am in Germany and TigerDirect will only take international orders by phone and they are very vague about additional handling and shipping charges, customs and taxes. Does anyone have advice about these topics, especially taxes and customs? I don't want to end up paying twice the price of the product in additional charges, just to get it to this side of the pond. That would kind of ruin the price/performance of the deal. Are there online services which handle international shippings for a reasonable fee and provide the knowhow upfront? Damn, I really want one.
More and more providers seem to like being a nanny. Port blockades, proxies, on-server mailviruscheckers, etc. wouldn't be so annoying if providers offered an easy and reliable way to turn these features off. There is a chance to circumvent the restrictions though: Make your traffic opaque, encrypt it with IPSec. After that, all that is left of your connections are target IP addresses and seemingly random data. No more ports and no more chances to interfere with your data through proxies. The provider can only allow all your connections or none.
The first day of the robot contest is over and Heise Online has posted a report on some of the most memorable events. Here's the translation of the linked page:
Today was the first day of the c't robot challenge at the Systems in Munich and a rather simple construction based on Lego Mindstorms made for a genuine surprise: The roughly soccer ball sized robot, which isn't equipped with any kind of camera sensors und is thus called "Stevie Wonder" by its creators, completed the "cleaning bot" challenge in only 6 minutes and 24 seconds on its second attempt. On the first attempt it had erroneously moved one of the garbage objects out of the target area and on the third attempt it drove onto a wooden garbage object of the "cigarette box" type, making the tracks lose ground contact. According to the contest rules, the best - the second - attempt counts. "Stevie's" success was above all astounding because, though being a relatively simple robot compared to other constructions in this contest, it unerringly moved the red cans and wooden cuboids, which served as garbage objects, into the blue "recycling area". The construction team around Wolfgang Lang from Roetenbach first anticipated problems connected to the cigarette box sized wooden objects: The "Stevie" creators had expected real cigarette boxes and had therefore built the robot with that kind of less sturdy lightweight objects in mind. Then however, two hair ties, which they borrowed from a Heise employee, were successfully used to tune "Stevie's" mechanics.
This day's other candidates had to face more or less serious disappointments: Wolfgang Draxinger's Fishertechnik Computing based "Mr. Proper" suffered from blasted power transistors, which forced the student to do without the intelligently designed gathering mechanics of his robot and instead redesign concept and software to use only pushing. Since some contestants didn't show up, the referees could grant him some time for reprogramming, but in the end, "Mr. Propper" wouldn't move at all, or required interactive commands respectively, so Draxinger had to give up with a heavy heart.
Team Metavox from Meerbusch had some more success: Their swivel flap equipped "Metabot", which is based on a self-made platform, showed a heart-warming tendency to hug walls, but at least mastered part of the challenge. In the first two rounds, it didn't manage to leave the starting room. In order to create a secondary ranking for those robots, which couldn't find the target area on their own, the referees allowed the camera equipped "Metabot" to start right in the doorway to the room with the target area, after which it moved five of six garbage objects into the recycling area after all. In order to find the direction of choice, the robot turns in small steps and tries to estimate distances. Then it rushes in the direction with the most free space at rather high speed. In doing so it hit walls and corners several times. The design team around Christian Winkgen shrugged off the orientation weaknesses of the rolling speed merchant: "Metabot" had been built for fun, and maybe it simply had a hard time correctly telling the colors of the contest arena apart.
The contest continues. Right now, "Stevie" is not only the audience's favorite, it is also on top of the current rankings. But nothing is decided yet: If all contestants who have been assigned a timeslot show up, there will be 33 more performances. Today three of the candidates where absent without notice, so it can't be predicted what will happen in the next days, but the contest will certainly stay exciting.
The 5-4-3 rule is for unswitched ethernet. That rule treats a switch as an endpoint. And you don't have to worry about collisions either, because you'll probably go for full duplex. Of course fiber is still preferable because you get longer range without active components.
Congratulations, Bin Laden.
We ourselves destroy what is our most precious good: our freedom. It was based on trust and you destroyed not only the World Trade Center and the lives of thousands, with the help of so many journalists and politicians you also destroyed that trust which made our societies free. Now that everyone has become a suspect, when will Stasi personal be in the high ranks again, or are they already? Elected politicians complete what you started, and you made them do it.
Talking about utopian dreams and inspiring people to take on cool projects: Watch "The Big Dig". One day I'm gonna grab a shovel and start...
The guide is for public LAN-parties, where some people may need guidance because it's their first time to such an event. For public events, a certain level of planning is necessary to avoid problems which just don't need to be considered when all attendants know eachother and the location and have met for a LAN before.
Besides, switches are cheap, the added cost compared to hubs is negligible. Switches avoid problems with network cards which are set to full duplex, perform much better in mixed 10mbit/100mbit situations and provide better overall bandwith. If new hardware needs to be bought anyway, don't go for hubs. If you got the hubs already, keep them.
It is worth noting though that other LAN guides recommend against DHCP because conflicting DHCP servers are hard to find unless you have manageable network equipment. And the more participants there are, the more likely someone will have forgotten that he's running such a server.
If I am misled, shown useless, false, or irrelevant content, then I will go away to a competing website, and likely never return.
Why would a content provider drive away its potential customers by acting like a jackass?
So what, if all the ad-blocking jackasses leave for the competitor's site? They'll cost the competitor's money for bandwith and give him nothing in return. A content provider who wants to reach as many people as possible, no matter what the costs are, wouldn't think about implementing ad-blocker-detection software in the first place.
The content provider might show reduced content to ad-blocking websurfers. At first, it looks like a stupid idea to present yourself in a worse way than you are capable of, but think about it: The content is what people want. Not giving the content away for free is what the content provider wants. By not delivering the full product to ad-blockers, these websurfers miss out on what they want, while the content provider still gives an overview of the content. This isn't a new scheme at all: Some games render themselves more and more useless the further you get into the game when they detect a copied cd. Some applications are known to produce false data if they reliably detect a copyright violation. This sort of selective product quality takes reliability away, and that is what many people will not accept "just to remove the ads".
The "one programmer, many users" assumption works only if the evolutionary steps don't occur too often and people have a fast and reliable way of detecting when it's time to upgrade. The arms race has just started and I think there are countless options for both sides, requiring considerable effort on the user side due to the necessary frequent upgrades. See "Punkbuster" for a similar situation. That, combined with the uncertainty that you might not be seeing what those see who don't block the ads, makes "normal" people shy away from joining the arms race. If you're religious about avoiding banner ads, that's a completely different story.
There's a difference between superstitiousness and banners: You don't have to believe in the effectiveness of banners to be affected by them.
While the interested geeks develop anti-content-blocker schemes, MediaBeam realized what it's really about: The masses want "free" content. The effort, which is necessary to fool the webserver into thinking that ads are shown while in reality they are cleverly redirected to the bitbucket, is bigger than just downloading and suffering through them. Is websurfing really fun when you always have to question wether that page you see is really the one you wanted? Or is it just the heavily abbreviated version for ad-blockers? Why aren't people complaining about this slow server? Maybe it's fast for them cause they don't block the ads? Nah, that can't be it. My ad-blocker behaves just like the real browser, it just doesn't show the ads. Or are they using new methods of detection and my blocker doesn't fool them anymore? Guess I'll have to upgrade. Again.
Fileplanet was well known. It was good. Fileplanet changed. It became less "fast downloads, no hassle" and more "so so downloads, big hassle". People stop using Fileplanet. Some may have problems with letting go and try all sorts of stunts to get to their fileplanet files without giving their mail-address, but that is a temporary effect. Over time, people will (hopefully) realize that there are more efficient ways of data distribution than individual downloads from a few central servers.
The number of FilePlanet mirrors decreases, they are not faster than other sites (on the contrary, at least not for me), the pop-ups drive me mad, their download-queue scheme prevents me from comfortably using a download manager - in other words: they suck. Don't go there anymore.
Anyone mentioning the terrorist attacks on the World Trade Center, the Pentagon or the fourth crashed plane in an attempt to justify a change of law is not acting in the tradition of a free country. Using the terrorist attacks to finally get what Big Brothers always wanted should anger every free citizen.
Nimda uses several attack vectors and not all of them involve an IIS. A machine infected by Nimda isn't necessarily running Win2K Server.
And they can kiss that extra revenue goodbye when IPSec or IPv6 come into widespread use. No more ports, just opaque traffic. Of course, with things being like they are right now, IPSec is certain to become an outlaw's tool...
You're right, the ISP undersubscribes on bandwith. And yes, prices would be higher if they didn't do that. But no, dropping someone's packets to reduce the bandwith bill for the provider is not acceptable unless it's legitimized by the terms of service. Estimating the average bandwith requirements and deciding how much reserve you put on top of that is part of the ISP's job. If that estimate is proven wrong by a sudden increase in bandwith requirements due to worm proliferation or a new bandwith eating killer application, then that is the ISP's problem, not the customer's. Blocking ports of uninfected machines is an unacceptable measure. On the other hand, cutting off crackers in action (read: infected machines) is most likely covered by the ToS.
the bandwidth belongs to the ISP. They have to protect it.
Actually, no. The bandwith belongs to those who pay for it and that is the customer. Internet providers really have no business keeping packets of the net to save bandwith. They do however have the right to stop crackers and spammers if that is in their terms of service and I bet it is. If the service provider is nice, they can also try to protect their customers from crackers but as long as the actions are not covered by ToS, they should be prepared to stop nannying.