That sounds right to me. Maybe I'm a user that signed on to a mailing list because I actually wanted those weekly product listings from them. How does the ISP (which is what the university is in this case) know whether or not the e-mail is wanted or not? I'll filter it myself. *
* - exception - if the ISP (university) wants to charge users for how much e-mail they recieve, then spam blocking does have to be done before the user sees the mail, otherwise he gets charged for everything, solicited or not.
True, but the whole point of this discussion is that even when the traffic hitting you is "syn, udp, echo packets, etc", the ISP still bills you for it the same as if it had been real traffic.
The USA has had the same constitutional government for most of its history. This has the ironic affect that while as a nation the USA is very young, and as a culture the USA is very young, as a continuous unbroken government, the USA is the oldest in the world (Well, maybe the second behind the UK - it's hard to pick a specific date for the start of the UK's current government since it mutated a little at a time to become what it is today, rather than having a specific starting point with a clean break.)
This means the US government has some very archaic constructs not present in the newer constitutions of other countries. One of them is the voting system. Because the country was first founded as a loose joining of individual colonies that thought of themselves as independant countries joining an alliance, the national voting system only dictates how each state has to send its results to the national tally, and how many votes each state gets. Originally each state got to pick for itself how it was going to handle the voting process within its borders. That's still true today. The only thing the federal government mandates is that the voting age must be 18. (And there was still some wriggle-room in the wording of the constitution that allowed some states to add extra rules designed to supress the black voters, until the civil rights movement of the '60s got some amendments added to close those loopholes.) But everything else about the details of the voting process is still up to the state. So that's why the tallying system is not uniform. In some states, the entire state uses a uniform system. In other states, each small voting district gets to do it their own way. That was the case in Florida. The unfairness with this system is that the funding to pay for the vote tallying systems comes from the local taxes even though the vote is fufilling a national rather than a local need. Since voting districts are very small areas, sometimes only a few square miles, this means that poor sections of town tend to have lower quality voting systems with larger margins of error. This is why Gore was so adamant about a recount in Florida - the areas where people are poorer tend to vote more for Democrats, and those are the areas where the greatest possibility for error were. Today it's impossible to tell what the vote actually ended up being, since the data gathering system itself is flawed to the point where the data available, even if it is counted and re-counted, isn't enough.
[...] it might be interesting if, for example, republicans voted more in the morning than in the afternoon and democrats voted more in the afternoon
(Actually, Republican votes DO tend to come in earlier in the day that Democrat votes. I think I know the reason why - mostly having to do with the fact that almost all white-collar jobs have the same normal get-up-in-the-morning schedule, while blue-collar jobs tend to have schedules that vary a lot in their times. Since most voters vote on their way to or from work, the schedule of their work tends to affect when they vote.)
This is also why I am strongly opposed to the television news broadcasting their guesses as to the results (based on exit polls) before the polls are closed. Knowing which candidate is winning can change who you vote for, giving an advantage to the later-in-the-day voter. For example, if you really like candidate A, think B is kind of okay, but really despise C, and you find that candidate A has no hope of winning, but B does have some hope of winning, you might decide to vote for B instead of A just to help oppose C. If you had voted before knowing the results so far you would have been voting for A. (And since this country spans 6 seperate time zones, that ends up meaning that people voting in the west get to find out the results of the polls in the east while their polls are still open, and since this election had signifigant showings for a third-party candidate, I'm sure that this A/B/C example above realy did happen in a lot of cases.)
There will, no doubt, be a small number of Arab people who would not accept any settlement that didn't include the slaying of the Israelis to the last child.
DON'T PUT WORDS IN MY MOUTH BY IMPLICATION. I never said "slayed". I said "removed". I was thinking of deportation, actually.
I don't think a generation washes away guilt.
Niether do I. You can't wash away guilt you never had.
NO, that's precisely how flood denial of service attacks work. The attacker comprimises many other machines first, and used them to attack you from multiple locations. That makes it a pain to firewall the attack away.
Just because the traffic hits you from more than one machine doesn't necessarily mean it's accidental normal traffic.
No. You've got a cell phone that gets billed by the minute. In order to cut down on incoming calls you set up an answering machine, or "firewall" to screen out the junk. Okay, so you ignore the unwanted calls...
NO. That isn't the situation. The situation is that you never pick up the phone when the person is calling from a known blacklist of numbers, never let those numbers switch over to voice-mail, and STILL are billed for the mere attempt to contact you.
Your ISP is selling you access, not protection. It'll probably sell you its own firewalling service, too, and that's the only way you're going to keep from recieving those unwanted packets.
Making it so that you may only avoid charges for OTHER PEOPLE using your line is to buy an extra service from the ISP sounds exactly like protection to me - the Mafia type of protection. "Buy our special protection service, or there's no telling what might just so happen to get billed to your account when strangers like Vinnie and Eddie here flood your computer."
It looks like you don't know that data is still sent to your computer to find out that port 80 isn't listening. There are still billable bits involved in that interaction.
One more fair approach would be for the ISP to install an IP-aware counter that only bills you for connections actually established. The only problem there is that not all traffic uses connection sockets. You could be downloading via UDP traffic which looks the same as someone trying to connect to port 80 and failing - it's a packet that comes in and never gets a connection.
That would be a great argument if the world was fair. It's not. The customer is going to be held responsible for paying for the bandwith by default whether he notifed the ISP or not.
If instead you show me a contract saying that they offer you burstable bandwidth at x dollars per y units of data transfer used, then I'm sorry, but you agreed to those terms.
No. What you agreed to, in that case, is to pay for the bandwith you use. If someone floods me with traffic, that isn't bandwith I'm using. That's bandwith he's using.
Well, there are delta-wings that combine the function of the ailerons with elevators into one set of flaps, with more complex rules governing their movement. Some canard planes also use this system too, since the wing is at the rear in a canard plane, leaving the canard as a fixed structure.
But that's an odd exception. Most planes separate the elevator from the aileron.
In the US, the machinery varies by location. Where I vote (Wisconsin) there's a simple thick paper sheet with several uncompleted arrows pointing at the choices. You use a black marker to complete one of the arrows in each category, and then YOU feed it into the reading machine yourself on the way out. If there is some reading error that makes the machine reject it, or the machine mangles it, you get to see it happen while you are standing right there, so you can try again with a new ballot. If you filled in the arrow for a write-in candidate blank, the machine sorts your ballot into a seperate pile to be read by humans later - otherwise the machine tells you right away whether or not your ballot was scannable and the vote tallied. That doesn't have the same problem Florida's system did - where you don't find out the ballot is rejected until long after the voter has gone home and become anonymous. Here the voter himself gets to watch the machine count his ballot. A green light and a beep means everything worked.
At the end of the day, the machines from seperate polling places are brought together and their sub-totals are added together.
Here in Wisconsin, we *also* had a margin as close as the one in Florida, but the machinery was deemed reliable enough that the losing side realized a recount was guaranteed to give the same results, so they didn't bother.
The problem is that they didn't even KNOW their employer was going to engage in such activity. Punish those responsible, absolutely. The grunts at the bottom of the company aren't the ones responsible. It's not like holding nazis responsible for attrocities they committed in death camps - it's like holding Joe-average German citizen responsible when he didn't even know what was going on.
I don't get it. What is the actual alleged IP they claim IBM stole? There seemed to be some implications not actually said that I missed. The only thing I can think of is:
1 - IBM licenses the UNIX name, and agrees to some NDA about it. 2 - IBM implements AIX, under the UNIX name. 3 - IBM later on ends up getting behind Linux. 4 - Some of Linux's development is now coming from IBM employees. 5 - Those employees probably previously worked on AIX. 6 - Thus IBM may be taking AIX techniques and putting them into Linux, which violates their agreement #1 at the top of this list.
I think that's SCO's angle, maybe. It's hard to tell from their press release. Of course those of us in the know understand that licensing the right to use a name has nothing to do with copying the technology. Linux is as UNIX-compatable as any other of the UNIX implementations, and that was already true well before IBM's involvement in it. It just can't legally call itself UNIX.
It seems to me that IBM could avoid this by no longer calling AIX a version of "UNIX", but instead calling it "UNIX compatable", just like Linux does, and thus neatly avoid SCO's idiotic case in a manner that doesn't really hurt them in any material way.
I'd argue that the internet *is* TCP/IP. That's the thing that distinguished it from the LANs and WANs that came before, and it's the low-level piece on top of which HTML and HTTP are built. TCP/IP without HTML and HTTP is still something useful, while HTML and HTTP without TCP/IP is something that doesn't work. They are not just an "array of" seperate technologies. They are a heirarchy of seperate technologies, and TCP/IP is the crucial lynchpin of that heirarchy. There are a number of seperate low level networking protocols, but they all exist to support TCP/IP on top of them, and then all the higher-level protocols exist on top of TCP/IP. (Yes, there is UDP, but it is still on top of IP.)
The system has always been as bad as it was in the last election. Ballots lost on the way to the counting center, polling stations running out of ballots, ballots getting jammed in the counting machines, people not understanding what they were doing. It's always been crap. The margin of error was always one or two percent. It's not that people got stupider, it's that this was the first time the margin was close enough that this always-existant problem became relevant.
"Their upstream" is the ISP. If the ISP drags their feet on blocking the traffic, the customer pays for it. Thus there is incentive for the iSP to drag their feet.
I just got an idea for a software project to solve this problem. It would be like the Windows program "ZoneAlarm", except that instead of configuring the firewall locally on your own machine, it would remotely administrate a firewall on the ISP for your traffic. The interface would have to be OS-agnostic (or I'd be pissed), so something like a small java application would work. (And the interface program doesn't have to do much but monitor an open socket and ask you whatever questions the server tells it to ask, so the bloatness of Java wouldn't matter much.)
That way the ISP could adopt a policy of charging you for traffic you explicitly allowed, and denying all traffic you didn't, and it would be simple enough to use that you don't need to be a computer expert to use it.
For the business customer with a full-time sysadmin, the same kind of firewall config at the ISP could occur, but the business customer could automate the configuration of said firewall by being given a programattic way of sending it commands in scripts.
That's not a good soultion either, because ISP's should be able to charge customers when they choose to download something large. Let's say you download some ISO images for a linux distro. That's traffic you *wanted*. Let's say you visit a porn site and download large quicktime movies. That's traffic you *wanted*. I think a good solution would be to do the following:
1 - Provide an automatic system where each customer can tell the ISP how to configure it's firewall (since the packets have to be blocked at the ISP, before they get counted against you.) The default is "block everything" when you first sign up for service, and you have to explicitly tell it otherwise.
2 - Only charge the user if they recieve traffic they have not chosen to block using #1 above.
NO, NO, NO. To use the cellphone analogy, it's like someone repeatedly attempting to call your phone, and you choosing not to answer, but still getting charged per attempt. It is NOT like what cell phone companies actually do, which is to charge you WHEN YOU PICK UP THE PHONE AND ACCEPT THE CALL, and not a moment before.
The ISP is measuring the incoming traffic before it even reaches your company's firewall.
Slashdot Mirror
That sounds right to me. Maybe I'm a user that signed on to a mailing list because I actually wanted those weekly product listings from them. How does the ISP (which is what the university is in this case) know whether or not the e-mail is wanted or not? I'll filter it myself. *
* - exception - if the ISP (university) wants to charge users for how much e-mail they recieve, then spam blocking does have to be done before the user sees the mail, otherwise he gets charged for everything, solicited or not.
From the efficient screensaver dept.: POKE 53281,0:POKE 53280,0:POKE 646,0
Okay, I know poke 5328[01], 0 will set the screen black, but what does poke 646,0 do?
I don't see them giving a fine to the victim, charging him for taking their time.
True, but the whole point of this discussion is that even when the traffic hitting you is "syn, udp, echo packets, etc", the ISP still bills you for it the same as if it had been real traffic.
This means the US government has some very archaic constructs not present in the newer constitutions of other countries. One of them is the voting system. Because the country was first founded as a loose joining of individual colonies that thought of themselves as independant countries joining an alliance, the national voting system only dictates how each state has to send its results to the national tally, and how many votes each state gets. Originally each state got to pick for itself how it was going to handle the voting process within its borders. That's still true today. The only thing the federal government mandates is that the voting age must be 18. (And there was still some wriggle-room in the wording of the constitution that allowed some states to add extra rules designed to supress the black voters, until the civil rights movement of the '60s got some amendments added to close those loopholes.) But everything else about the details of the voting process is still up to the state. So that's why the tallying system is not uniform. In some states, the entire state uses a uniform system. In other states, each small voting district gets to do it their own way. That was the case in Florida. The unfairness with this system is that the funding to pay for the vote tallying systems comes from the local taxes even though the vote is fufilling a national rather than a local need. Since voting districts are very small areas, sometimes only a few square miles, this means that poor sections of town tend to have lower quality voting systems with larger margins of error. This is why Gore was so adamant about a recount in Florida - the areas where people are poorer tend to vote more for Democrats, and those are the areas where the greatest possibility for error were. Today it's impossible to tell what the vote actually ended up being, since the data gathering system itself is flawed to the point where the data available, even if it is counted and re-counted, isn't enough.
(Actually, Republican votes DO tend to come in earlier in the day that Democrat votes. I think I know the reason why - mostly having to do with the fact that almost all white-collar jobs have the same normal get-up-in-the-morning schedule, while blue-collar jobs tend to have schedules that vary a lot in their times. Since most voters vote on their way to or from work, the schedule of their work tends to affect when they vote.)
This is also why I am strongly opposed to the television news broadcasting their guesses as to the results (based on exit polls) before the polls are closed. Knowing which candidate is winning can change who you vote for, giving an advantage to the later-in-the-day voter. For example, if you really like candidate A, think B is kind of okay, but really despise C, and you find that candidate A has no hope of winning, but B does have some hope of winning, you might decide to vote for B instead of A just to help oppose C. If you had voted before knowing the results so far you would have been voting for A. (And since this country spans 6 seperate time zones, that ends up meaning that people voting in the west get to find out the results of the polls in the east while their polls are still open, and since this election had signifigant showings for a third-party candidate, I'm sure that this A/B/C example above realy did happen in a lot of cases.)
DON'T PUT WORDS IN MY MOUTH BY IMPLICATION. I never said "slayed". I said "removed". I was thinking of deportation, actually.
Niether do I. You can't wash away guilt you never had.
NO, that's precisely how flood denial of service attacks work. The attacker comprimises many other machines first, and used them to attack you from multiple locations. That makes it a pain to firewall the attack away.
Just because the traffic hits you from more than one machine doesn't necessarily mean it's accidental normal traffic.
NO. That isn't the situation. The situation is that you never pick up the phone when the person is calling from a known blacklist of numbers, never let those numbers switch over to voice-mail, and STILL are billed for the mere attempt to contact you.
Making it so that you may only avoid charges for OTHER PEOPLE using your line is to buy an extra service from the ISP sounds exactly like protection to me - the Mafia type of protection. "Buy our special protection service, or there's no telling what might just so happen to get billed to your account when strangers like Vinnie and Eddie here flood your computer."
It looks like you don't know that data is still sent to your computer to find out that port 80 isn't listening. There are still billable bits involved in that interaction.
One more fair approach would be for the ISP to install an IP-aware counter that only bills you for connections actually established. The only problem there is that not all traffic uses connection sockets. You could be downloading via UDP traffic which looks the same as someone trying to connect to port 80 and failing - it's a packet that comes in and never gets a connection.
That would be a great argument if the world was fair. It's not. The customer is going to be held responsible for paying for the bandwith by default whether he notifed the ISP or not.
No. What you agreed to, in that case, is to pay for the bandwith you use. If someone floods me with traffic, that isn't bandwith I'm using. That's bandwith he's using.
And when the police can't find the actual mugger, should they just arrest the victim for lack of anything better to do?
HTTP is still tied to the URL, which includes the assumption that the host is reachable using a DNS hostname or an IP address.
Well, there are delta-wings that combine the function of the ailerons with elevators into one set of flaps, with more complex rules governing their movement. Some canard planes also use this system too, since the wing is at the rear in a canard plane, leaving the canard as a fixed structure.
But that's an odd exception. Most planes separate the elevator from the aileron.
In the US, the machinery varies by location. Where I vote (Wisconsin) there's a simple thick paper sheet with several uncompleted arrows pointing at the choices. You use a black marker to complete one of the arrows in each category, and then YOU feed it into the reading machine yourself on the way out. If there is some reading error that makes the machine reject it, or the machine mangles it, you get to see it happen while you are standing right there, so you can try again with a new ballot. If you filled in the arrow for a write-in candidate blank, the machine sorts your ballot into a seperate pile to be read by humans later - otherwise the machine tells you right away whether or not your ballot was scannable and the vote tallied. That doesn't have the same problem Florida's system did - where you don't find out the ballot is rejected until long after the voter has gone home and become anonymous. Here the voter himself gets to watch the machine count his ballot. A green light and a beep means everything worked.
At the end of the day, the machines from seperate polling places are brought together and their sub-totals are added together.
Here in Wisconsin, we *also* had a margin as close as the one in Florida, but the machinery was deemed reliable enough that the losing side realized a recount was guaranteed to give the same results, so they didn't bother.
The problem is that they didn't even KNOW their employer was going to engage in such activity. Punish those responsible, absolutely. The grunts at the bottom of the company aren't the ones responsible. It's not like holding nazis responsible for attrocities they committed in death camps - it's like holding Joe-average German citizen responsible when he didn't even know what was going on.
I don't get it. What is the actual alleged IP they claim IBM stole? There seemed to be some implications not actually said that I missed. The only thing I can think of is:
1 - IBM licenses the UNIX name, and agrees to some NDA about it.
2 - IBM implements AIX, under the UNIX name.
3 - IBM later on ends up getting behind Linux.
4 - Some of Linux's development is now coming from IBM employees.
5 - Those employees probably previously worked on AIX.
6 - Thus IBM may be taking AIX techniques and putting them into Linux, which violates their agreement #1 at the top of this list.
I think that's SCO's angle, maybe. It's hard to tell from their press release. Of course those of us in the know understand that licensing the right to use a name has nothing to do with copying the technology. Linux is as UNIX-compatable as any other of the UNIX implementations, and that was already true well before IBM's involvement in it. It just can't legally call itself UNIX.
It seems to me that IBM could avoid this by no longer calling AIX a version of "UNIX", but instead calling it "UNIX compatable", just like Linux does, and thus neatly avoid SCO's idiotic case in a manner that doesn't really hurt them in any material way.
Look at any aircraft, and the main movement is governent by these four:
Throttle.
Ailerons (via "wing warping).
Elevator.
Rudder.
That basic configuration hasn't changed since Orville and Wilber used it in 1903.
I'd argue that the internet *is* TCP/IP. That's the thing that distinguished it from the LANs and WANs that came before, and it's the low-level piece on top of which HTML and HTTP are built. TCP/IP without HTML and HTTP is still something useful, while HTML and HTTP without TCP/IP is something that doesn't work. They are not just an "array of" seperate technologies. They are a heirarchy of seperate technologies, and TCP/IP is the crucial lynchpin of that heirarchy. There are a number of seperate low level networking protocols, but they all exist to support TCP/IP on top of them, and then all the higher-level protocols exist on top of TCP/IP. (Yes, there is UDP, but it is still on top of IP.)
The system has always been as bad as it was in the last election. Ballots lost on the way to the counting center, polling stations running out of ballots, ballots getting jammed in the counting machines, people not understanding what they were doing. It's always been crap. The margin of error was always one or two percent. It's not that people got stupider, it's that this was the first time the margin was close enough that this always-existant problem became relevant.
That is precisely what the ISP in question is doing wrong. They are not taking action against the culprits, only the victims.
"Their upstream" is the ISP. If the ISP drags their feet on blocking the traffic, the customer pays for it. Thus there is incentive for the iSP to drag their feet.
I just got an idea for a software project to solve this problem. It would be like the Windows program "ZoneAlarm", except that instead of configuring the firewall locally on your own machine, it would remotely administrate a firewall on the ISP for your traffic. The interface would have to be OS-agnostic (or I'd be pissed), so something like a small java application would work. (And the interface program doesn't have to do much but monitor an open socket and ask you whatever questions the server tells it to ask, so the bloatness of Java wouldn't matter much.)
That way the ISP could adopt a policy of charging you for traffic you explicitly allowed, and denying all traffic you didn't, and it would be simple enough to use that you don't need to be a computer expert to use it.
For the business customer with a full-time sysadmin, the same kind of firewall config at the ISP could occur, but the business customer could automate the configuration of said firewall by being given a programattic way of sending it commands in scripts.
That's not a good soultion either, because ISP's should be able to charge customers when they choose to download something large. Let's say you download some ISO images for a linux distro. That's traffic you *wanted*. Let's say you visit a porn site and download large quicktime movies. That's traffic you *wanted*. I think a good solution would be to do the following:
1 - Provide an automatic system where each customer can tell the ISP how to configure it's firewall (since the packets have to be blocked at the ISP, before they get counted against you.)
The default is "block everything" when you first sign up for service, and you have to explicitly tell it otherwise.
2 - Only charge the user if they recieve traffic they have not chosen to block using #1 above.
NO, NO, NO. To use the cellphone analogy, it's like someone repeatedly attempting to call your phone, and you choosing not to answer, but still getting charged per attempt. It is NOT like what cell phone companies actually do, which is to charge you WHEN YOU PICK UP THE PHONE AND ACCEPT THE CALL, and not a moment before.
The ISP is measuring the incoming traffic before it even reaches your company's firewall.