You mentioned subway systems, which reminded me of how that can go wrong...ever notice that Verizon Wireless is the only cell carrier that has transmitters in Washington DC's Metro system? It's because Verizon (the phone company) simply won't let competing tower operators connect to their subway infrastructure.
I used to work for a printing company that did color proof output on Canon color copiers. We had exactly the problem you described once when we tried to print a job that had a partial currency image on each page - the printer shut itself down after printing five pages with a black box over the image.
Apparently things are even more insiduous than you think...the tech told me that each time the currency detection code is triggered, the algorithm adjusts its sensitivity upwards (to thwart folks who try to "experiment" to beat it).
Even worse, every time there's a service call to re-activate a printer that shut down in this manner, a secret service report must be filed, along with a report (or hard copy, preferably) of what exactly what the user was trying to copy. As the tech explained it, it's either 1. Canon reports this data, or 2. The Secret Service comes over directly to ask you about it. The former is certainly a far better option, IMO...
That may be true, but I'm sure there are many, many other ways that Microsoft can make life difficult for an OEM who ships machines with a competing OS.
Just like how Apple can "persuade" their resellers to never sell below MSRP. They can't legally demand that outright, but they can certainly say "hey, it looks like your inventory order has been delayed. We'll get right on that...".
What this means is that the prospect of selling OS X is appealing enough to Dell that they're willing to weather whatever retaliation Microsoft may have in store. Can't say I disagree.
Here's my question: is it the tag overhead, or the fact that it's an ASCII file, the reason that people complain about XML file sizes? If it's the latter, one would think that a simple gzip action at load/save would solve that problem...
Covad actuallly provides what you would call "Layer 2" connectivity between their equipment at the local telco's CO and the ISP's equipment using ATM. The ISP will provision both upstream connectivity to the internet and then an ATM trunk to Covad, who feeds all of that ISP's customers terminating at their DSLAMs (where your copper pair is split off to) at the various LEC's central offices and "concentrates" them using ATM PVCs (permanent virtual circuits) onto to the ISP's trunk. The ISP then feeds the data between the internet pipes and the ATM trunk.
Another time, we deduced that someone else had signed up the person in question (the person's last name was recorded in the database as "Assface").
You obviously didn't have a confirmed opt-in system in place then...if you had, the address in question wouldn't have gotten on the list, he would have gotten one email asking him to confirm his subscription, and nothing else if he didn't reply to it.
They are not acting on "our" behalf unless you are the author of the GPLed software being used in violation of the license. Remember, the GPL does not bequeath ownership of the code to the public domain. If it did, then the code would no longer be subject to the GPL.
Why not do 3.x, 4.x,... like every other software developer in the world (well, except Microsoft and Apple...)?
Honestly, I don't understand the insistence on keeping everything at 2.x, 2.x.y, etc. If someone can explain the rationale to me, I'd be quite interested.
Generally, that's not actually the case. The industry standard for usage-based bandwidth billing is a formula called "95p". What happens is that the circuit load is constantly measured (typically in 5-minute averages) throughout the billing period. The highest 5% of those measurements are then thrown out, and the highest measurement remaining is what the month's bill is based on.
This has two implications: First, a customer doesn't get bent over if they get a traffic spike or DoS that is short-lived...you can push you pipe to capacity for almost a day and a half before you actually get billed for running the connection at that level.
Second, given that most pipes see their highest usage during business hours, weekend traffic rarely exceeds the current 95p value, and as such is effectively "free". That's why most companies do their network-based backups at night and on weekends.
As such, it's highly unlikely that the/. effect is putting a significant dent in Cornell's bandwidth bill. If this had been posted at 2pm on a weekday, and folks were hammering the link for days, there might be a different story.
Latency by itself (and router hops by itself) will not kill a VoIP call.
What *will* kill a VoIP call is jitter (wide variations in latency), dropped packets, and out-of-order delivery.
Mind you, the more router hops, the higher the chance that one of those devices is behaving badly and is dropping/queuing/reordering packets. But in general I wouldn't point to router hops alone as an indication of a bad connection.
Also, keep in mind that many providers now run MPLS backbones, which means that intermediate router hops don't show up in traceroutes...
It's not uncommon for controversial sites to be DoS magnets, to the point that the cost of dealing with the DoS traffic far exceeds the revenue generated by hosting the site. I've seen it happen, and sad but true, the financial concerns will trump the free-speech concerns every time.
As far as the breach of contract accusation goes, every ISP has a clause that allows them to take immediate action to protect their network with no notice. That may have been the clause invoked here.
And at the time, you couldn't use a standard PC keyboard on a Mac thanks to the ADB interface, which I'm sure was a big reason people were screaming about that so loudly. Mac keyboards were much more expensive.
The standardization to USB makes that much less of an issue today than it was back then.
Your response shows that you have no experience with this whatsoever. Just because it's UDP and not TCP doesn't mean that QoS is not an issue. Just try to make a VoIP call while running a BitTorrent client without bandwidth limits.
While it is true that dropping a packet in a TCP media stream hurts the audio quality worse than if it was a UDP stream (UDP won't stop and try to retransmit like TCP does), there is still a very real need for QoS, as dropped UDP packets *do* affect audio quality. WonderShaper is your friend.
Somewhat OT, but wasn't there a company putting up WiFi APs into buildings that were adjacent to Starbucks locations (making sure the signal reached the store) and selling WiFi at half of T-Mobile's price?
The fundamental difference between MS and Apple is that MS already has a monopoly in desktop operating systems, and US antitrust laws prohibit leveraging one's monopoly status in one market to monopolize another. Which is exactly what MS did with IE and with Office.
Apple, on the other hand, is not a monopoly, and does not fall under such rules.
True, but what you're doing is working around a fundamental weakness of both IPv4 (the relative scarcity of addresses) and your ISP's policy (not willing to allocate you a static IP for each machine, partly due to #1). With IPv6, both issues are moot.
Slashdot Mirror
You mentioned subway systems, which reminded me of how that can go wrong...ever notice that Verizon Wireless is the only cell carrier that has transmitters in Washington DC's Metro system? It's because Verizon (the phone company) simply won't let competing tower operators connect to their subway infrastructure.
I used to work for a printing company that did color proof output on Canon color copiers. We had exactly the problem you described once when we tried to print a job that had a partial currency image on each page - the printer shut itself down after printing five pages with a black box over the image.
Apparently things are even more insiduous than you think...the tech told me that each time the currency detection code is triggered, the algorithm adjusts its sensitivity upwards (to thwart folks who try to "experiment" to beat it).
Even worse, every time there's a service call to re-activate a printer that shut down in this manner, a secret service report must be filed, along with a report (or hard copy, preferably) of what exactly what the user was trying to copy. As the tech explained it, it's either 1. Canon reports this data, or 2. The Secret Service comes over directly to ask you about it. The former is certainly a far better option, IMO...
Well, I'm sure it has no more hooks into the Windows kernel than "bash:# cat foo > /proc/bar" has into a UNIX kernel...
That may be true, but I'm sure there are many, many other ways that Microsoft can make life difficult for an OEM who ships machines with a competing OS.
Just like how Apple can "persuade" their resellers to never sell below MSRP. They can't legally demand that outright, but they can certainly say "hey, it looks like your inventory order has been delayed. We'll get right on that...".
What this means is that the prospect of selling OS X is appealing enough to Dell that they're willing to weather whatever retaliation Microsoft may have in store. Can't say I disagree.
...it just means that Steve is a big Kraftwerk fan.
Here's my question: is it the tag overhead, or the fact that it's an ASCII file, the reason that people complain about XML file sizes? If it's the latter, one would think that a simple gzip action at load/save would solve that problem...
Covad actuallly provides what you would call "Layer 2" connectivity between their equipment at the local telco's CO and the ISP's equipment using ATM. The ISP will provision both upstream connectivity to the internet and then an ATM trunk to Covad, who feeds all of that ISP's customers terminating at their DSLAMs (where your copper pair is split off to) at the various LEC's central offices and "concentrates" them using ATM PVCs (permanent virtual circuits) onto to the ISP's trunk. The ISP then feeds the data between the internet pipes and the ATM trunk.
Or the classic Onion headline that went something like: "African Tribesman Uses IBM Modem To Crush Nut"...
I do wonder how PearPC manages to do it, then...
Another time, we deduced that someone else had signed up the person in question (the person's last name was recorded in the database as "Assface").
You obviously didn't have a confirmed opt-in system in place then...if you had, the address in question wouldn't have gotten on the list, he would have gotten one email asking him to confirm his subscription, and nothing else if he didn't reply to it.
They are not acting on "our" behalf unless you are the author of the GPLed software being used in violation of the license. Remember, the GPL does not bequeath ownership of the code to the public domain. If it did, then the code would no longer be subject to the GPL.
Why not do 3.x, 4.x, ... like every other software developer in the world (well, except Microsoft and Apple...)?
Honestly, I don't understand the insistence on keeping everything at 2.x, 2.x.y, etc. If someone can explain the rationale to me, I'd be quite interested.
Generally, that's not actually the case. The industry standard for usage-based bandwidth billing is a formula called "95p". What happens is that the circuit load is constantly measured (typically in 5-minute averages) throughout the billing period. The highest 5% of those measurements are then thrown out, and the highest measurement remaining is what the month's bill is based on.
/. effect is putting a significant dent in Cornell's bandwidth bill. If this had been posted at 2pm on a weekday, and folks were hammering the link for days, there might be a different story.
This has two implications: First, a customer doesn't get bent over if they get a traffic spike or DoS that is short-lived...you can push you pipe to capacity for almost a day and a half before you actually get billed for running the connection at that level.
Second, given that most pipes see their highest usage during business hours, weekend traffic rarely exceeds the current 95p value, and as such is effectively "free". That's why most companies do their network-based backups at night and on weekends.
As such, it's highly unlikely that the
Better make sure your altitude stays below 2,000 feet...
Latency by itself (and router hops by itself) will not kill a VoIP call.
What *will* kill a VoIP call is jitter (wide variations in latency), dropped packets, and out-of-order delivery.
Mind you, the more router hops, the higher the chance that one of those devices is behaving badly and is dropping/queuing/reordering packets. But in general I wouldn't point to router hops alone as an indication of a bad connection.
Also, keep in mind that many providers now run MPLS backbones, which means that intermediate router hops don't show up in traceroutes...
The track on "Original Fire" is a remix of the track "Original Control" which originally appeared on the 1991 Satyricon album.
:)
Also, you referred to MBM as "bad club music". What say we take this outside, 'k?
It's not uncommon for controversial sites to be DoS magnets, to the point that the cost of dealing with the DoS traffic far exceeds the revenue generated by hosting the site. I've seen it happen, and sad but true, the financial concerns will trump the free-speech concerns every time.
As far as the breach of contract accusation goes, every ISP has a clause that allows them to take immediate action to protect their network with no notice. That may have been the clause invoked here.
And at the time, you couldn't use a standard PC keyboard on a Mac thanks to the ADB interface, which I'm sure was a big reason people were screaming about that so loudly. Mac keyboards were much more expensive.
The standardization to USB makes that much less of an issue today than it was back then.
True, but try to send a fax using G.729 or have your TiVO dial in for the program info...
Your response shows that you have no experience with this whatsoever. Just because it's UDP and not TCP doesn't mean that QoS is not an issue. Just try to make a VoIP call while running a BitTorrent client without bandwidth limits.
While it is true that dropping a packet in a TCP media stream hurts the audio quality worse than if it was a UDP stream (UDP won't stop and try to retransmit like TCP does), there is still a very real need for QoS, as dropped UDP packets *do* affect audio quality. WonderShaper is your friend.
Somewhat OT, but wasn't there a company putting up WiFi APs into buildings that were adjacent to Starbucks locations (making sure the signal reached the store) and selling WiFi at half of T-Mobile's price?
The fundamental difference between MS and Apple is that MS already has a monopoly in desktop operating systems, and US antitrust laws prohibit leveraging one's monopoly status in one market to monopolize another. Which is exactly what MS did with IE and with Office.
Apple, on the other hand, is not a monopoly, and does not fall under such rules.
I think the IETF's web site is http://www.ietf.org.
True, but what you're doing is working around a fundamental weakness of both IPv4 (the relative scarcity of addresses) and your ISP's policy (not willing to allocate you a static IP for each machine, partly due to #1). With IPv6, both issues are moot.
Wrong. T-Mobile is a division of Deutche Telekom.