They can be tracked, but not necessarily will be. This has little to do with "poorer countries".
Phone companies make a profit from the stealing of cell phones. It would be foolish to try to combat it (at the cost of maintaining some blacklist server). A stolen phone needs replacement, and the stolen phone is put on the network by someone else, bringing in revenue. They only setup a blacklisting system when they are somehow forced into it, and they get some guarantee that all competitors will get one as well. If not, all phones stolen will be operated on the network of their competitor.
There are also countries where phone companies and law enforcement are passing on the buck to eachother. Phone companies say it is not their task to do something about stealing, law enforcement does not see it as a priority to solve theft of inexpensive items and only act when the stealing was done in robbery-fashion.
Do you think Microsoft would release such patches when there would not be people that would write a worm to exploit the vulnerability? I don't think so. They would just leave the system open to attack by lower-profile hackers and those who gave anxious remarks about that would be sent away with "it is no longer a maintained operating system, please spend money again to buy our new product".
Large-scale attacks are what keeps commercial software vendor's attention to the problem alive.
I think what you describe here will not happen in practice.
In the OS-supported software RAID implementations in Windows and Linux, the system will compare and resync the entire disc after an unexpected system shutdown, and the disks will be consistent with another. Of course there is the issue of inconsistent filesystems and defeated journalling, but it is not a problem of RAID and not related to RAID.
I have no experience with BIOS-and-driver based Software RAID, but I would expect them to behave similarly. Just assuming the parts of a RAID set are consistent after a crash is indeed dangerous.
Don't assume a "hardware RAID" is always better, though. A "hardware RAID" is just a software RAID running on a separate processor, and the software may be written just as poorly as any other implementation. It could be fitted with a battery backup to cover unexpected shutdowns (but so could be your entire system). I have seen Compaq RAID controllers copy the new disk over the valid one after a RAID-1 member disk was swapped after failure, and suggesting the entire array needed to be rebuilt after an onfortunate series of events on a RAID-5 (requiring me to recover the data manually external to the unit). This has reduced my trust in "hardware RAID" a bit, relative to software implementations I can study in source form.
While this is true, it is not very much related to RAID. The conditions you describe determine what happens during a powerloss with nonwritten data. People use RAID (RAID-1 at least) not to recover from those situations, but to recover from a crashed disk. That will still work.
I think you should count yourself lucky when you are in a country where HDTV is even considered.
Over here, aside from the state broadcasters who have their budget cut further every year, we have a bunch of commercial broadcasters who really don't give a d*mn about picture quality. Even programme quality is of no concern to them.
Their business is to take a night's worth of airtime, put as many commercials on it as possible, and then fill the remaining time with material to avoid the viewer zapping away too quickly. Thus their only metric is viewer count.
As the "average viewer" does not own high-end equipment, they are not interested in providing good picture quality as that would not significantly increase the number of viewers. Just like they are not interested in providing subtitling for the deaf, because the average viewer is not deaf.
The fact that those channels still air movies in 16:9 or even 2.35:1 format as letterboxed 4:3 is a very clear proof of that situation. When a letterboxed picture is considered acceptable, don't hold your breath for HDTV...
This sure is a point of concern. I think the design was made this way to put the designers/manufacturers under pressure. Right now, there are some rights management features in DVD players, but the manufacturers just say "up my *ass". Players are supposed to be region-bound but it is hard to find a player that does not have some publicly known "secret" code to disable the region coding check.
With this new system, manufacturers that design hackable players will be punished. That this punishment is via the retailers, and those retailers could become a secondary victim, is bad.
It could make Walmart decide to only sell A-brand stuff and no obscure chinese players that could turn out to be hackable. This will increase prices for the consumer, but then it might also increase the store's profit.
Such "balanced" systems have been in local use here in the past, but our standard 230V system has a neutral and a 230V live wire. The distribution network is 3-phase, delivering 3 230V phases with 400V between the phases. Higher-power motors and equipment that consumes more than about 3kW uses a 3-phase connection, and can use either the 3 230V "star" phases or the 3 400V inter-phase voltages in a "triangle" configuration.
Fortunately here in Europe we have 230V mains. Typical hourse circuits are 16A. We have 3 or more of those in a house. Hungry appliances like an AC would be on a separate breaker. You rarely trip a breaker because of overload here.
Too much AC line ripple to supply stepper motors??? I don't believe you.
Ok, let's check. Yesterday I added two more diskdrives to my system for a total of 5 (plus tape, dvd, cd, 2.5GHz P4, 1GB etc). This is run from an Enermax 430W ATX supply.
Ripple is 5mV on +5 and +3.3, 25mV on +12.
I think you are exaggerating, calling this "way too much".
The same article says the world average is 2.2 kW.
I would not be very proud of burning 5 times more fuel than average. But then I am not the one that looks sheepishly in the camera and utters out that this is needed to provide the citizens with jobs...
Often a drive will not start spinning immediately after power is applied. There is some delay, maybe partly because of drive reset logic and selftest routines.
Modern drives have a jumper setting to turn off the automatic spin up (just like SCSI drives have had for years), and the RAID controller or the BIOS will spin them up one by one when it is polling the buses for available drives.
An IDE drive typically consumes 10W during normal operation and this may increase to 25-30W during the first second of spinup. So in a system with 8 drives you may need 150W of excess capacity during the first second. Not really a reason for a 1kW supply.
You forget to mention that apart from this basic "rectification" part of the power supply, there is also the switchmode converter that actually has a feedback loop to control the converter depending on the output loading. So while you will see more ripple on the capacitor when the load increases, the regulation loop will compensate for that and you still get a stable output on your +5 that does not depend on the ripple.
Of course the output of the converter is again rectified (but not full-wave), but this converter is running at a high frequency so ripple is not the real design limit here. If it were, you could improve the supply by increasing the capacitor values rather than the power rating.
That is the main problem of any opposition against an excess of "rights management" laws. The lobbyists from globalistic large corporations are seen as respectable persons that represent economic wealth and growth. The representatives of small companies and citizens, who are against such a system, are seen as activists and idealists whose voice you should hear but not necessarily take serious. There has been a victory in the case of european software patents, but no doubt they will be back. Lobbying in the individual member states and also finding a new way to enforce their ideas upon the european parliament. In the long run, there is no way "we" can keep defeating those who have the money on their side. Governments are not very much interested in civilian rights and the wellbeing of small businesses, that has been shown over and over again. Politicians cannot please everyone, and they more like to please those that bring the economic growth they are accounted for, and their next job after their political career is over, than they like to please the 4-man company or the man in the street.
1. Installing apps sucks on Linux. And don't say "oh, well if you take a couple hours and learn all the commands and how to work out the dependencies, it's really not that bad." It shouldn't be bad at all. It should be a non-issue in 2005. I heard there are some simpler ways that are gaining popularity, but I won't learn about them till next year.
Apart from the difference between distributions, which is something the Windows community really needs to do something about, I think installation of software is better managed on Linux than on Windows.
Microsoft released its.msi installer format much too late, and in the meantime too many ways of installing applications had appeared. This does not matter too much when you are a home user and just to a "click setup NEXT NEXT NEXT NEXT" type of installation, but when you are an administrator of a business network you want a unified way of installing and upgrading software without user intervention. This is no problem with Linux, but with Windows it requires you to spend time for each new package, to peel the onion of self-unpacking installers with more installers hidden inside them.
When I started with Linux, Windows did not even exist as an operating system (there was a GUI running on top of DOS that was called Windows as well). Before Linux, I used Unix. I was very pleased that I finally could run a decent OS on my home system. Before that, the alternative was MS-DOS.
When Windows appeared as an OS (first Windows 95, later NT) I found myself in the same situation as you are now. With Linux, I could do anything, to use Windows beyond "install it and click on some things" I needed to read a lot. And there was not even sourcecode to read, only "user manuals" that often spent more than 60% of content on basic principles like how to insert a floppy disk.
So indeed, it is not really a difference between Linux and Windows, just a difference between what you know and what is new.
SNTP usually means "send a query, receive a response, put that time into the computer clock, wait some fixed amout of time, repeat".
This is not good on a WAN link because each query may be delayed a different amount of time (mainly depending on queue length on a maxed-out local connection) and thus the actual time set into the clock may vary so much that the clock sometimes steps backward.
A complete NTP implementation avoids this by using filtering and by steering the rate of the clock rather than its actual time setting.
For timesyncing on a fast LAN it does not matter too much, as the variation in delay is very small.
The problem of naive SNTP implementations is that varying delays on your link can make the clock jump backward in time regularly. That is not a good thing.
I am aware of such possibilities, but I thought the poster claimed that W2K somehow has usable NTP support builtin. I never found it...
I run ntpd on a Linux system to keep accurate time, and let the Windows sytems get their time from it over the LAN. SNTP is usable for that (not for syncing over a WAN link that is also used for other purposes). For that, a Windows version of ntpd is a much better choice.
How do you enable true NTP support in Windows (any version)? All I ever found was SNTP and other names for "regularly poll some real server and smack the result into the clock". No PLL, no filtering, no multiple sources, no local clock support, nothing.
Slashdot Mirror
They can be tracked, but not necessarily will be.
This has little to do with "poorer countries".
Phone companies make a profit from the stealing of cell phones. It would be foolish to try to combat it (at the cost of maintaining some blacklist server). A stolen phone needs replacement, and the stolen phone is put on the network by someone else, bringing in revenue.
They only setup a blacklisting system when they are somehow forced into it, and they get some guarantee that all competitors will get one as well. If not, all phones stolen will be operated on the network of their competitor.
There are also countries where phone companies and law enforcement are passing on the buck to eachother.
Phone companies say it is not their task to do something about stealing, law enforcement does not see it as a priority to solve theft of inexpensive items and only act when the stealing was done in robbery-fashion.
Cringely is not a person. It is an alias for the columnist they have hired at that moment.
Do you think Microsoft would release such patches when there would not be people that would write a worm to exploit the vulnerability?
I don't think so. They would just leave the system open to attack by lower-profile hackers and those who gave anxious remarks about that would be sent away with "it is no longer a maintained operating system, please spend money again to buy our new product".
Large-scale attacks are what keeps commercial software vendor's attention to the problem alive.
I think what you describe here will not happen in practice.
In the OS-supported software RAID implementations in Windows and Linux, the system will compare and resync the entire disc after an unexpected system shutdown, and the disks will be consistent with another. Of course there is the issue of inconsistent filesystems and defeated journalling, but it is not a problem of RAID and not related to RAID.
I have no experience with BIOS-and-driver based Software RAID, but I would expect them to behave similarly. Just assuming the parts of a RAID set are consistent after a crash is indeed dangerous.
Don't assume a "hardware RAID" is always better, though. A "hardware RAID" is just a software RAID running on a separate processor, and the software may be written just as poorly as any other implementation. It could be fitted with a battery backup to cover unexpected shutdowns (but so could be your entire system).
I have seen Compaq RAID controllers copy the new disk over the valid one after a RAID-1 member disk was swapped after failure, and suggesting the entire array needed to be rebuilt after an onfortunate series of events on a RAID-5 (requiring me to recover the data manually external to the unit).
This has reduced my trust in "hardware RAID" a bit, relative to software implementations I can study in source form.
While this is true, it is not very much related to RAID.
The conditions you describe determine what happens during a powerloss with nonwritten data. People use RAID (RAID-1 at least) not to recover from those situations, but to recover from a crashed disk. That will still work.
Why is software RAID not RAID?
I think you should count yourself lucky when you are in a country where HDTV is even considered.
Over here, aside from the state broadcasters who have their budget cut further every year, we have a bunch of commercial broadcasters who really don't give a d*mn about picture quality. Even programme quality is of no concern to them.
Their business is to take a night's worth of airtime, put as many commercials on it as possible, and then fill the remaining time with material to avoid the viewer zapping away too quickly. Thus their only metric is viewer count.
As the "average viewer" does not own high-end equipment, they are not interested in providing good picture quality as that would not significantly increase the number of viewers.
Just like they are not interested in providing subtitling for the deaf, because the average viewer is not deaf.
The fact that those channels still air movies in 16:9 or even 2.35:1 format as letterboxed 4:3 is a very clear proof of that situation. When a letterboxed picture is considered acceptable, don't hold your breath for HDTV...
This sure is a point of concern.
I think the design was made this way to put the designers/manufacturers under pressure. Right now, there are some rights management features in DVD players, but the manufacturers just say "up my *ass". Players are supposed to be region-bound but it is hard to find a player that does not have some publicly known "secret" code to disable the region coding check.
With this new system, manufacturers that design hackable players will be punished. That this punishment is via the retailers, and those retailers could become a secondary victim, is bad.
It could make Walmart decide to only sell A-brand stuff and no obscure chinese players that could turn out to be hackable. This will increase prices for the consumer, but then it might also increase the store's profit.
While the extra space of a blue-ray disc relative to a normal DVD is always welcome, it does little to solve the backup problem.
Today a large HD is 400GB or so, and it doesn't matter too much if 100 or 25 discs are required to backup, it is impractical anyway.
Such "balanced" systems have been in local use here in the past, but our standard 230V system has a neutral and a 230V live wire.
The distribution network is 3-phase, delivering 3 230V phases with 400V between the phases.
Higher-power motors and equipment that consumes more than about 3kW uses a 3-phase connection, and can use either the 3 230V "star" phases or the 3 400V inter-phase voltages in a "triangle" configuration.
Fortunately here in Europe we have 230V mains.
Typical hourse circuits are 16A. We have 3 or more of those in a house. Hungry appliances like an AC would be on a separate breaker.
You rarely trip a breaker because of overload here.
Too much AC line ripple to supply stepper motors???
I don't believe you.
Ok, let's check. Yesterday I added two more diskdrives to my system for a total of 5 (plus tape, dvd, cd, 2.5GHz P4, 1GB etc).
This is run from an Enermax 430W ATX supply.
Ripple is 5mV on +5 and +3.3, 25mV on +12.
I think you are exaggerating, calling this "way too much".
The same article says the world average is 2.2 kW.
I would not be very proud of burning 5 times more fuel than average. But then I am not the one that looks sheepishly in the camera and utters out that this is needed to provide the citizens with jobs...
Often a drive will not start spinning immediately after power is applied. There is some delay, maybe partly because of drive reset logic and selftest routines.
Modern drives have a jumper setting to turn off the automatic spin up (just like SCSI drives have had for years), and the RAID controller or the BIOS will spin them up one by one when it is polling the buses for available drives.
An IDE drive typically consumes 10W during normal operation and this may increase to 25-30W during the first second of spinup. So in a system with 8 drives you may need 150W of excess capacity during the first second. Not really a reason for a 1kW supply.
You forget to mention that apart from this basic "rectification" part of the power supply, there is also the switchmode converter that actually has a feedback loop to control the converter depending on the output loading.
So while you will see more ripple on the capacitor when the load increases, the regulation loop will compensate for that and you still get a stable output on your +5 that does not depend on the ripple.
Of course the output of the converter is again rectified (but not full-wave), but this converter is running at a high frequency so ripple is not the real design limit here. If it were, you could improve the supply by increasing the capacitor values rather than the power rating.
That is the main problem of any opposition against an excess of "rights management" laws.
The lobbyists from globalistic large corporations are seen as respectable persons that represent economic wealth and growth. The representatives of small companies and citizens, who are against such a system, are seen as activists and idealists whose voice you should hear but not necessarily take serious.
There has been a victory in the case of european software patents, but no doubt they will be back. Lobbying in the individual member states and also finding a new way to enforce their ideas upon the european parliament.
In the long run, there is no way "we" can keep defeating those who have the money on their side. Governments are not very much interested in civilian rights and the wellbeing of small businesses, that has been shown over and over again. Politicians cannot please everyone, and they more like to please those that bring the economic growth they are accounted for, and their next job after their political career is over, than they like to please the 4-man company or the man in the street.
1. Installing apps sucks on Linux. And don't say "oh, well if you take a couple hours and learn all the commands and how to work out the dependencies, it's really not that bad." It shouldn't be bad at all. It should be a non-issue in 2005. I heard there are some simpler ways that are gaining popularity, but I won't learn about them till next year.
.msi installer format much too late, and in the meantime too many ways of installing applications had appeared.
Apart from the difference between distributions, which is something the Windows community really needs to do something about, I think installation of software is better managed on Linux than on Windows.
Microsoft released its
This does not matter too much when you are a home user and just to a "click setup NEXT NEXT NEXT NEXT" type of installation, but when you are an administrator of a business network you want a unified way of installing and upgrading software without user intervention.
This is no problem with Linux, but with Windows it requires you to spend time for each new package, to peel the onion of self-unpacking installers with more installers hidden inside them.
When I started with Linux, Windows did not even exist as an operating system (there was a GUI running on top of DOS that was called Windows as well).
Before Linux, I used Unix. I was very pleased that I finally could run a decent OS on my home system. Before that, the alternative was MS-DOS.
When Windows appeared as an OS (first Windows 95, later NT) I found myself in the same situation as you are now. With Linux, I could do anything, to use Windows beyond "install it and click on some things" I needed to read a lot. And there was not even sourcecode to read, only "user manuals" that often spent more than 60% of content on basic principles like how to insert a floppy disk.
So indeed, it is not really a difference between Linux and Windows, just a difference between what you know and what is new.
I was not really looking for any add-on product. When I want a credible NTP implementation I can install ntpd (isc.org).
I wondered if Windows really had a builtin NTP client, not the simple NTP (SNTP) that I know about. But I think it hasn't.
I don't see how NTP is better on a WAN link.
SNTP usually means "send a query, receive a response, put that time into the computer clock, wait some fixed amout of time, repeat".
This is not good on a WAN link because each query may be delayed a different amount of time (mainly depending on queue length on a maxed-out local connection) and thus the actual time set into the clock may vary so much that the clock sometimes steps backward.
A complete NTP implementation avoids this by using filtering and by steering the rate of the clock rather than its actual time setting.
For timesyncing on a fast LAN it does not matter too much, as the variation in delay is very small.
The problem of naive SNTP implementations is that varying delays on your link can make the clock jump backward in time regularly.
That is not a good thing.
I am aware of such possibilities, but I thought the poster claimed that W2K somehow has usable NTP support builtin. I never found it...
I run ntpd on a Linux system to keep accurate time, and let the Windows sytems get their time from it over the LAN. SNTP is usable for that (not for syncing over a WAN link that is also used for other purposes). For that, a Windows version of ntpd is a much better choice.
But rdate does not use the NTP protocol...
How do you enable true NTP support in Windows (any version)?
All I ever found was SNTP and other names for "regularly poll some real server and smack the result into the clock".
No PLL, no filtering, no multiple sources, no local clock support, nothing.
Or do you mean an xntpd port to Windows?
You don't understand antenna gain and EIRP.
Discussion closed.