It depends. First give me the address, if my navigation device has problems, I'll ask you for directions. When I talk on the phone, I do not usually have a pen and paper nearby (and reading the directions when driving may be a problem) but I can remember the address at least until I enter it in the navigation device.
Can we outlaw headphones and ear buds too? I see far too many people using them.
How about cellphones? If they're just staring down at the screen, they won't see the cars.
Well, in that case, the person will get hit by a car learn a lesson about looking for cars if he/she survives. Looking at a cellphone while crossing a street puts the user (who is making the decision to ignore the risk) at risk more than the driver. On the other hand, cars with stealth mode put the pedestrian at risk more than the driver (who is the one making the decision to use a quiet car), so they may be a problem.
Crossing the road legally involves looking for oncoming vehicles unless there is a traffic light or sign which requires them to stop regardless of pedestrians. There is nothing whatsoever in this arrangement that involves the pedestrian hearing the vehicle.
And locks are not needed because there are laws against theft.
Yes, the law says that a car must yield to pedestrians and that pedestrians must cross the street at specific locations. In practice, however, there are drunk drivers and drivers, who think that since pedestrian is not a tree, he can get out of the way so there is no reason to stop. Blind people cannot see an incoming car, but they can hear it, especially if the car is with an internal combustion engine. Electric motors are quieter than gas or diesel engines, so hearing the car might be a problem.
On the other hand, I have never been near a Prius, so I can't say if it is too quiet or not, but some of the newer cars with gasoline engines are very quiet.
In my network, DNS server is on the router too (though the router is a PC instead of some small box). If it goes down, I can still use IP addresses for internal nodes (my laptop has a dynamic IP, but I usually don't run services that I need remote access to on my laptop and all other computers have static IPs, so I can use the IP or hostname if I want to).
Also, there is another problem in having public IPs in internal network if you have multiple ISPs. I'll use an IPv4 example here, but I assume the same would be with IPv6:
Suppose I have two ISPs, both of which give me a/28 (14 useful hosts). One ISP gives me 1.0.0.0/28 and the other 2.0.0.0/28. I assign them to the computers inside my network, do not use NAT and so on, however, I only use one connection at a time (let's say that ISP #2 charges for time or data or there is some other reason why that connection is not always on, just when the main connection dies).
My network works happily until the internet connection fails. The router automatically connects to ISP #2 and distributes new IPs to my hosts (2.... instead of 1...). Let's say I was transferring a file over my local network. The connection would break because both computers got new IPs and their old IPs are no longer valid.
Contrast this with a network behind NAT - internal IPS stay constant and only the router cares that the external IP changes.
Currently, the same server -- even the same process (dnsmasq) -- provides both of these, and it's been stable and reliable, but again, if it's down, that's an immediate priority.
A lot of things in my network become an immediate priority when they fail, but if I don't have a spare part to replace the broken one I might need to wait a few days until I get it. So, I have made my network so that no single failure can bring it all down (well, except the switch, but I have a 100mbps switch and a 10mbps hub in that case). Degraded - yes, completely down - no.
To get around that, you'd need a VPN -- but if you're willing to expose a VPN, why not just expose the machines?
Depends on what I am exposing. For example, I use VPN to connect to my home network when I am not at home. That would not change with IPv6 because now only the VPN service is exposed (a.k.a. anyone can try to hack it), but the internal services can be unencrypted because if someone hacks the VPN then he most likely would be able to hack whatever services there are. I would not want to use Windows file sharing over a public network, but VPN is OK.
Compare that to something like SIP, which could be difficult to forward, but becomes simple and natural by simply giving each phone in the house a real, Internet-addressable IP.
It looks like the designers of SIP were actually trying to make it impossible to use with NAT.
I would still try to use NAT with IPv6 (if iptables or other sotware supported it), if not for anything else then to have my internal IPs constant and independent of the ISP (I assume my external IP(s) would still change if I changed the ISP or had to use the backup connection).
One ISP I know, offers 3 IPs with its fastest plan. I honestly would not be able to make use of them, well, I could assign separate IPs for 2 of my computers, but that would not become more useful to me than just forwarding ports.
Yes, NAT is a kludge, yes IPv6 would be better, but NAT works well enough and IPv6 is not (currently) much better than IPv4+NAT.
Is DRM futile? Depends on who you ask, I guess. I think these results show that DRM can work effectively if sufficient effort is put into it.
Well, as you said, it depends. "DRM" of an online-only game (like WoW) most likely cannot be cracked (if you want to play on the official server). DRM on a closed platform can also be quite strong. DRM on an open platform, like PC will be much weaker. There are no PS3 emulators, but there are PC emulators, so PC is easier to debug.
This is for interactive stuff, like games and applications. It will always be possible to copy non-interactive things like audio and video, at least using analog means.
Which gives me the idea for a new DRM scheme...has anyone ever made a USB hardware dongle for a software license that has an internal CPU performing complex calculations needed for the host software to work properly? If enough of the software depended on this internal CPU, how could you crack it?
You can find out what the processor is doing and emulate this or manufacture clone dongles. Yes, it is much more difficult than cracking Securom or whatever, but possible. On the other hand, the dongle itself is expensive, so the prices of games would have to increase and they are already the maximum what people would pay.
Ok, let's say that I have a network, connected to a router that is connected to two ISPs (one connection is much faster than the other, which is used as a backup). Currently, the router uses the main connection for everything. If the connection goes down, the router starts using the other connection (it may have to dial it, if it is dial-up or DSL with PPPoE). All currently established TCP connections to outside break, but new ones connect successfully (since they now use the other link). No change happens for the computers on the network.
Does each computer get two IP addresses with IPv6? If so, how does it know which source IP to use (the backup connection may be slower, or I may need to pay for each GB transferred)? Do the PCs have to have a third IP address, this time for the internal network so that the internal network connections are not disturbed when one link to outside goes down?
For example, try using FTP over IPv6 through a NAT
I haven't, but wouldn't it work just like with IPv4? Client uses PASV (only outgoing connections) and the NAT router in front of the server has a port range forwarded.
There MAY be niche scenarios where non-routable IPs are desirable in the IPv6 world, I honestly can't think of any. Can you?
With NAT, if I make a mistake in config, the result is usually that the port remains closed. With only firewall, it is easier to make a mistake and have a port open, for example to allow a certain port to all machines instead of just the one.
I assume that with IPv6 the IP addresses (or the subnet) will be assigned by the ISP, so if I change ISPs (or I have two connections and one goes down so I have to use the other one), the IP addresses in my local network will change, which is not that great. With NAT, I can keep the same IP addresses in my internal network, while the external IP address(es) can change and only the router cares about them.
With NAT, I can make it appear that I have only one (two etc) machine when in fact I have more than one.
Unless I want to have two web servers (both at port 80 but different physical machines) or something like that, I do not need more than one external IP address. Longer IP addresses would not be more useful to me and I would not be able to remember them, so I would have to write them down or print them.
But other than the conceptual "a IP for each machine", having it is not any better than NAT. Well, unless you are at risk of running out of ports, which a lot of people aren't.
Does Windows 2000 support IPv6? In any case, my printer does not support IPv6 it does support IPv4 and IPX.
Uhhhhh, apache 1.3 worked with patches, so post '98, a mere dozen years ago, is OK with minimal effort.
I meant servers as in sites, email servers etc. I don't think there are a lot of sites that you can only reach by IPv6. And there is a good reason - no company would be stupid enough to prevent almost all people from connecting to its (public) servers.
IPv4 devices will not be able to access IPv6 devices, which means that if you have devices with old OS in your network, you will have to use both v4 and v6.
Also, there is that chicken-and-egg problem. There is only a small amount of servers that support v6 and even smaller number of them support only v6.
1.There is no reason for a user to upgrade to IPv6 (they may need to change their router, the new IP address is almost impossible to remember etc) because little would be gained from it (very few servers support only IPv6). A IPv6-only connection wold be almost useless. 2.There is no reason for a company to upgrade to IPv6 because all of their clients can use IPv4. Goto 1.
How the fuck can you enjoy broadcast radio and have an IQ above 80?
Simple - 1.My favorite radio station at night broadcasts music non stop. I like a lot of songs they broadcast, but do not know the artists and song titles (since the radio station never says them).
2.Radio requires least involvement to listen (turn the tuner on, tune to a station, that's it), compared to records, tapes, CDs or files. Useful when I want to do something that requires a lot of concentration, but still want the music to play in the background.
Depends. Maybe the same hardware costs the same, but AFAIK, Apple sells only complete computers computers (case + CPU + motherboard + RAM etc). When I want to upgrade, I can just buy the part(s) I want to upgrade and keep the rest (for example, I can upgrade the motherboard+RAM+CPU and keep the old video card, hard drives, CD drive and sound card).
Also, I have much more options when building a Windows/Linux PC than I do if I were to buy a complete computer from Apple or someone else.
I built my current PC (2x Opteron 270, 4GB RAM) ~3 years ago and then the OS choice was either XP or Vista. I probably would have tried 7 if it was available at that time.
That they are pretty does not mean that I like keeping them on, but I understand why Microsoft included them. They run on graphics hardware, but my video card (ATI HD900XT) downclocks in desktop mode to save power, if Aero causes it to enter 3D mode then my PC will use ~100W more power.
Why is it that a linux system isn't configured well if it doesn't run compiz, but at the same time aero on windows is a bloated monstrosity that needs disabling?
I never said that.
Anyway, the main reason to run windows 7 is responsiveness. XP regularly blocks if you're multi-tasking, because it has lousy CPU scheduling and no IO scheduling.
I didn't know that. Anyway, not enough of a reason for me to format and reinstall. My 2x dual core PC rarely block anyway.
However, 7 seems to use more resources compared to XP or 2003. I have 7 on a virtual machine (to experiment with it and also to find out how to mod the UI) in a server with 3x 700MHz CPUs and it is so much slower than 2003 running in another virtual machine on the same physical machine, even though 7 gets 1GB RAM and 2003 only 384MB.
PC started randomly turning off, no shutdown, crash or anything, as if the plug was pulled (it wasn't). Temperatures were high, but not too high, also, the problem did not seem to be dependent on the temperatures. Setting the CPU clock to half of original (1GHz instead of 2) seemed to decrease the frequency of the problem, while playing a game almost always resulted in instant off.
So, maybe the power supply is failing. The 3 year old 550W power supply was operating near capacity, also, the PC turned off when the fan of the power supply failed and the power supply overheated (but the fan worked this time). Oh well, I'll get a 750W PSU, that one will last.
A week later and 144EUR paid, I got the new PSU, installed it and the problem remained. Also, during the week the problem worsened - the PC was turning off a few minutes after turning on.
I pulled out all the cards (except VGA) and disconnected all drives. The problem remained.
I replaced the video card, ATI Radeon HD2900XT with a 15 year old PCI card (I have a few newer cards, but all are AGP). The PC booted up and was working. If I add the HD290XT back, the PC starts and works until Knoppix (no reason to try to boot Windows and risk damaging the file system) starts enumerating all devices.
I checked the capacitors, all looked normal.
So, maybe the video card was bad. I ordered a GTX260 (if I'm buying a new card, might as well upgrade). 200EUR. Until I actually got it, I kept the computer with the old PCI card, well, it was at least usable. GTX260arrived, I installed it and the problem came back. Now a bunch of capacitors on the motherboard had bulging out tops. I got the motherboard repaired (15EUR) and my PC works fine.
Oh, and the GTX260 most likely has low quality analog circuitry - video using the DVI->VGA adapter (I have a good CRT monitor) was blurry and the blurriness increased with the resolution and refresh rate (1600x1200@60Hz = quite OK, 1600x1200@85Hz = bad, 1920x1440@85Hz = really bad), so I had to go back to using the HD2900XT and now am in the process of returning the GTX260.
So, if a capacitor looks normal, it does not mean that the capacitor still works.
I used to be like that some time ago. Then I started reinstalling Windows only if I absolutely could not get the old installation to work, because reinstalling all of the programs takes a long time and makes the PC semi-unusable (oh, I know, I'll just use program X here, wait, last time I used it was a year ago, before the reinstall, now where is that setup file...).
I can play with different OSs on virtual machines and an unused PC that I have (now it has Windows 98 installed, because I wanted to play System Shock 2).
I don't have any major problems with Windows 7, well other than the UI changes from XP, but I also see no point in formatting a drive with a working OS on it that is still good enough for me and installing a different OS, and then reinstalling all of my programs, some of which may not work with the new OS. And all of that for what? Pretty graphics (Aero) that will be turned off right after Autorun?
He said "cheap". An dumb hub is cheaper than a managed switch.
Re:Network meltdown due to hub cross-connects
on
Stupid Data Center Tricks
·
· Score: 3, Informative
This should work quite OK with hubs. A hub, after all, sends the packet to every port except the one where it came from. So two hubs in a loop should just forward the same packet back and forth all the time.
Slashdot Mirror
It depends. First give me the address, if my navigation device has problems, I'll ask you for directions. When I talk on the phone, I do not usually have a pen and paper nearby (and reading the directions when driving may be a problem) but I can remember the address at least until I enter it in the navigation device.
Can we outlaw headphones and ear buds too? I see far too many people using them.
How about cellphones? If they're just staring down at the screen, they won't see the cars.
Well, in that case, the person will get hit by a car learn a lesson about looking for cars if he/she survives. Looking at a cellphone while crossing a street puts the user (who is making the decision to ignore the risk) at risk more than the driver. On the other hand, cars with stealth mode put the pedestrian at risk more than the driver (who is the one making the decision to use a quiet car), so they may be a problem.
Crossing the road legally involves looking for oncoming vehicles unless there is a traffic light or sign which requires them to stop regardless of pedestrians. There is nothing whatsoever in this arrangement that involves the pedestrian hearing the vehicle.
And locks are not needed because there are laws against theft.
Yes, the law says that a car must yield to pedestrians and that pedestrians must cross the street at specific locations. In practice, however, there are drunk drivers and drivers, who think that since pedestrian is not a tree, he can get out of the way so there is no reason to stop. Blind people cannot see an incoming car, but they can hear it, especially if the car is with an internal combustion engine. Electric motors are quieter than gas or diesel engines, so hearing the car might be a problem.
On the other hand, I have never been near a Prius, so I can't say if it is too quiet or not, but some of the newer cars with gasoline engines are very quiet.
Because some sound alert is needed. Not everyone can see, out of those people that can see, not everyone can see a car clearly from far enough.
In my network, DNS server is on the router too (though the router is a PC instead of some small box). If it goes down, I can still use IP addresses for internal nodes (my laptop has a dynamic IP, but I usually don't run services that I need remote access to on my laptop and all other computers have static IPs, so I can use the IP or hostname if I want to).
Also, there is another problem in having public IPs in internal network if you have multiple ISPs. I'll use an IPv4 example here, but I assume the same would be with IPv6:
Suppose I have two ISPs, both of which give me a /28 (14 useful hosts). One ISP gives me 1.0.0.0/28 and the other 2.0.0.0/28. I assign them to the computers inside my network, do not use NAT and so on, however, I only use one connection at a time (let's say that ISP #2 charges for time or data or there is some other reason why that connection is not always on, just when the main connection dies).
My network works happily until the internet connection fails. The router automatically connects to ISP #2 and distributes new IPs to my hosts (2.... instead of 1...). Let's say I was transferring a file over my local network. The connection would break because both computers got new IPs and their old IPs are no longer valid.
Contrast this with a network behind NAT - internal IPS stay constant and only the router cares that the external IP changes.
Currently, the same server -- even the same process (dnsmasq) -- provides both of these, and it's been stable and reliable, but again, if it's down, that's an immediate priority.
A lot of things in my network become an immediate priority when they fail, but if I don't have a spare part to replace the broken one I might need to wait a few days until I get it. So, I have made my network so that no single failure can bring it all down (well, except the switch, but I have a 100mbps switch and a 10mbps hub in that case). Degraded - yes, completely down - no.
What? Why are you using IPs directly? That's what DNS is for.
IPs work even when the DNS server is down/unreachable. Oh, and you have to enter the IP of the DNS server.
To get around that, you'd need a VPN -- but if you're willing to expose a VPN, why not just expose the machines?
Depends on what I am exposing. For example, I use VPN to connect to my home network when I am not at home. That would not change with IPv6 because now only the VPN service is exposed (a.k.a. anyone can try to hack it), but the internal services can be unencrypted because if someone hacks the VPN then he most likely would be able to hack whatever services there are. I would not want to use Windows file sharing over a public network, but VPN is OK.
Compare that to something like SIP, which could be difficult to forward, but becomes simple and natural by simply giving each phone in the house a real, Internet-addressable IP.
It looks like the designers of SIP were actually trying to make it impossible to use with NAT.
I would still try to use NAT with IPv6 (if iptables or other sotware supported it), if not for anything else then to have my internal IPs constant and independent of the ISP (I assume my external IP(s) would still change if I changed the ISP or had to use the backup connection).
One ISP I know, offers 3 IPs with its fastest plan. I honestly would not be able to make use of them, well, I could assign separate IPs for 2 of my computers, but that would not become more useful to me than just forwarding ports.
Yes, NAT is a kludge, yes IPv6 would be better, but NAT works well enough and IPv6 is not (currently) much better than IPv4+NAT.
Is DRM futile? Depends on who you ask, I guess. I think these results show that DRM can work effectively if sufficient effort is put into it.
Well, as you said, it depends. "DRM" of an online-only game (like WoW) most likely cannot be cracked (if you want to play on the official server). DRM on a closed platform can also be quite strong. DRM on an open platform, like PC will be much weaker. There are no PS3 emulators, but there are PC emulators, so PC is easier to debug.
This is for interactive stuff, like games and applications. It will always be possible to copy non-interactive things like audio and video, at least using analog means.
Which gives me the idea for a new DRM scheme...has anyone ever made a USB hardware dongle for a software license that has an internal CPU performing complex calculations needed for the host software to work properly? If enough of the software depended on this internal CPU, how could you crack it?
You can find out what the processor is doing and emulate this or manufacture clone dongles. Yes, it is much more difficult than cracking Securom or whatever, but possible. On the other hand, the dongle itself is expensive, so the prices of games would have to increase and they are already the maximum what people would pay.
How is that supposed to work?
Ok, let's say that I have a network, connected to a router that is connected to two ISPs (one connection is much faster than the other, which is used as a backup). Currently, the router uses the main connection for everything. If the connection goes down, the router starts using the other connection (it may have to dial it, if it is dial-up or DSL with PPPoE). All currently established TCP connections to outside break, but new ones connect successfully (since they now use the other link). No change happens for the computers on the network.
Does each computer get two IP addresses with IPv6? If so, how does it know which source IP to use (the backup connection may be slower, or I may need to pay for each GB transferred)? Do the PCs have to have a third IP address, this time for the internal network so that the internal network connections are not disturbed when one link to outside goes down?
For example, try using FTP over IPv6 through a NAT
I haven't, but wouldn't it work just like with IPv4? Client uses PASV (only outgoing connections) and the NAT router in front of the server has a port range forwarded.
There MAY be niche scenarios where non-routable IPs are desirable in the IPv6 world, I honestly can't think of any. Can you?
With NAT, if I make a mistake in config, the result is usually that the port remains closed. With only firewall, it is easier to make a mistake and have a port open, for example to allow a certain port to all machines instead of just the one.
I assume that with IPv6 the IP addresses (or the subnet) will be assigned by the ISP, so if I change ISPs (or I have two connections and one goes down so I have to use the other one), the IP addresses in my local network will change, which is not that great. With NAT, I can keep the same IP addresses in my internal network, while the external IP address(es) can change and only the router cares about them.
With NAT, I can make it appear that I have only one (two etc) machine when in fact I have more than one.
Unless I want to have two web servers (both at port 80 but different physical machines) or something like that, I do not need more than one external IP address. Longer IP addresses would not be more useful to me and I would not be able to remember them, so I would have to write them down or print them.
But other than the conceptual "a IP for each machine", having it is not any better than NAT. Well, unless you are at risk of running out of ports, which a lot of people aren't.
Only if you use stuff much older than, say, Y2K.
Does Windows 2000 support IPv6?
In any case, my printer does not support IPv6 it does support IPv4 and IPX.
Uhhhhh, apache 1.3 worked with patches, so post '98, a mere dozen years ago, is OK with minimal effort.
I meant servers as in sites, email servers etc. I don't think there are a lot of sites that you can only reach by IPv6. And there is a good reason - no company would be stupid enough to prevent almost all people from connecting to its (public) servers.
Incompatibility with v4.
IPv4 devices will not be able to access IPv6 devices, which means that if you have devices with old OS in your network, you will have to use both v4 and v6.
Also, there is that chicken-and-egg problem. There is only a small amount of servers that support v6 and even smaller number of them support only v6.
1.There is no reason for a user to upgrade to IPv6 (they may need to change their router, the new IP address is almost impossible to remember etc) because little would be gained from it (very few servers support only IPv6). A IPv6-only connection wold be almost useless.
2.There is no reason for a company to upgrade to IPv6 because all of their clients can use IPv4. Goto 1.
How the fuck can you enjoy broadcast radio and have an IQ above 80?
Simple -
1.My favorite radio station at night broadcasts music non stop. I like a lot of songs they broadcast, but do not know the artists and song titles (since the radio station never says them).
2.Radio requires least involvement to listen (turn the tuner on, tune to a station, that's it), compared to records, tapes, CDs or files. Useful when I want to do something that requires a lot of concentration, but still want the music to play in the background.
Depends. Maybe the same hardware costs the same, but AFAIK, Apple sells only complete computers computers (case + CPU + motherboard + RAM etc). When I want to upgrade, I can just buy the part(s) I want to upgrade and keep the rest (for example, I can upgrade the motherboard+RAM+CPU and keep the old video card, hard drives, CD drive and sound card).
Also, I have much more options when building a Windows/Linux PC than I do if I were to buy a complete computer from Apple or someone else.
Oh well, at least I can play more games and my hardware costs less.
I built my current PC (2x Opteron 270, 4GB RAM) ~3 years ago and then the OS choice was either XP or Vista. I probably would have tried 7 if it was available at that time.
15 Euros for repair (replacing bad caps), not a new motherboard.
That they are pretty does not mean that I like keeping them on, but I understand why Microsoft included them. They run on graphics hardware, but my video card (ATI HD900XT) downclocks in desktop mode to save power, if Aero causes it to enter 3D mode then my PC will use ~100W more power.
Why is it that a linux system isn't configured well if it doesn't run compiz, but at the same time aero on windows is a bloated monstrosity that needs disabling?
I never said that.
Anyway, the main reason to run windows 7 is responsiveness. XP regularly blocks if you're multi-tasking, because it has lousy CPU scheduling and no IO scheduling.
I didn't know that. Anyway, not enough of a reason for me to format and reinstall. My 2x dual core PC rarely block anyway.
However, 7 seems to use more resources compared to XP or 2003. I have 7 on a virtual machine (to experiment with it and also to find out how to mod the UI) in a server with 3x 700MHz CPUs and it is so much slower than 2003 running in another virtual machine on the same physical machine, even though 7 gets 1GB RAM and 2003 only 384MB.
It was fun to me too:
PC started randomly turning off, no shutdown, crash or anything, as if the plug was pulled (it wasn't). Temperatures were high, but not too high, also, the problem did not seem to be dependent on the temperatures. Setting the CPU clock to half of original (1GHz instead of 2) seemed to decrease the frequency of the problem, while playing a game almost always resulted in instant off.
So, maybe the power supply is failing. The 3 year old 550W power supply was operating near capacity, also, the PC turned off when the fan of the power supply failed and the power supply overheated (but the fan worked this time). Oh well, I'll get a 750W PSU, that one will last.
A week later and 144EUR paid, I got the new PSU, installed it and the problem remained. Also, during the week the problem worsened - the PC was turning off a few minutes after turning on.
I pulled out all the cards (except VGA) and disconnected all drives. The problem remained.
I replaced the video card, ATI Radeon HD2900XT with a 15 year old PCI card (I have a few newer cards, but all are AGP). The PC booted up and was working. If I add the HD290XT back, the PC starts and works until Knoppix (no reason to try to boot Windows and risk damaging the file system) starts enumerating all devices.
I checked the capacitors, all looked normal.
So, maybe the video card was bad. I ordered a GTX260 (if I'm buying a new card, might as well upgrade). 200EUR. Until I actually got it, I kept the computer with the old PCI card, well, it was at least usable. GTX260arrived, I installed it and the problem came back. Now a bunch of capacitors on the motherboard had bulging out tops. I got the motherboard repaired (15EUR) and my PC works fine.
Oh, and the GTX260 most likely has low quality analog circuitry - video using the DVI->VGA adapter (I have a good CRT monitor) was blurry and the blurriness increased with the resolution and refresh rate (1600x1200@60Hz = quite OK, 1600x1200@85Hz = bad, 1920x1440@85Hz = really bad), so I had to go back to using the HD2900XT and now am in the process of returning the GTX260.
So, if a capacitor looks normal, it does not mean that the capacitor still works.
I used to be like that some time ago. Then I started reinstalling Windows only if I absolutely could not get the old installation to work, because reinstalling all of the programs takes a long time and makes the PC semi-unusable (oh, I know, I'll just use program X here, wait, last time I used it was a year ago, before the reinstall, now where is that setup file...).
I can play with different OSs on virtual machines and an unused PC that I have (now it has Windows 98 installed, because I wanted to play System Shock 2).
I don't have any major problems with Windows 7, well other than the UI changes from XP, but I also see no point in formatting a drive with a working OS on it that is still good enough for me and installing a different OS, and then reinstalling all of my programs, some of which may not work with the new OS. And all of that for what? Pretty graphics (Aero) that will be turned off right after Autorun?
He said "cheap". An dumb hub is cheaper than a managed switch.
This should work quite OK with hubs. A hub, after all, sends the packet to every port except the one where it came from. So two hubs in a loop should just forward the same packet back and forth all the time.