The one that Jews believe in? The one that Muslims believe in? The one that Christians believe in (who is kinda similar to the Jewish one)? The gods that ancient Greeks believed in? The gods that ancient Egyptians believed in? Whatever Scientologists believe in?
There are a lot of religions, most of them are mutually exclusive ("this religion is the only one true religion and our God(s) is/are the only true God(s), everybody else is wrong"). So, if the fact that people believed in something made it true, then all those religions should be true, but at the same time all would be false. So, which one religion is the only one true religion and how do you prove that?
Also, a lot of people believed (and some still do) that the Earth was flat, that didn't make it true.
Do you have any power-saving features enabled at all ?
CPU frequency throttling makes sound crackle when the frequency is changed, which happens a lot when I am watching a movie, so I disabled it. I also disabled turning HDDs off. In short - no.
Also, your server probably does not have one of the most power-hungry video cards (well, it now has been outdone by GTX470 and dual GPU cards).
I thought 250W was kind of normal for a two socket PC, since the CPUs are rated 85W each, so together they draw 170W, the video card at idle also draws some power and the PSU and all other voltage regulators are not 100% efficient.
In Lithuania it's yyyy mm dd or yyyy-mm-dd. Makes the most sense, slowest changing number is the first and fastest changing number is last, like in regular numbers: 100 -> 101 -> 102...
The point is not that these relays generate no heat, but that they can survive up to 500C, while modern CPUs survive about 100C, but start working incorrectly. Discrete transistors and low density logic chips (like the 4xxx series) survive up to 125C. Assuming the power dissipation and ambient temperature is the same, you need much smaller heatsink to keep the device at 400C than you need to keep it at 60C.
Atom at 500kHz would probably use less power, it still would melt if the ambient temperature was 150C.
Common 4:3 resolutions are 1024x768(low), 1280x1024 (technically 5:4, but a lot of 4:3 LCDs have it), 1600x1200 and 1920x1440.
16:10 resolutions usually are 1280x800, 1680x1050, 1920x1200.
My 21" CRT can display up to 2048x1536 (75Hz) or 1920x1440 (85Hz), but I usually use 1600x1200, so a big LCD with 1920x1200 resolution would have the same amount of vertical pixels as my regular resolution and less than the max resolution I use. Even though it would have bigger screen.
I don't know, but skype usually says "less than 1kB/s download, less than 1kB/s upload" I have seen it grow to ~5KB/s once.
Also, if I do not forward the port and send/receive file from somebody who hasn't either, then Skype says that my transfer "is being relayed" and the speed is never over 5KB/s.
Skype has a different solution: When you actually have reasonably fast direct internet access while running Skype, you act as a supernode. This means that you act as a proxy when two NAT-using persons want to talk to each other. If you pay for your data transfer (and you always do, even when it's indirectly), you pay for two persons you don't even know being able to audio/video chat.
So, it works even without forwarding the port. I have forwarded the port and I don't really care that 10KB/s or whatever is used to relay the data of two NAT-using people, after all, I seed torrents at 4-7MB/s.
As for everything else, I now have to agree, v6 would no worse than v4 for configuration, still there's the problem that old devices do not support it, but maybe someone will figure out some kind of NAT so that v4 device can talk to v6 device.
You can have multiple machines running ssh or web servers without screwing around with forwarding different nonstandard ports on the firewall/nat box through to the different internal machines.
That may be useful to some, but I always assign non standard ports. I have a HTTP server, but port 0 is not forwarded. Even when I had one computer, I used nonstandard ports. Now I assign ports based on what PC is running the service, so one PC may get ports 11xxx, another 12xxx and so on.
Internal ipv6 IPs stay constant too, and you should use link-local or site-local addresses for your ipv6 services just as you use RFC1918 addresses for ipv4 services.
Well, I didn't know that. This is good, but having two or three IPs per host on the same network makes configuration a bit difficult, especially since (again, I don't know how v6 handles this) if DNS returns all addresses, the host may pick the public one to connect to an internal host.
Still, to me it seems more of a hack than each host having a single IP address and entire LAN having NAT. Especially since v6 IPs are harder to remember. Also, those applications that send the IP address as payload are badly designed. The IP address already is in the header of the packet, why send a copy of it in the data section? Yes, FTP does this wrong, though I used a few FTP servers and they could find out the external IP by either DNS or some other method. IM, like Skype, already just works, I just need to forward the required port (the same amount of configuration as would be to allow that port in the firewall).
But, they will each ALSO have a locally assigned, non-routable ("site-local") address that you can use as an unchanging address on your LAN.
And how will they choose which IP to use? If all computers in my network have two IPs then they'll probably use one of them randomly, which means that some established internal connections will break after the public IP changes.
I'll probably never understand the need for every device to have a public IP, or use services at default ports. I have a HTTP server, but my port 80 is not forwarded to anything, some other port goes to the HTTP server.
Comcast is going to poll every assignned/64 to see how many of the 2^64 addresses are in use -- you're pretty much free to do as you like with your block of addresses.
They don't need to, they can just count different source IPs in your subnet.
With NAT, I can keep my internal IPs constant even if I switch ISPs (be it a backup connection or whatever). Only the NAT router needs to know that the external IP changed, other PCs usually don't care. And internal IPs stay constant.
IIRC, BGP is a routing protocol. How would it help with keeping internal IPs constant? I don't think any of the ISPs would allow me to rewrite their routing tables so I keep the same IPs no matter which ISP I am currently using.
In short: IPv4+NAT allows me to do this without any help (or knowledge) from the ISPs. It also works quite well. I have only ~6 ports forwarded, so there are a lot still left.
Me too. I look forward to having no NAT and changing the IPs in my internal network every time I use a different ISP.
"Hmm, my internet connection failed, better connect the backup one. OK, now this ISP gives me xxx:yyy:zzz:xxyz::0 IP, so I now have to go and change the addresses of all my PCs, since they won't be able to access the internet. If only there could be some way to keep the internal IPs constant..."
Currently, the internal IPs of my computers do not depend on which ISP I am connected to.
Last time I checked, you need to enter the IP address of the DNS server. 8.8.8.8 (or even the IPs of the DNS of my ISP) is easy to remember, v6 addresses are not.
I do not dedicate drives to music, videos and so on. Doing this would result in a situation where a drive (let's say "music") has some free space but I cannot use it because I want to write a different type of file (let's say "video").
Sometimes a GUI is easier for very simple operations like you mention, however, as others have exemplified further up and down the thread, when complexity is added, the CLI is much faster and easier.
Yes, I never said that GUI is the answer to everything. However it does make simple operations easier, while the CLI is still there for complex operations. Also, the vast majority of operations on a computer are the simple ones. I have moved or copied files between hard drives a lot of times, but I only needed to use a CLI once.
Same thing with configuration - changing the IP address of a Windows PC is easy using GUI (and probably CLI) and I do not have to remember the exact command that I would need to use. If I want to do some complex operation I'll probably have to RTFM or google and then I can also use CLI, but I don't want to remember some long command for a simple task.
That falls apart when you want to download some video file, but your "video" drive is full, even though you have 50GB left in the "music" drive. You either have to delete some other video files to make the space or put some of the video files in the music drive.
That's why I have some part of my music/video/whatever on every drive. In this case I can put the new file to whatever drive that has free space.
Also, Windows too has this feature. While you have to have at least one drive letter (C:\ or whatever), if the file system is NTFS you can mount a hard drive as a folder. So, I could make C:\movies be a separate drive from C:\.
Well, Windows can do RAID too, so that's no different. I can have drive R:\ that's a huge RAID array, or I can have/RAID (or/home or whatever) on Linux.
As for the mount points, they can be anything, well, let's see... Oh, I know, how about this -/movies1,/movies2,/movies3...? If I have 10 hard drives (100-750GB) and use all of them for data and not system files, what do you think the mount points should be?
ls - lists the contents of a folder, ls -la lists them with permissions and file size (does not display folder size). head - shows first few lines of a file (or of another programs output). mv - move cut - can show the middle of a file.
I don't even know how to check free space on Linux. "dir" on Windows shows the free space, but in my case it would be something like...
dir x:\movies/p *great, I think I can move this 700MB file to some other drive* dir e: *200MB left, nope* dir h: *500MB left, nope* dir j: *100MB left, nope* dir m: *1GB left, OK* move x:\movies\SomeMovie.xvid.ac3.avi m:\movies
So when you add a new hard drive and your stuff installed on C: is now on D:, your registry is pointing to a file that doesn't exist.
When I install a new drive, I choose the letter for it. Windows usually take the first available one, but I might be using a network drive with that letter, so I go to Computer Management and change the letter to whichever I want. Windows NT/2000/XP (and possibly Vista/7) do not change the drive letters of existing drives when installing a new one. Windows 2003 does not even assign a drive letter until I assign it manually.
Really, the whole C:, D:, E: thing was fucking moronic. The Unix scheme of mounting seamlessly as a folder under/, in a way which renders the underlying volumes invisible is far saner
Does it also magically make 10 hard drives appear as one, so you don't have to move files between drives to make enough free space for one bigger file?
Windows also has that - it's called RAID and can be used with a hardware controller or in software. Neither version works if the hard drives are not identical (I have drives from 100GB up to 750GB), on different computers and were not bought (or installed) at the same time, but oer the years as I ran out of space.
So, on Windows it's C:\, D:\, E:\,... Z:\, on Linux it would be/mnt/hda1,/mnt/hdb1,/mnt/smb1,.../mnt/smb5
(and/mnt/hda1 is longer than C:\)
2) Tab completion. I don't think I've ever typed more than 5 letters of a filename, of any length.
Which is good, but still less useful as GUI. A very common problem to me is that I want to download or copy some big file and while I have enough free space combines, none of the drives has enough space for that file, so I need to move some smaller files to make enough space for the big one. I can use "dir" to see the files in each folder, find the ones I can move then use "dir" multiple times to find out which drive has enough free space for those files then move them.
Or, I can open "My Computer" and see the free space of each drive, then go to some folder to look for files that can be moved or use some GUI tool like SpaceMonger to see what are the largest files and folder in the whole drive at once. This way is much less typing and much faster.
move "Some TV show S02E04.mkv" "E:\TV Shows\Some TV Show\Season2\"
I would not want to have to type that to just copy a file.
As far as copying files goes, the only time I used a command line copy tool was when I wanted to have my MP3 collection on one hard drive, so I had to copy from *:\MP3 to x:\MP3, I made a.bat file to do it, so I would not have to start copying from each hard drive separately.
Other times it's a single file, a single folder or a bunch of files with similar names (foe example, all episodes of "Some TV Show".
ever listen to music while say browsing the web? yes, dual cores help with things like that.
Yes, they would. After all, decoding 128kbps stereo MP3 would take more than 50% of a 486 100MHz, you definitely need one more core to browse the web.
Oh wait, on a 1.6GHz Turion64 CPU, playing MP3 takes probably 1% of CPU, never mind...
you might not *need* dual cores now, but it won't hurt and will almost certainly help in the future.
However, in the future, when the user actually needs two cores, they will be much cheaper than now. Quad cores too.
that's kind of like choosing an equally priced SDTV over an HDTV because well you don't have HDTV service, right now.
Well, if both TVs are equally priced and differ only in the resolution, then sure, there is no point in buying the lower resolution TV. Just as there is no point in buying a new TV if some transistor (cost: $1) in the old one burned out. Just replace the transistor and use the saved money elsewhere. Unless you do want a new TV, in which case you can just buy it and not even wait for your old TV to fail.
Slashdot Mirror
Ok, I'll bite. Which God?
The one that Jews believe in? The one that Muslims believe in? The one that Christians believe in (who is kinda similar to the Jewish one)? The gods that ancient Greeks believed in? The gods that ancient Egyptians believed in? Whatever Scientologists believe in?
There are a lot of religions, most of them are mutually exclusive ("this religion is the only one true religion and our God(s) is/are the only true God(s), everybody else is wrong"). So, if the fact that people believed in something made it true, then all those religions should be true, but at the same time all would be false. So, which one religion is the only one true religion and how do you prove that?
Also, a lot of people believed (and some still do) that the Earth was flat, that didn't make it true.
Do you have any power-saving features enabled at all ?
CPU frequency throttling makes sound crackle when the frequency is changed, which happens a lot when I am watching a movie, so I disabled it. I also disabled turning HDDs off. In short - no.
Also, your server probably does not have one of the most power-hungry video cards (well, it now has been outdone by GTX470 and dual GPU cards).
I thought 250W was kind of normal for a two socket PC, since the CPUs are rated 85W each, so together they draw 170W, the video card at idle also draws some power and the PSU and all other voltage regulators are not 100% efficient.
CPU: 2x AMD Opteron 270 (dual core 2GHz);
MB: Tyan Thunder K8WE;
RAM: 4x512MB + 2x 1GB DDR1 400MHz;
VGA: 1x ATI HD2900XT;
HDD: 2x 15kRPM, 1x 5.4kRPM.
Power consumption as indicated by my UPS (I also have a few other computers plugged in):
This PC turned off, (monitor is on): ~500W
This PC turned on and idle, CPUs at full speed: ~750W
This PC turned on and I am playing a game: 900W
So, with CPUs idle, my main PC uses 250W, when playing a GPU and CPU intensive game it's 400W.
In Lithuania it's yyyy mm dd or yyyy-mm-dd. Makes the most sense, slowest changing number is the first and fastest changing number is last, like in regular numbers: 100 -> 101 -> 102 ...
The point is not that these relays generate no heat, but that they can survive up to 500C, while modern CPUs survive about 100C, but start working incorrectly. Discrete transistors and low density logic chips (like the 4xxx series) survive up to 125C. Assuming the power dissipation and ambient temperature is the same, you need much smaller heatsink to keep the device at 400C than you need to keep it at 60C.
Atom at 500kHz would probably use less power, it still would melt if the ambient temperature was 150C.
But when you need to enter it, you cannot google it.
well, it doesn't say "Full" HD. So it probably was labeled HD because it's 16:9 or because the resolution is higher than SDTV (720x576).
Depends on what you compare it to.
Common 4:3 resolutions are 1024x768(low), 1280x1024 (technically 5:4, but a lot of 4:3 LCDs have it), 1600x1200 and 1920x1440.
16:10 resolutions usually are 1280x800, 1680x1050, 1920x1200.
My 21" CRT can display up to 2048x1536 (75Hz) or 1920x1440 (85Hz), but I usually use 1600x1200, so a big LCD with 1920x1200 resolution would have the same amount of vertical pixels as my regular resolution and less than the max resolution I use. Even though it would have bigger screen.
I don't know, but skype usually says "less than 1kB/s download, less than 1kB/s upload" I have seen it grow to ~5KB/s once.
Also, if I do not forward the port and send/receive file from somebody who hasn't either, then Skype says that my transfer "is being relayed" and the speed is never over 5KB/s.
Skype has a different solution: When you actually have reasonably fast direct internet access while running Skype, you act as a supernode. This means that you act as a proxy when two NAT-using persons want to talk to each other. If you pay for your data transfer (and you always do, even when it's indirectly), you pay for two persons you don't even know being able to audio/video chat.
So, it works even without forwarding the port. I have forwarded the port and I don't really care that 10KB/s or whatever is used to relay the data of two NAT-using people, after all, I seed torrents at 4-7MB/s.
As for everything else, I now have to agree, v6 would no worse than v4 for configuration, still there's the problem that old devices do not support it, but maybe someone will figure out some kind of NAT so that v4 device can talk to v6 device.
You can have multiple machines running ssh or web servers without screwing around with forwarding different nonstandard ports on the firewall/nat box through to the different internal machines.
That may be useful to some, but I always assign non standard ports. I have a HTTP server, but port 0 is not forwarded. Even when I had one computer, I used nonstandard ports. Now I assign ports based on what PC is running the service, so one PC may get ports 11xxx, another 12xxx and so on.
Internal ipv6 IPs stay constant too, and you should use link-local or site-local addresses for your ipv6 services just as you use RFC1918 addresses for ipv4 services.
Well, I didn't know that. This is good, but having two or three IPs per host on the same network makes configuration a bit difficult, especially since (again, I don't know how v6 handles this) if DNS returns all addresses, the host may pick the public one to connect to an internal host.
Still, to me it seems more of a hack than each host having a single IP address and entire LAN having NAT. Especially since v6 IPs are harder to remember. Also, those applications that send the IP address as payload are badly designed. The IP address already is in the header of the packet, why send a copy of it in the data section? Yes, FTP does this wrong, though I used a few FTP servers and they could find out the external IP by either DNS or some other method. IM, like Skype, already just works, I just need to forward the required port (the same amount of configuration as would be to allow that port in the firewall).
But, they will each ALSO have a locally assigned, non-routable ("site-local") address that you can use as an unchanging address on your LAN.
And how will they choose which IP to use? If all computers in my network have two IPs then they'll probably use one of them randomly, which means that some established internal connections will break after the public IP changes.
I'll probably never understand the need for every device to have a public IP, or use services at default ports. I have a HTTP server, but my port 80 is not forwarded to anything, some other port goes to the HTTP server.
Comcast is going to poll every assignned /64 to see how many of the 2^64 addresses are in use -- you're pretty much free to do as you like with your block of addresses.
They don't need to, they can just count different source IPs in your subnet.
OK, how about this:
With NAT, I can keep my internal IPs constant even if I switch ISPs (be it a backup connection or whatever). Only the NAT router needs to know that the external IP changed, other PCs usually don't care. And internal IPs stay constant.
IIRC, BGP is a routing protocol. How would it help with keeping internal IPs constant? I don't think any of the ISPs would allow me to rewrite their routing tables so I keep the same IPs no matter which ISP I am currently using.
In short: IPv4+NAT allows me to do this without any help (or knowledge) from the ISPs. It also works quite well. I have only ~6 ports forwarded, so there are a lot still left.
Me too. I look forward to having no NAT and changing the IPs in my internal network every time I use a different ISP.
"Hmm, my internet connection failed, better connect the backup one. OK, now this ISP gives me xxx:yyy:zzz:xxyz::0 IP, so I now have to go and change the addresses of all my PCs, since they won't be able to access the internet. If only there could be some way to keep the internal IPs constant..."
Currently, the internal IPs of my computers do not depend on which ISP I am connected to.
Last time I checked, you need to enter the IP address of the DNS server. 8.8.8.8 (or even the IPs of the DNS of my ISP) is easy to remember, v6 addresses are not.
I do not dedicate drives to music, videos and so on. Doing this would result in a situation where a drive (let's say "music") has some free space but I cannot use it because I want to write a different type of file (let's say "video").
Sometimes a GUI is easier for very simple operations like you mention, however, as others have exemplified further up and down the thread, when complexity is added, the CLI is much faster and easier.
Yes, I never said that GUI is the answer to everything. However it does make simple operations easier, while the CLI is still there for complex operations. Also, the vast majority of operations on a computer are the simple ones. I have moved or copied files between hard drives a lot of times, but I only needed to use a CLI once.
Same thing with configuration - changing the IP address of a Windows PC is easy using GUI (and probably CLI) and I do not have to remember the exact command that I would need to use. If I want to do some complex operation I'll probably have to RTFM or google and then I can also use CLI, but I don't want to remember some long command for a simple task.
That falls apart when you want to download some video file, but your "video" drive is full, even though you have 50GB left in the "music" drive. You either have to delete some other video files to make the space or put some of the video files in the music drive.
That's why I have some part of my music/video/whatever on every drive. In this case I can put the new file to whatever drive that has free space.
Also, Windows too has this feature. While you have to have at least one drive letter (C:\ or whatever), if the file system is NTFS you can mount a hard drive as a folder. So, I could make C:\movies be a separate drive from C:\.
Well, Windows can do RAID too, so that's no different. I can have drive R:\ that's a huge RAID array, or I can have /RAID (or /home or whatever) on Linux.
As for the mount points, they can be anything, well, let's see... Oh, I know, how about this - /movies1, /movies2, /movies3 ...? If I have 10 hard drives (100-750GB) and use all of them for data and not system files, what do you think the mount points should be?
IIRC -
ls - lists the contents of a folder, ls -la lists them with permissions and file size (does not display folder size).
head - shows first few lines of a file (or of another programs output).
mv - move
cut - can show the middle of a file.
I don't even know how to check free space on Linux. "dir" on Windows shows the free space, but in my case it would be something like...
dir x:\movies /p
*great, I think I can move this 700MB file to some other drive*
dir e:
*200MB left, nope*
dir h:
*500MB left, nope*
dir j:
*100MB left, nope*
dir m:
*1GB left, OK*
move x:\movies\SomeMovie.xvid.ac3.avi m:\movies
this way is slower.
So when you add a new hard drive and your stuff installed on C: is now on D:, your registry is pointing to a file that doesn't exist.
When I install a new drive, I choose the letter for it. Windows usually take the first available one, but I might be using a network drive with that letter, so I go to Computer Management and change the letter to whichever I want. Windows NT/2000/XP (and possibly Vista/7) do not change the drive letters of existing drives when installing a new one. Windows 2003 does not even assign a drive letter until I assign it manually.
Really, the whole C:, D:, E: thing was fucking moronic. The Unix scheme of mounting seamlessly as a folder under /, in a way which renders the underlying volumes invisible is far saner
Does it also magically make 10 hard drives appear as one, so you don't have to move files between drives to make enough free space for one bigger file?
Windows also has that - it's called RAID and can be used with a hardware controller or in software. Neither version works if the hard drives are not identical (I have drives from 100GB up to 750GB), on different computers and were not bought (or installed) at the same time, but oer the years as I ran out of space.
So, on Windows it's C:\, D:\, E:\, ... Z:\, on Linux it would be /mnt/hda1, /mnt/hdb1, /mnt/smb1, ... /mnt/smb5
(and /mnt/hda1 is longer than C:\)
2) Tab completion. I don't think I've ever typed more than 5 letters of a filename, of any length.
Which is good, but still less useful as GUI. A very common problem to me is that I want to download or copy some big file and while I have enough free space combines, none of the drives has enough space for that file, so I need to move some smaller files to make enough space for the big one. I can use "dir" to see the files in each folder, find the ones I can move then use "dir" multiple times to find out which drive has enough free space for those files then move them.
Or, I can open "My Computer" and see the free space of each drive, then go to some folder to look for files that can be moved or use some GUI tool like SpaceMonger to see what are the largest files and folder in the whole drive at once. This way is much less typing and much faster.
GUI lets me have longer file and directory names.
move "Some TV show S02E04.mkv" "E:\TV Shows\Some TV Show\Season2\"
I would not want to have to type that to just copy a file.
As far as copying files goes, the only time I used a command line copy tool was when I wanted to have my MP3 collection on one hard drive, so I had to copy from *:\MP3 to x:\MP3, I made a .bat file to do it, so I would not have to start copying from each hard drive separately.
Other times it's a single file, a single folder or a bunch of files with similar names (foe example, all episodes of "Some TV Show".
ever listen to music while say browsing the web? yes, dual cores help with things like that.
Yes, they would. After all, decoding 128kbps stereo MP3 would take more than 50% of a 486 100MHz, you definitely need one more core to browse the web.
Oh wait, on a 1.6GHz Turion64 CPU, playing MP3 takes probably 1% of CPU, never mind...
you might not *need* dual cores now, but it won't hurt and will almost certainly help in the future.
However, in the future, when the user actually needs two cores, they will be much cheaper than now. Quad cores too.
that's kind of like choosing an equally priced SDTV over an HDTV because well you don't have HDTV service, right now.
Well, if both TVs are equally priced and differ only in the resolution, then sure, there is no point in buying the lower resolution TV. Just as there is no point in buying a new TV if some transistor (cost: $1) in the old one burned out. Just replace the transistor and use the saved money elsewhere. Unless you do want a new TV, in which case you can just buy it and not even wait for your old TV to fail.