Depending on the quality of the ISP their DNS servers will refer to the root servers but this is an unlikely scenario.
Of course not all DNS queries will result in a query sent to a root DNS server. But some of them will need to go all the way to the roots. The root DNS servers are the only servers authoritative for "." and you can only cache lookups for a limited amount of time. So no matter how many different DNS servers can use the same response, eventually it will expire, and by then they must query the roots again. In addition to that all queries for nonexisting top level domains must be send to a root DNS server, unless of course someone recently queried the same nonexisting domain. Actually I think trying to lookup a local hostname that can be found in your DNS search path will result in an attempt to lookup that as a top level domain name first. So when my/etc/resolv.cof says "search daimi.au.dk" and I do a lookup of the hostname kalashnikov, this will result in an attempt to lookup a top level domain named kalashnikov, and only when that fails it will lookup kalashnikov.daimi.au.dk.
Clearly the growing internet does not cause the load on the root DNS servers to decrease. Quite the contrary, it causes more load on the root DNS servers. But fortunately caching and forwarding queries to other DNS servers prevents the load from increasing linearily in the size of the internet. But considering how the default for some DNS software is to queriy the root servers directly when something is not in its own cache, you can still expect a significant load on the root DNS servers. I changed my own caching DNS server from doing that to just sending every query to my provider's DNS server.
Gee, get your quotes right. I know this quote (because I use it in my.signature when I post on usenet). The exact words are: Would you like fries with that? and they are from a network driver in Linux:
sunhme.c
Way more frequently used that SysReq or Scroll Lock.
I use SysRq sometimes. Unfortunately I have multiple keyboards where SysRq doesn't work reliably. Just because some systems doesn't use the key, they shouldn't start producing keyboards where the key doesn't work.
The Scroll Lock key is the most functionally overloaded key on my keyboard. It is used as hotkey for my KVM switch and it is also used to stop scrolling. It is really annoying when you start shutting down one computer, and as you are switching to another you accidentially stop scrolling and thereby preventing the computer from shutting down.
The Windows keys are some of the least used keys on my keyboard, they really serves no purpose at all. And BTW I don't want a single key to do Ctrl+C - killing the running program isn't something you want to do by accident.
The DNS servers aren't being spread across the world, there are just plain more of them.
It is about root DNS servers, not DNS servers in general. There were a good technical reason for having no more than thirteen root DNS servers. IIRC the reason was that this was the largest number you could fit into a single UDP packet. And until people started making workarounds, the number of root DNS servers did not change.
Then, I believe, users clamoured for more convenience, so keyboard manufacturers started duplicating the ctrl and alt keys (and/or moved the del key) so that the salute could be performed one-handed.
Actually some keyboards have gone a bit too far along that road. I have seen a keyboard, where you could do CTRL+ALT+DEL with just a single keypress. Yup, they really added an extra key to do just that. Some people must have used CTRL+ALT+DEL too much.
The number of countries outside the U.S. outnumber the number of countries inside the U.S.
I would have moderated parent Insightful, but unfortunately it was already on the limit. The US is not all the world, in fact it is not even one third of the world. My first thought when I read the head line was: "That was about time". We don't want too much control over the internt to lie in the hands of one country, no matter which country it is. I know the way DNS is designed it is still possible to mess up the system with control over less than one third of the root DNS servers, but so far the problems we have seen with root DNS servers were DoS attacks, not malicious responses. So I think spreading the root DNS servers across the world is a good start.
Good point. But then you must need to store your key somewhere. Actually each archive you create should contain the secret key encrypted under your password, because you don't want to eventually lose your secret key and then be unable to decrypt your backup copies. So on your harddisk you must keep the encrypted secret key along with the public key. Could you explain in a litle more detail how you do this? And is that the same as PKZip does?
I doubt that PKZip is based only on RSA. RSA is an asymmetric encryption. For some purposes this is nice, but it is inefficient. For that reason you almost always use asymmetric encryption together with a symmetric encryption. You generate a one time symmetric encryption key. The data is encrypted with the symmetric key, typically in CBC or CFB mode. Then only the symmetric encryption key is encrypted asymmetrically, which means much better speed.
Actually I think this is one of the cases, where there is no need for asymmetric encryption at all. So AES sounds like a better idea. Can anybody explain why PKZip use RSA? And which symmetric cipher is it combined with?
Most moniters (as far as I know) have to use the same resolution for the whole screen.
The vertical resolution have to be the same everywhere. The horisontal doesn't have to. Most GFX chips does however impose limitations on the horisontal resolution. What the Amiga could do was to change the horisontal resolution so it didn't have to be the same on every horisontal line, it could also change other stuff including palette and mouse position. The price was one empty horisontal line where the change happened. But this empty line was due to limitations in the OS software. With special tricks in software you could do better. AFAIK the two player version of Lemmings changed the image position in the midle of the screen, so the left half and right half could scroll independently.
Given that in this case the GPL'd code was not for sale or under any other commercial license
It could be. There are dual licensed projects. And if you have written some code and released it under GPL, you still have the right to release it under any other license you would like to. This of course requires that you have written all of it yourself, or at least the acceptance by every contributor. The larger a project, the more difficult it will be to get every contributor accept a sale under a different license. But in fact the possibility exists, even if there isn't any kind of standard price from previous sales.
What are the damages you claim when you give away your software for free?
That is a good question. But if you by that argumentation can avoid paying damages, then how about this one: All the people copying music and software for their own personal use shouldn't pay damages either. Because they are copying stuff they would never have bought if they had to pay for it. Thuss there are no damages here.
If you think about it, there isn't much difference. Both cases you can buy/make the copies legally if you comply with a license. And in both cases we are dealing with people who would never comply with that license, and (illegaly) make copies anyway.
I will argue, that there are in fact damages, even though the software being copied could have been gotten for free. Why? I have indirect costs from the widespread usage of closed source software, in particular Windows. Illegal copies of free software will weaken it's position in the market and to some extent strengthen closed source software. Here are a list of indirect costs:
I have to pay for software I don't need when I buy a new computer.
I have to spend a lot of time dealing with people using me using Linux as a bad excuse for not wanting to fix their own systems.
I have to spend a lot of time dealing with hardware, either because of missing drivers, defective hardware that the supplier refuse to repair because I run Linux.
I sometimes have costs of buying extra hardware because of previously mentioned problems.
Right now I'm facing the need of switching to a different bank because the current one will not fix their home banking system. That is not without cost to me.
The value of my own software gets smaller. Since I decide to give the software to a lot of people, who are also affected by the problem, the total damage is larger. The fact that I give some of the value I earn to other people shouldn't count as a reason for the company not to pay damages.
And finally the software does have a price (free means freedom, not that you don't have to pay), I expect to get something in return for the software, that is full source for the modifications. As Linus pointed out, the value of expecting to get software in return is explicitly mentioned by copyright law in some country.
Replace the core of their system with a Unix system like MacOS did.
If that is too much alternatively they can just vanish, I wouldn't miss them. And BTW given their history, no matter what their next step is, I won't trust them.
... i'm really eager to see what is going to happen when the GPL goes to court.
Some time ago I attended a talk given by a Danish lawyer, who said he would love to be the first to take such a case to court. And in a case like this the first thing he would do if KISS Technology refuses to publish the code would be to have the bailiff go and pick it up.
It is exactly six minutes and fourty one seconds into the file. Only reason I could find it so quickly is, that I heard the entire file just before I read the story on slashdot.
The quoted translation is more or less correct. I listened very carefully to his words in the published.rm file starting at 6:41 and ending at 7:02. (Ironically I couldn't make mplayer play the file, so I used Real Player instead). Here is my translation (with help from Gyldendals Rode Ordboger):
We don't use mplayer, we have our own player. Simply a player like we know Real Player - Microsoft Player, the application that does that video is presented. And that is of course a completely fundamental thing for our player because that is what made us known, that we are able to replay a number of formats.
And I think what he is trying to tell us is, that they have a program similar in function to Real Player and Windows Media Player.
What's with 174? I don't recall anything special about 174. And about 192, some people think it is resrved, but in fact only 192.168.0.0/16 is reserved by
RFC 1918. All the people not understanding RFC 1918 of course cause some problems for everybody with an IP adress in 192.0.0.0/24.
224-255 are for multicast.
In fact only 224-239 are for multicast. 240-254 are reserved for future extensions, and 255 is reserved for something I don't remember.
Taking reserved addresses into account means it is more like 1/221st of the address space. Only 1-223 in the first octet are used for host addresses of these 10 and 127 are reserved for special purposes.
I'm curious to know what are those 16 bytes in the ROM...must have been highly optimized:)
The trick is, that when writing an emulator, you don't need to write a BIOS in 16 bit code. Instead a BIOS implementation is written in 32 bit code, that can execute outside the precious low end address space. Then I just need enough entry points from 16 bit code to 32 bit code. An entry points requires an instruction that will trap from virtual 86 mode to 32 bit user mode. I decided to use the HLT instruction which is only one byte. Because of the segment:offset addressing there are 4096 different ways to address this single byte, that means I have 4096 entry points which is a lot more than I need. The entry point for reboot is sometimes accessed through at least two different addressings, so I avoided to place my HLT instruction there and instead placed the conventional five bytes long far jump instruction there, which jumps to one of my entry points. After this five bytes instruction are eight bytes reserved for the BIOS date written as month / date / year. The last three bytes are three single byte instructions HLT IRET RETF. The HLT and IRET are actually used, the RETF I just placed there because it might come in handy. Because of the DOS memory management and the reboot entry point, there is no way to make the ROM smaller than 16 bytes.
To actually protect the ROM against writing I mark the entire page read only, though it is only the last 16 bytes I really need to protect. This means any write to the first 4050 bytes of this page will trap, those are the traps that I needed to fix in the kernel because they would Oops if triggered by a stack access by an instruction emulated in the kernel. All those traps of course slows down execution, so I might want to sacrifice the last 4050 bytes for a bit of performance. I'm still looking for an efficient way to access the last bytes. If I could put an upper limit to the address accessed by virtual 86 mode, I could switch between a limit just below the ROM and a write protected page, which I belive would speed up execution. Together with my emulator I have put a GPL'ed UMM driver that works with my emulator, quite conveniently this driver does not support the last 4050 bytes of UMB that have been causing problems anyway. EMM386 doesn't work with my emulator, and never will because of braindead Intel design.
Actually I have written my own PC emulator, but it is far from as usable as DOSemu. I wanted to test a way to do the emulation with only 16 bytes used for ROM. As long as it was fun I kept coding. But eventually I ran into some problems. If I actually wanted to use all the available 255KB of UMB the kernel would Oops when the stack was on the same page as my ROM. I fixed the kernel bugs together with Manfred Spraul and Stas Sergeev. But I never got back to coding on my emulator.
Slashdot Mirror
Depending on the quality of the ISP their DNS servers will refer to the root servers but this is an unlikely scenario.
/etc/resolv.cof says "search daimi.au.dk" and I do a lookup of the hostname kalashnikov, this will result in an attempt to lookup a top level domain named kalashnikov, and only when that fails it will lookup kalashnikov.daimi.au.dk.
Of course not all DNS queries will result in a query sent to a root DNS server. But some of them will need to go all the way to the roots. The root DNS servers are the only servers authoritative for "." and you can only cache lookups for a limited amount of time. So no matter how many different DNS servers can use the same response, eventually it will expire, and by then they must query the roots again. In addition to that all queries for nonexisting top level domains must be send to a root DNS server, unless of course someone recently queried the same nonexisting domain. Actually I think trying to lookup a local hostname that can be found in your DNS search path will result in an attempt to lookup that as a top level domain name first. So when my
Clearly the growing internet does not cause the load on the root DNS servers to decrease. Quite the contrary, it causes more load on the root DNS servers. But fortunately caching and forwarding queries to other DNS servers prevents the load from increasing linearily in the size of the internet. But considering how the default for some DNS software is to queriy the root servers directly when something is not in its own cache, you can still expect a significant load on the root DNS servers. I changed my own caching DNS server from doing that to just sending every query to my provider's DNS server.
Clerk: do you want any frees with that?
.signature when I post on usenet). The exact words are: Would you like fries with that? and they are from a network driver in Linux:
sunhme.c
Gee, get your quotes right. I know this quote (because I use it in my
Way more frequently used that SysReq or Scroll Lock.
I use SysRq sometimes. Unfortunately I have multiple keyboards where SysRq doesn't work reliably. Just because some systems doesn't use the key, they shouldn't start producing keyboards where the key doesn't work.
The Scroll Lock key is the most functionally overloaded key on my keyboard. It is used as hotkey for my KVM switch and it is also used to stop scrolling. It is really annoying when you start shutting down one computer, and as you are switching to another you accidentially stop scrolling and thereby preventing the computer from shutting down.
The Windows keys are some of the least used keys on my keyboard, they really serves no purpose at all. And BTW I don't want a single key to do Ctrl+C - killing the running program isn't something you want to do by accident.
The DNS servers aren't being spread across the world, there are just plain more of them.
It is about root DNS servers, not DNS servers in general. There were a good technical reason for having no more than thirteen root DNS servers. IIRC the reason was that this was the largest number you could fit into a single UDP packet. And until people started making workarounds, the number of root DNS servers did not change.
Then, I believe, users clamoured for more convenience, so keyboard manufacturers started duplicating the ctrl and alt keys (and/or moved the del key) so that the salute could be performed one-handed.
Actually some keyboards have gone a bit too far along that road. I have seen a keyboard, where you could do CTRL+ALT+DEL with just a single keypress. Yup, they really added an extra key to do just that. Some people must have used CTRL+ALT+DEL too much.
The number of countries outside the U.S. outnumber the number of countries inside the U.S.
I would have moderated parent Insightful, but unfortunately it was already on the limit. The US is not all the world, in fact it is not even one third of the world. My first thought when I read the head line was: "That was about time". We don't want too much control over the internt to lie in the hands of one country, no matter which country it is. I know the way DNS is designed it is still possible to mess up the system with control over less than one third of the root DNS servers, but so far the problems we have seen with root DNS servers were DoS attacks, not malicious responses. So I think spreading the root DNS servers across the world is a good start.
I always use asymmetric crypto in my backups.
Good point. But then you must need to store your key somewhere. Actually each archive you create should contain the secret key encrypted under your password, because you don't want to eventually lose your secret key and then be unable to decrypt your backup copies. So on your harddisk you must keep the encrypted secret key along with the public key. Could you explain in a litle more detail how you do this? And is that the same as PKZip does?
I doubt that PKZip is based only on RSA. RSA is an asymmetric encryption. For some purposes this is nice, but it is inefficient. For that reason you almost always use asymmetric encryption together with a symmetric encryption. You generate a one time symmetric encryption key. The data is encrypted with the symmetric key, typically in CBC or CFB mode. Then only the symmetric encryption key is encrypted asymmetrically, which means much better speed.
Actually I think this is one of the cases, where there is no need for asymmetric encryption at all. So AES sounds like a better idea. Can anybody explain why PKZip use RSA? And which symmetric cipher is it combined with?
Most moniters (as far as I know) have to use the same resolution for the whole screen.
The vertical resolution have to be the same everywhere. The horisontal doesn't have to. Most GFX chips does however impose limitations on the horisontal resolution. What the Amiga could do was to change the horisontal resolution so it didn't have to be the same on every horisontal line, it could also change other stuff including palette and mouse position. The price was one empty horisontal line where the change happened. But this empty line was due to limitations in the OS software. With special tricks in software you could do better. AFAIK the two player version of Lemmings changed the image position in the midle of the screen, so the left half and right half could scroll independently.
\.(com|org|net)
Oh I would love all spammers to do like that.
Given that in this case the GPL'd code was not for sale or under any other commercial license
It could be. There are dual licensed projects. And if you have written some code and released it under GPL, you still have the right to release it under any other license you would like to. This of course requires that you have written all of it yourself, or at least the acceptance by every contributor. The larger a project, the more difficult it will be to get every contributor accept a sale under a different license. But in fact the possibility exists, even if there isn't any kind of standard price from previous sales.
SCO, in the attempt to protect its trademark
Who's trademark?
That is a good question. But if you by that argumentation can avoid paying damages, then how about this one: All the people copying music and software for their own personal use shouldn't pay damages either. Because they are copying stuff they would never have bought if they had to pay for it. Thuss there are no damages here.
If you think about it, there isn't much difference. Both cases you can buy/make the copies legally if you comply with a license. And in both cases we are dealing with people who would never comply with that license, and (illegaly) make copies anyway.
I will argue, that there are in fact damages, even though the software being copied could have been gotten for free. Why? I have indirect costs from the widespread usage of closed source software, in particular Windows. Illegal copies of free software will weaken it's position in the market and to some extent strengthen closed source software. Here are a list of indirect costs:
- I have to pay for software I don't need when I buy a new computer.
- I have to spend a lot of time dealing with people using me using Linux as a bad excuse for not wanting to fix their own systems.
- I have to spend a lot of time dealing with hardware, either because of missing drivers, defective hardware that the supplier refuse to repair because I run Linux.
- I sometimes have costs of buying extra hardware because of previously mentioned problems.
- Right now I'm facing the need of switching to a different bank because the current one will not fix their home banking system. That is not without cost to me.
- The value of my own software gets smaller. Since I decide to give the software to a lot of people, who are also affected by the problem, the total damage is larger. The fact that I give some of the value I earn to other people shouldn't count as a reason for the company not to pay damages.
And finally the software does have a price (free means freedom, not that you don't have to pay), I expect to get something in return for the software, that is full source for the modifications. As Linus pointed out, the value of expecting to get software in return is explicitly mentioned by copyright law in some country.Because this time, the fake-patenters got overzealous and attacked someone who actually has the legal resources to fight back.
Are you talking about SCO?
- Comply with standards
- Open the source
- Replace the core of their system with a Unix system like MacOS did.
If that is too much alternatively they can just vanish, I wouldn't miss them. And BTW given their history, no matter what their next step is, I won't trust them.Did you know that becuae your name "/Wegge" has a slash in it it seems to be impossible to visit your user page.
There is a way.
... i'm really eager to see what is going to happen when the GPL goes to court.
Some time ago I attended a talk given by a Danish lawyer, who said he would love to be the first to take such a case to court. And in a case like this the first thing he would do if KISS Technology refuses to publish the code would be to have the bailiff go and pick it up.
Didn't listen that far??
It is exactly six minutes and fourty one seconds into the file. Only reason I could find it so quickly is, that I heard the entire file just before I read the story on slashdot.
And 174 and 192...
What's with 174? I don't recall anything special about 174. And about 192, some people think it is resrved, but in fact only 192.168.0.0/16 is reserved by RFC 1918. All the people not understanding RFC 1918 of course cause some problems for everybody with an IP adress in 192.0.0.0/24.
224-255 are for multicast.
In fact only 224-239 are for multicast. 240-254 are reserved for future extensions, and 255 is reserved for something I don't remember.
Taking reserved addresses into account means it is more like 1/221st of the address space. Only 1-223 in the first octet are used for host addresses of these 10 and 127 are reserved for special purposes.
2038 is a valid concern. But if DNS servers compare serial numbers according to RFC 1982 it is not going to be a problem.
I'm curious to know what are those 16 bytes in the ROM...must have been highly optimized :)
The trick is, that when writing an emulator, you don't need to write a BIOS in 16 bit code. Instead a BIOS implementation is written in 32 bit code, that can execute outside the precious low end address space. Then I just need enough entry points from 16 bit code to 32 bit code. An entry points requires an instruction that will trap from virtual 86 mode to 32 bit user mode. I decided to use the HLT instruction which is only one byte. Because of the segment:offset addressing there are 4096 different ways to address this single byte, that means I have 4096 entry points which is a lot more than I need. The entry point for reboot is sometimes accessed through at least two different addressings, so I avoided to place my HLT instruction there and instead placed the conventional five bytes long far jump instruction there, which jumps to one of my entry points. After this five bytes instruction are eight bytes reserved for the BIOS date written as month / date / year. The last three bytes are three single byte instructions HLT IRET RETF. The HLT and IRET are actually used, the RETF I just placed there because it might come in handy. Because of the DOS memory management and the reboot entry point, there is no way to make the ROM smaller than 16 bytes.
To actually protect the ROM against writing I mark the entire page read only, though it is only the last 16 bytes I really need to protect. This means any write to the first 4050 bytes of this page will trap, those are the traps that I needed to fix in the kernel because they would Oops if triggered by a stack access by an instruction emulated in the kernel. All those traps of course slows down execution, so I might want to sacrifice the last 4050 bytes for a bit of performance. I'm still looking for an efficient way to access the last bytes. If I could put an upper limit to the address accessed by virtual 86 mode, I could switch between a limit just below the ROM and a write protected page, which I belive would speed up execution. Together with my emulator I have put a GPL'ed UMM driver that works with my emulator, quite conveniently this driver does not support the last 4050 bytes of UMB that have been causing problems anyway. EMM386 doesn't work with my emulator, and never will because of braindead Intel design.
Actually I have written my own PC emulator, but it is far from as usable as DOSemu. I wanted to test a way to do the emulation with only 16 bytes used for ROM. As long as it was fun I kept coding. But eventually I ran into some problems. If I actually wanted to use all the available 255KB of UMB the kernel would Oops when the stack was on the same page as my ROM. I fixed the kernel bugs together with Manfred Spraul and Stas Sergeev. But I never got back to coding on my emulator.
You got a point. I'd better stop reading slashdot and do some coding. Somebody mail me when this farce is over and SCO is history.