Yes, at 128-bits. Variable-width addresses would bog down routers, because now they have to parse the length out of the packet. With fixed-width addreesses, it's just an XOR and bit shift, or maybe an lookup in an array of bytes (depending on what the implementers did their work). 128-bits is absurdadly huge (on the order of the number of atoms in the universe), so nobody worries about running out.
"there are sixteen trillion addys, but my entire workplace gets one - why?"
IPv6 ISPs are required to give each customer an entire subnet to themselves (a/48, IIRC). That gives you 2**80 addresses to play with--several powers more than there are available IPv4 addresses.
Is the god-awful port-numbering system still there?
Ports are handled by higher-layer protocols, like TCP or UDP. Neither IPv4 or IPv6 have an concept of what a port is. I imagine, though, that a string-based port system would be too computationally expensive on high-traffic hosts and routers.
Getting this inverse is considered to be hard - but this is not proven yet.
Nobody has proven you can't have an efficent algorithm for factoring large prime numbers, either. Though in that case, people have been trying to solve the problem for centries, and the added incentive of breaking RSA hasn't produced a breakthrough, either.
Technically true. The question is if you'll finish searching the entire keyspace before the universe blows up.
It was estimated that in 1993, you could take $1 million and build a special-purpose computer and break any 56-bit DES key in three hours. Given Moore's Law, you could probably get a few of your freinds today w/GHz-class systems and break it in a few days. However, as the bit size increases, the keyspace grows exponentially. We'd need some fundamental advances in computers to brute-force a 160-bit key before all the stars become black holes.
Homeworld is made by Relic, not Valve (they both happen to publish under Siera, but are not the same dev house).
I was hoping for porting, too, but this might take a lot of work. Parts of the code that Relic doesn't actualy own aren't included, so you'd need to do some hacking just to get a faithful version of the executable and DLLs. Then you get to wade through all that DirectX crap and convert it to something that actually runs elsewhere.
So this looks like it'll be mostly a thing to help modders, not those hoping for a port. I will be delighted to be wrong, though:)
Don't forget, too, that the moral-high-ground this company takes is equally applicable to Pinkerton. When a company needed to break up a worker's union meeting around the 1900's, they'd call Pinkerton. The Mafia would have had these guys on their speed-dial, had the technology existed at the time. Founded by the same fellow who started the US Secret Service, who have a such a stellar record of civil rights abuses of their own.
I find the Pinkerton analogy to have a beautiful double-meaning in the context of the above article. I haven't figured out yet if BayTSP intended it that way or not.
Re: Quantum Computers will NOT make crypto obsolet
on
Practical Cryptography
·
· Score: 1
You will note that I said "become practical within that timeframe" and "if they do, then . . ". I'm well aware of the breakthroughs required to make that happen.
There was going to be a Babylon 5-based flight sim with Star Furies, which sounds like it would have been pretty close. You could at least spin around while still moving in the same direction, which is a lot more acurate than the "fighting in a gravity-less atmosphere" that Star Wars/Trek tend to use.
Unfortunately, that game got killed off with the production company ran out of cash.
I would not want to see what would happen if you taught them assembly first. I have seen one college in Ontario, Canada that teaches assembly in their CMP program in the last semester.
I'd love to see that. Call it "thinning the hurd".
As another poster pointed out, it is likely that their memories were wiped. Beyond being standard practice for droid maintence, it is probably likely that they had to wipe them to keep Luke & Leia's locations a secret from their father.
There is one exception: OEM licenses. As far as the law is concerned, an OEM license applies only to the machine it came with (at least, that's my understanding. IANAL.) The donator could still give the school the machine without turning over the license, though the donator cannot (legaly) use that license for another machine.
Of course, none of this applies to shrink-wraped copies of the OS, or (usually) computers bought through mom-and-pop stores, which are full licenses.
I would also wonder how this would apply to a large corperation, or even another school, who usually get special, bulk licenses that are to be used within the orginization and cannot be transfered. According to Microsoft, they couldn't donate the computers at all!
I think it's the humans who aren't with the program. Sci-fi (mostly Roddenbery) likes to pat humanity on the back, which is fine from time to time, but I think a far more likely circumstance is that humanity would be just another species. In Farscape, the lone human has some very important information that is vital to the two great powers of the galaxy he's stuck in, but most of that information was gained accidently. Otherwise, that human are just a dopey race that (understandably) doesn't know basic things about the galaxy around him, and has bad eyesight.
Note: One of my favorite lines is where Chrition says "My eyes are fine! I have perfect 20/20 vision, and they're blue!"
Personally, I just download the individual packages (surely your firewall doesn't block both HTTP and FTP?) and burn them for OpenBSD. If I have some cash on hand, I send a little to OpenBSD with a little note saying "I burned my own CD. Here's some cash because I'm not an arrogant jerk like Theo":)
I prefer OpenBSD when security matters, FreeBSD when performance matters, and GNU/Linux when I want something well-rounded.
What it does say, is that there can be no generalized "obfuscator" that you run your program through and voila you've got an obfuscated version. Hoever, program obfuscation is possible on a per program basis.
If I understand this right, you can create an obfuscator that will work specifically on track 3 of "Oops, I did it again", but that same obfuscator will fail on every other track on that CD, or any other series of bits ever made. Correct?
There was a Java emulator of the PDP-1 around, where you could play a game which was exactly like the orginal spacewars except for a few lines of code. The KDE game KSpaceDuel is also an acceptable alternative.
Clustering simply means to join several computers together in order to combine their resources. A "resource" could be bandwidth, CPU cycles, hard drive space, whatever. There is nothing "OS specific" about this definition.
As many others have pointed out, Microsoft is typically more known for doing clustering for network applications (like web servers), but they have been doing some recent advances in doing more Beowulf-style computing as well.
There is also more to the issue of price than the licensing that a lot of posters are mentioning. Most indiviual computers in a Beowulf cluster run headless (e.g., without a monitor or keyboard). I know NT 4 refused to boot when headless, and AFAIK, Win2k and XP won't either (but I'm not sure about that). The money you use buying monitors could be used to get a few extra machines in your cluster (even if you get bargin-basement monitors). Another consideration is the space for that extra stuff; could you imagine how much more space this would take up if every machine needed it's own monitor and keyboard?
Anyway I think it is a great project and put all my (big)file releases up as a public KSK, simply beacuse it's a good, clean and simple way to share files.
Sorry, but I'm about to make this a whole lot more complex:) In the Freenet development cicle we have a saying:
DON'T USE A KSK!!!
KSKs are highly insecure, because you have no means of validating that the data is really the data that you orginally inserted. I suggest you migrate your old KSK data over to a subspace.
Until now Freenet has no popularity in both areas.
The MAME community has distributed quite a number of ROM images via Freenet. That has to count for something.
Yeah, crypto is easy. Knowing what a "CHK" is without looking it up is easy. Figuring out how to stop Man-in-the-middle attacks is easy. Solving the initial node announcment problem is easy. It's just that gosh darn Java that is getting in the way.
Incedently, there are many people working on a GCJ-compiled Freenet, which would allow you to run a Java node as a native program. Why don't you go help them out instead of whining on Slashdot?
C would have been the language of choice simply because more people know C than java, porting would have been faster.
Freenet did not initialy start using Java because of being cross-platform. It was used because Java is a nice language for prototyping, and because Java is very nice to use for network-centric programs.
someone needs to port freenet to C right now, if its ported to C people will develop for it.
Are you talking about the node or the clients? The clients have FCP, an easy protocol to implement, so there is no reason why you can't write your client in C. As for the node, do you really expect Freenet to throw away all the work that has been done for the last few years and start over in C? (We actualy had this debate on the Freenet development lists recently).
Further, while lack of documentation certainly doesn't help, Freenet is a inheirently complex beast that isn't for the faint of heart to attempt understanding of. Knowing a lot about crypto will give you a head start, but even so, the typical time to overcome Freenet's learning curve is about six months, if you push it. Porting the node to C isn't going to help people who are having problems getting over basic Freenet concepts, such as CHKs.
If you don't like the lack of documentation, why don't you write it yourself. The developers are quite responsive to anyone needing help with documentation.
I think by cowboyneal's clear victory in this poll, we can conclude that most slashdotters don't watch farscape. Heck, most of us probably don't even get it.
Personally, I really couldn't choose between these characters, so CowboyNeal was the easy way out.
nobody I know gets both scifi and comedy central
I do, although everything else about my local cable monopoly sucks.
Slashdot Mirror
is the IP address still fixed-length?
Yes, at 128-bits. Variable-width addresses would bog down routers, because now they have to parse the length out of the packet. With fixed-width addreesses, it's just an XOR and bit shift, or maybe an lookup in an array of bytes (depending on what the implementers did their work). 128-bits is absurdadly huge (on the order of the number of atoms in the universe), so nobody worries about running out.
"there are sixteen trillion addys, but my entire workplace gets one - why?"
IPv6 ISPs are required to give each customer an entire subnet to themselves (a /48, IIRC). That gives you 2**80 addresses to play with--several powers more than there are available IPv4 addresses.
Is the god-awful port-numbering system still there?
Ports are handled by higher-layer protocols, like TCP or UDP. Neither IPv4 or IPv6 have an concept of what a port is. I imagine, though, that a string-based port system would be too computationally expensive on high-traffic hosts and routers.
Getting this inverse is considered to be hard - but this is not proven yet.
Nobody has proven you can't have an efficent algorithm for factoring large prime numbers, either. Though in that case, people have been trying to solve the problem for centries, and the added incentive of breaking RSA hasn't produced a breakthrough, either.
Technically true. The question is if you'll finish searching the entire keyspace before the universe blows up.
It was estimated that in 1993, you could take $1 million and build a special-purpose computer and break any 56-bit DES key in three hours. Given Moore's Law, you could probably get a few of your freinds today w/GHz-class systems and break it in a few days. However, as the bit size increases, the keyspace grows exponentially. We'd need some fundamental advances in computers to brute-force a 160-bit key before all the stars become black holes.
Homeworld is made by Relic, not Valve (they both happen to publish under Siera, but are not the same dev house).
I was hoping for porting, too, but this might take a lot of work. Parts of the code that Relic doesn't actualy own aren't included, so you'd need to do some hacking just to get a faithful version of the executable and DLLs. Then you get to wade through all that DirectX crap and convert it to something that actually runs elsewhere.
So this looks like it'll be mostly a thing to help modders, not those hoping for a port. I will be delighted to be wrong, though :)
Don't forget, too, that the moral-high-ground this company takes is equally applicable to Pinkerton. When a company needed to break up a worker's union meeting around the 1900's, they'd call Pinkerton. The Mafia would have had these guys on their speed-dial, had the technology existed at the time. Founded by the same fellow who started the US Secret Service, who have a such a stellar record of civil rights abuses of their own.
I find the Pinkerton analogy to have a beautiful double-meaning in the context of the above article. I haven't figured out yet if BayTSP intended it that way or not.
You will note that I said "become practical within that timeframe" and "if they do, then . . ". I'm well aware of the breakthroughs required to make that happen.
Depends on if Quantum Computers become practical in that timeframe. If they do, RSA (and a lot of other crypto algorithms) become useless.
There was going to be a Babylon 5-based flight sim with Star Furies, which sounds like it would have been pretty close. You could at least spin around while still moving in the same direction, which is a lot more acurate than the "fighting in a gravity-less atmosphere" that Star Wars/Trek tend to use.
Unfortunately, that game got killed off with the production company ran out of cash.
I would not want to see what would happen if you taught them assembly first. I have seen one college in Ontario, Canada that teaches assembly in their CMP program in the last semester.
I'd love to see that. Call it "thinning the hurd".
if you don't like it don't use it
I don't. The last CD I bought was the "Mission: Impossible 2" soundtrack, and that was a mistake.
If you don't want to pay for music, find like minded artists and listen to them.
I do. I bought the CD of a small, local celtic band, even though they make their music freely available.
Think of it as parrellism.
As another poster pointed out, it is likely that their memories were wiped. Beyond being standard practice for droid maintence, it is probably likely that they had to wipe them to keep Luke & Leia's locations a secret from their father.
There is one exception: OEM licenses. As far as the law is concerned, an OEM license applies only to the machine it came with (at least, that's my understanding. IANAL.) The donator could still give the school the machine without turning over the license, though the donator cannot (legaly) use that license for another machine.
Of course, none of this applies to shrink-wraped copies of the OS, or (usually) computers bought through mom-and-pop stores, which are full licenses.
I would also wonder how this would apply to a large corperation, or even another school, who usually get special, bulk licenses that are to be used within the orginization and cannot be transfered. According to Microsoft, they couldn't donate the computers at all!
This seems to be a pretty good intro to resistor logic.
I think it's the humans who aren't with the program. Sci-fi (mostly Roddenbery) likes to pat humanity on the back, which is fine from time to time, but I think a far more likely circumstance is that humanity would be just another species. In Farscape, the lone human has some very important information that is vital to the two great powers of the galaxy he's stuck in, but most of that information was gained accidently. Otherwise, that human are just a dopey race that (understandably) doesn't know basic things about the galaxy around him, and has bad eyesight.
Note: One of my favorite lines is where Chrition says "My eyes are fine! I have perfect 20/20 vision, and they're blue!"
Personally, I just download the individual packages (surely your firewall doesn't block both HTTP and FTP?) and burn them for OpenBSD. If I have some cash on hand, I send a little to OpenBSD with a little note saying "I burned my own CD. Here's some cash because I'm not an arrogant jerk like Theo" :)
I prefer OpenBSD when security matters, FreeBSD when performance matters, and GNU/Linux when I want something well-rounded.
What it does say, is that there can be no generalized "obfuscator" that you run your program through and voila you've got an obfuscated version. Hoever, program obfuscation is possible on a per program basis.
If I understand this right, you can create an obfuscator that will work specifically on track 3 of "Oops, I did it again", but that same obfuscator will fail on every other track on that CD, or any other series of bits ever made. Correct?
(yes, that is my normal .sig)
There was a Java emulator of the PDP-1 around, where you could play a game which was exactly like the orginal spacewars except for a few lines of code. The KDE game KSpaceDuel is also an acceptable alternative.
Clustering simply means to join several computers together in order to combine their resources. A "resource" could be bandwidth, CPU cycles, hard drive space, whatever. There is nothing "OS specific" about this definition.
As many others have pointed out, Microsoft is typically more known for doing clustering for network applications (like web servers), but they have been doing some recent advances in doing more Beowulf-style computing as well.
There is also more to the issue of price than the licensing that a lot of posters are mentioning. Most indiviual computers in a Beowulf cluster run headless (e.g., without a monitor or keyboard). I know NT 4 refused to boot when headless, and AFAIK, Win2k and XP won't either (but I'm not sure about that). The money you use buying monitors could be used to get a few extra machines in your cluster (even if you get bargin-basement monitors). Another consideration is the space for that extra stuff; could you imagine how much more space this would take up if every machine needed it's own monitor and keyboard?
Everyone who couldn't stop laughing at the irony when they first saw "Sonic the Hedghog" for a Nintendo system, raise your hand.
Anyway I think it is a great project and put all my (big)file releases up as a public KSK, simply beacuse it's a good, clean and simple way to share files.
Sorry, but I'm about to make this a whole lot more complex :) In the Freenet development cicle we have a saying:
DON'T USE A KSK!!!
KSKs are highly insecure, because you have no means of validating that the data is really the data that you orginally inserted. I suggest you migrate your old KSK data over to a subspace.
Until now Freenet has no popularity in both areas.
The MAME community has distributed quite a number of ROM images via Freenet. That has to count for something.
Yeah, crypto is easy. Knowing what a "CHK" is without looking it up is easy. Figuring out how to stop Man-in-the-middle attacks is easy. Solving the initial node announcment problem is easy. It's just that gosh darn Java that is getting in the way.
Incedently, there are many people working on a GCJ-compiled Freenet, which would allow you to run a Java node as a native program. Why don't you go help them out instead of whining on Slashdot?
In comparison to what? Client development? Ease-of-use? Node implementation?
Freenet is pretty easy for client development and average for ease-of-use. However, node implementation is no easy task (just ask Adam Langly).
C would have been the language of choice simply because more people know C than java, porting would have been faster.
Freenet did not initialy start using Java because of being cross-platform. It was used because Java is a nice language for prototyping, and because Java is very nice to use for network-centric programs.
someone needs to port freenet to C right now, if its ported to C people will develop for it.
Are you talking about the node or the clients? The clients have FCP, an easy protocol to implement, so there is no reason why you can't write your client in C. As for the node, do you really expect Freenet to throw away all the work that has been done for the last few years and start over in C? (We actualy had this debate on the Freenet development lists recently).
Further, while lack of documentation certainly doesn't help, Freenet is a inheirently complex beast that isn't for the faint of heart to attempt understanding of. Knowing a lot about crypto will give you a head start, but even so, the typical time to overcome Freenet's learning curve is about six months, if you push it. Porting the node to C isn't going to help people who are having problems getting over basic Freenet concepts, such as CHKs.
If you don't like the lack of documentation, why don't you write it yourself. The developers are quite responsive to anyone needing help with documentation.
I think by cowboyneal's clear victory in this poll, we can conclude that most slashdotters don't watch farscape. Heck, most of us probably don't even get it.
Personally, I really couldn't choose between these characters, so CowboyNeal was the easy way out.
nobody I know gets both scifi and comedy central
I do, although everything else about my local cable monopoly sucks.