Since you're so keen on definitions, perhaps you should check what "taking" means. For property, it means the original owner no longer has possession of it. For services, it's a little more nebulous, but would apply where they have spent time and/or resources, and not received promised compensation.
When you download software for free, you haven't taken anything, because the original copy still exists. You have copied something. This is why despite the best efforts of the recording industries, copyright violation is not legally the same as theft.
How about, oh, copyright infringement? Which is what, legally, it is.
Seriously, there's a perfectly good and accurate name for it, and words like theft and piracy are intentionally used to create associations with much more serious crimes.
Yes, language changes, but in this case it's been deliberately manipulated to influence people's emotions.
True, but it depends on the complexity of the codes somewhat. The more data chunks are encoded into a block, the longer you have to wait to verify it. If some chunks are downloaded directly, not in an encoded form, this may reduce the wait.
I should add that my suggestion wasn't taken from the paper; it was a very quick (and flawed, as you've pointed out) way to handle the problem of having an exponential number of potential blocks. Network coding is rather far removed from my area of research, so I'm not sure what practical solutions exist.
Because there are an exponential number of codes. That said, once you've got enough blocks to work out one of the original chunks, you can verify it using the checksums in the torrent file (or equivalent). This also confirms that the blocks contributing to that chunk aren't poisoned.
I believe the paper does address this issue (certainly one of the comments on the blog says it does), but I didn't read it in great detail.
Quite. The point of Avalanche is to apply network coding to a Bittorrent-like system. The exact details of the system it's compared to aren't particularly important.
No sane researcher is going to simulate the latest version of Bittorrent in all its glory. Instead, it makes sense to compare a relatively simple protocol without network coding to the same protocol with it.
The precise way in which Bittorrent's throttling works is irrelevent, since the missing piece problem would still occur if all clients played by the rules and throttling were unnecessary.
In short, Cohen has attacked some minor side issues of the paper, and ignored the main result completely.
Since the 6GB model is the larger iPod Mini, I don't think it's a question of stinginess. But you're right about the song capacity. Most likely, the Sun hasn't done their research properly (or, knowing the Sun, at all).
Well, when bush started waffling about Operation Infinite Justice (demonstrating that he doesn't have a clue what infinite means, but I digress), what was the first thing I thought of?
Not being American, I'm not too well up on the constitution. Surely, though, if the government requires ISPs to retain records, then the constitution applies?
Otherwise, the government can get around anything they don't like in the constitution just by outsourcing.
In this case, yes, because the primary justification for war was that Iraq had weapons of mass destruction and that they could be deployed within 45 minutes.
Totally offtopic, I know. Being from the UK, I continue to find "no solicitors" amusing. Even more so coming from a lawyer.
For the sake of relevence, how do you propose to explicitly disallow spam? SMTP doesn't have a "no-spam" flag, and even if one were added, it would be difficult to enforce its use (all mail sending programs would need to be updated, quite apart from the fact that SMTP is just a protocol, and not legally enforcable). There's currently no email equivalent to "no soliciting", which makes it somewhat different to door-to-door sales.
Re:Pigeon-hole principle
on
SHA-1 Broken
·
· Score: 2, Interesting
To nit-pick further, the pigeon-hole principle says nothing about not reusing any slot. It states that if you place n items in to n slots, either every slot is filled, or (at least) one slot has more than one item.
Equivalently, if there are n+1 items, there must be a slot with more than one item. Your statement is a special case of the principle, but not as general.
It is possible to prove (by induction) that there are an infinite number of collisions for some hash value using this. However, proving that collisions exist for every hash value requires detailed knowledge of the algorithm, and doesn't follow directly from the pigeon-hole principle.
I would assume, since they're one of the bigger companies out there, that they think it will make them look good. If they don't crack down on the fraudsters, there's a risk that people will stop trusting Verisign. In which case, no more profits for them.
Not quite. Both keys consist of a product of two primes u and v, and another number (called d or e. I forget exactly how they're computed. I think they're both co-prime with uv, and related in some way modulo (u-1)(v-1)). The result is such that x^de = x (mod uv).
The difficulty of factoring ensures that you can't recover (u-1)(v-1) from uv - if that were possible, it would be easy to retrieve the private key from the public one, or vice-versa.
You can encrypt with the public key and decrypt with the private key, or encrypt with the private key and decrypt with the public key (this is how signing works). It's a trivial corollary that there's a 1-1 correspondence between public and private keys.
In most cases, yes. But it's also common for series to be rereleased as DVD rips later, once the Japanese DVDs are released. This is higher quality than the TV release. Also, extra material from the DVDs is also sometimes released (see, for example, Maria-sama ga Miteru).
Fansubs are also released of OVA series, which by their very nature are not broadcast on TV.
Why? Media Factory are well within their rights. When I first heard about this, my immediate reaction was to boycott their products as well. But fansubbing is technically illegal (although tolerated), and if a company isn't happy with fansub distribution, they're entitled to stop it.
One point that doesn't seem to have been brought up: since digisubbing has become popular, fansubs are no longer restricted to the English-speaking world. They are also available to Japanese fans, often at DVD quality. Even though it isn't the fansubbers' intention, they are making new anime available for free in Japan, as well as elsewhere.
If other companies follow Media Factory's lead, and the fansubbers' argument is correct, then nobody needs to boycott anything: sales will drop on their own. Otherwise, you're just skewing the figures.
The impression I get is that PalmOS and PocketPC came from two different directions and didn't quite meet in the middle: PalmOS evolved mainly from simpler electronic organisers and, earlier, the Filofax; PocketPC is closer to a cut-down laptop.
Having used both systems, the iPaq is nice, but PocketPC is too complicated for use as a genuine organiser. All I ask for is something with the power of an iPaq and the ease of use of PalmOS.
The way the form was produced, the question is completely useless for statistical purposes anyway. The point of the census is that it is compulsary; making a question optional makes the data invalid because it is impossible to estimate the religion of those that did not answer. Assuming the proportions are the same as those that did doesn't work.
As a result, nobody is actually going to use the results of this question for any planning (at least, it would be statistically flawed if they did), and it hardly seems worth putting it in in the first place.
As I remember (this was some time ago), it was two years maximum for not giving up keys. This is a major improvement on 30, or life, or whatever. In fact, it would make sense to go with seven years, allowing you to warn all your co-conspirators at the same time. Of course, at the time everyone was concerned about child porn, not terrorism.
Somebody seriously screwed up when they came up with that law.
As far as I can tell, that's the whole point - some (possibly most) of that 60% will buy a dubbed-only version when it's released, and then a subbed version as well. More money for the manufacturers. Especially if they release it in Japan first and get people to import it for a few months. Personally, I have to import most anime DVDs anyway (the UK selection is lousy); I'd only buy a dubbed version on VHS locally.
As it happens, this is illegal. The GPL specifically states that any derivative work must be licensed under the GPL - it is not possible to remove or modify the license once it has been applied (except potentially with the agreement of all contributors). It is unlikely that anyone would notice if you did this, but that doesn't make it legal.
Slashdot Mirror
Since you're so keen on definitions, perhaps you should check what "taking" means. For property, it means the original owner no longer has possession of it. For services, it's a little more nebulous, but would apply where they have spent time and/or resources, and not received promised compensation.
When you download software for free, you haven't taken anything, because the original copy still exists. You have copied something. This is why despite the best efforts of the recording industries, copyright violation is not legally the same as theft.
How about, oh, copyright infringement? Which is what, legally, it is.
Seriously, there's a perfectly good and accurate name for it, and words like theft and piracy are intentionally used to create associations with much more serious crimes.
Yes, language changes, but in this case it's been deliberately manipulated to influence people's emotions.
True, but it depends on the complexity of the codes somewhat. The more data chunks are encoded into a block, the longer you have to wait to verify it. If some chunks are downloaded directly, not in an encoded form, this may reduce the wait.
I should add that my suggestion wasn't taken from the paper; it was a very quick (and flawed, as you've pointed out) way to handle the problem of having an exponential number of potential blocks. Network coding is rather far removed from my area of research, so I'm not sure what practical solutions exist.
Because there are an exponential number of codes. That said, once you've got enough blocks to work out one of the original chunks, you can verify it using the checksums in the torrent file (or equivalent). This also confirms that the blocks contributing to that chunk aren't poisoned.
I believe the paper does address this issue (certainly one of the comments on the blog says it does), but I didn't read it in great detail.
Quite. The point of Avalanche is to apply network coding to a Bittorrent-like system. The exact details of the system it's compared to aren't particularly important.
No sane researcher is going to simulate the latest version of Bittorrent in all its glory. Instead, it makes sense to compare a relatively simple protocol without network coding to the same protocol with it.
The precise way in which Bittorrent's throttling works is irrelevent, since the missing piece problem would still occur if all clients played by the rules and throttling were unnecessary.
In short, Cohen has attacked some minor side issues of the paper, and ignored the main result completely.
Since the 6GB model is the larger iPod Mini, I don't think it's a question of stinginess. But you're right about the song capacity. Most likely, the Sun hasn't done their research properly (or, knowing the Sun, at all).
Well, when bush started waffling about Operation Infinite Justice (demonstrating that he doesn't have a clue what infinite means, but I digress), what was the first thing I thought of?
For great justice.
Not being American, I'm not too well up on the constitution. Surely, though, if the government requires ISPs to retain records, then the constitution applies?
Otherwise, the government can get around anything they don't like in the constitution just by outsourcing.
Yes. Although whether that's a comment on UK pricing or the value of the dollar is left as an exercise to the reader.
In this case, yes, because the primary justification for war was that Iraq had weapons of mass destruction and that they could be deployed within 45 minutes.
Totally offtopic, I know. Being from the UK, I continue to find "no solicitors" amusing. Even more so coming from a lawyer.
For the sake of relevence, how do you propose to explicitly disallow spam? SMTP doesn't have a "no-spam" flag, and even if one were added, it would be difficult to enforce its use (all mail sending programs would need to be updated, quite apart from the fact that SMTP is just a protocol, and not legally enforcable). There's currently no email equivalent to "no soliciting", which makes it somewhat different to door-to-door sales.
To nit-pick further, the pigeon-hole principle says nothing about not reusing any slot. It states that if you place n items in to n slots, either every slot is filled, or (at least) one slot has more than one item.
Equivalently, if there are n+1 items, there must be a slot with more than one item. Your statement is a special case of the principle, but not as general.
It is possible to prove (by induction) that there are an infinite number of collisions for some hash value using this. However, proving that collisions exist for every hash value requires detailed knowledge of the algorithm, and doesn't follow directly from the pigeon-hole principle.
I would assume, since they're one of the bigger companies out there, that they think it will make them look good. If they don't crack down on the fraudsters, there's a risk that people will stop trusting Verisign. In which case, no more profits for them.
Or maybe he's in the latter group?
Not quite. Both keys consist of a product of two primes u and v, and another number (called d or e. I forget exactly how they're computed. I think they're both co-prime with uv, and related in some way modulo (u-1)(v-1)). The result is such that x^de = x (mod uv).
The difficulty of factoring ensures that you can't recover (u-1)(v-1) from uv - if that were possible, it would be easy to retrieve the private key from the public one, or vice-versa.
You can encrypt with the public key and decrypt with the private key, or encrypt with the private key and decrypt with the public key (this is how signing works). It's a trivial corollary that there's a 1-1 correspondence between public and private keys.
In most cases, yes. But it's also common for series to be rereleased as DVD rips later, once the Japanese DVDs are released. This is higher quality than the TV release. Also, extra material from the DVDs is also sometimes released (see, for example, Maria-sama ga Miteru).
Fansubs are also released of OVA series, which by their very nature are not broadcast on TV.
Why? Media Factory are well within their rights. When I first heard about this, my immediate reaction was to boycott their products as well. But fansubbing is technically illegal (although tolerated), and if a company isn't happy with fansub distribution, they're entitled to stop it.
One point that doesn't seem to have been brought up: since digisubbing has become popular, fansubs are no longer restricted to the English-speaking world. They are also available to Japanese fans, often at DVD quality. Even though it isn't the fansubbers' intention, they are making new anime available for free in Japan, as well as elsewhere.
If other companies follow Media Factory's lead, and the fansubbers' argument is correct, then nobody needs to boycott anything: sales will drop on their own. Otherwise, you're just skewing the figures.
The impression I get is that PalmOS and PocketPC came from two different directions and didn't quite meet in the middle: PalmOS evolved mainly from simpler electronic organisers and, earlier, the Filofax; PocketPC is closer to a cut-down laptop.
Having used both systems, the iPaq is nice, but PocketPC is too complicated for use as a genuine organiser. All I ask for is something with the power of an iPaq and the ease of use of PalmOS.
The way the form was produced, the question is completely useless for statistical purposes anyway. The point of the census is that it is compulsary; making a question optional makes the data invalid because it is impossible to estimate the religion of those that did not answer. Assuming the proportions are the same as those that did doesn't work.
As a result, nobody is actually going to use the results of this question for any planning (at least, it would be statistically flawed if they did), and it hardly seems worth putting it in in the first place.
As I remember (this was some time ago), it was two years maximum for not giving up keys. This is a major improvement on 30, or life, or whatever. In fact, it would make sense to go with seven years, allowing you to warn all your co-conspirators at the same time. Of course, at the time everyone was concerned about child porn, not terrorism.
Somebody seriously screwed up when they came up with that law.
As far as I can tell, that's the whole point - some (possibly most) of that 60% will buy a dubbed-only version when it's released, and then a subbed version as well. More money for the manufacturers. Especially if they release it in Japan first and get people to import it for a few months. Personally, I have to import most anime DVDs anyway (the UK selection is lousy); I'd only buy a dubbed version on VHS locally.
As it happens, this is illegal. The GPL specifically states that any derivative work must be licensed under the GPL - it is not possible to remove or modify the license once it has been applied (except potentially with the agreement of all contributors). It is unlikely that anyone would notice if you did this, but that doesn't make it legal.