Why? I'm hoping that they get this granted. It won't make the situation worse for anyone in the tech sector who is already experiencing this kind of crap. After all, having a patent on this process don't make it more legitimate... unless they're going to sue other patent trolls for violation.
Then, it is truly a thing of beauty, and should be widely applauded.
Almost right. The GP is asking how to connect to the overlay network (in DHT terms) to find peers to query, rather that how to find the peers that match a query, which is where a DHT comes into it.
Without delving into the Tribble source I would guess the other reply below is correct that a hosts.txt lists some long-lived peers in the network to bootstrap the search from. Once one of them has been contacted then more local peers can be found, and then the DHT protocol can be run over the top.
In fact I didn't even check the Tribble faq to see if they are running a DHT, but it seems likely given that it's the fashionable shiney new thing.
Of course Louis argues against traditional peer review and elitism. In his case it has worked well at keeping his own brand of psuedo-science away from respectable outlets. For those of you arguing against him - don't bother. Just google him and read some of the bizarre rants that he has come up with in the past. Then remember that peer review was designed to keep him, and those cranks like him, away from mainstream academia.
No you've completely missed the point. Each user takes their own photograph of themselves. This never leaves their device. When a connection is to be setup the other user photographs them. There are now 4 photos in the set:
Alice's photo of Alice Alice's photo of Bob Bob's photo of Alice Bob's photo of Bob
Each user now has a pair of images that should be similar (but are not the same). Alice has a photo of her and Bob, Bob has different photos of Alice and himself.
The images are hashed in some sense to generate a seed for the key. The assumption is that the image hash is robust enough that the two different images of Alice generate the same seed, and the same for Bob.
So if I now take a third pair of pictures, one of Alice and one of Bob, and the hash is robust then I can recompute their seed, and derive their key. As I said originally, it's some interesting vision work to come with a robust hash like that, but it is not actually secure.
If you still don't believe me, then reread the article and consider that the security of the system relies on Alice's picture never leaving her device. The same applies for Bob. To perform key agreement that means the same seed must be derived from the separate images.
This sounds like interesting work as I'm sure that the hashing of the photos to generate the passwords is quite interesting research. But from the summary (on the uni site) the work is quite flawed as a security measure. If I see Alice and Bob taking pictures of each other in order to establish a secure link then all I need to do is photograph them both covertly and I can regenerate their password.
Thanks for writing that - it's a really insightful post. It's interesting that Greenspan was still trumpeting the successes of computer modeling for subprime debt in 2005. I think the first mention that I heard from an Actuary friend that a wave of shit was about to hit the fan was roughly 3-4 years ago over a poker game. Back then rumours were starting to circulate that the aaa rating on subprime securities was garbage.
Once of the basic difficulties in modeling markets is that the rules change. The rules get encoded into the simulation as the basis of the model. Over the timescales that investment houses operate on, it looks like these assumptions are static. But when they do move, on a longer 15-20 year timescale everything changes, and the old model just produces garbage.
The point that you make about the interconnectedness of uncorrelated risks is a great example of this. In the boom years those risks were uncorrelated, so the correct model was one that treated them separately. But the inevitable bust is a Black Swan - at any given point it has negligible probability, but we know that it will occur sooner or later. When it does the basic rules of the game change and the effects are huge.
OK, so your answer is that you don't think there should be large changes to the source in isolation from the main branch. I can buy that being possible in small projects. I can also buy making those changes switchable in some cases.
The reason that I say some cases is that sometimes a change cannot be isolated from a large swathe of code. When this happens every point that it touches the code becomes a point of failure. Isolating the changes in the source control guarantees that they cannot affect the stable release until the dev team is ready.
By adding switches to the code to choose versions you are moving that responsibility from source control into runtime code. Although that may work, it is also a potential source of failure.
I agree that it is a good aim to start small, and change things incrementally, but experience tells me that this is not always possible. Say for example you're working on a compiler, and after a few versions have been released it becomes apparent that one of the internal representations is wrong. Maybe it was a bad decision in the design, or maybe the functionality required has shifted over time. Whatever the reason this type of major surgery - changing a basic datatype that everything depends on does occur some times in big projects.
In those cases it doesn't seem feasible to change it all in one go, freezing the mainline code. Or to code conditionals in each bit of code affected - as this introduces too many potential points of failure.
This type of case is the rational for branching and merging in source control. It's there because there is a real need for it. Even if the theory says that you shouldn't get into that position, in practice is can be unavoidable. It is a major decider in which version control to use.
I use subversion day to day. The merging support doesn't seem to be fantastic - although it is better than I remember cvs being. Lots of people seem to recommend git because it has a strength in merges. I was curious if you had some silver bullet to prevent merges, but it would seem that you just have the same regular grade bullets as everyone else.:)
Imagine for a second that the timescale for developing a new feature is longer than your timescale for releasing minor builds on your stable branch. That's why the branch gets the name experimental. This is a common requirement in lots of software, and not simply bad management.
How would you deal with this without branching and merging?
So as a cmd-line user who only uses centralized repositories (with only a couple of other developers), what kind of benefits would git provide over subversion?
So how would you handle maintaining a stable branch with a small amount of day-to-day changes at the same time as an experimental branch where live dev work is going on?
There are plenty of reasons why this is necessary, and it drives people to branch merge big chunks of codebases. What alternative method would you use?
Actually wikipedia tries not to select a version of the truth. The reason that most talk pages are filled with arguments and claims that articles lack NPOV is that the aim is to write something that is Neutral Point Of View. Rather than have one person select the truth according to their knowledge and biases, the idea is that all the people who argue about a topic can contribute something.
As long as they can back it up with a citation.
It's a low bar, but it's high enough to cut out 90% of the crap. As long as citations are provided then multiple view points can coexist in the same article. The point is to prevent somebody deciding on truth, to present the argument so that the reader can make up their own mind. To do so in such a way that there are links into the background literature that can be followed, or debate on the talk page to see which areas were the most contentious.
In short, it's the complete opposite of what you describe.
Why would you blame "greedy" individuals? They were behaving rationally and trying to maximise their own profits. Surely the people who rigged the game to make their bets look one way have a lot more blame coming their way. Expecting people to make sub-optimal financial decisions so that they can collectively improve the situation at a macro-scale is silly. Besides, if anyone wanted to they are following a strongly dominated strategy and would be drowned out by the hoards of people searching for equilibrium. Those people were stupid, but that doesn't make them to blame for the current crisis. Hopefully there will be a massive wave of fraud charges amongst the bankers and estate agents who deliberately rigged the game. Sadly the central bankers and politicians who helped them do will get away with it.
As far as reaching bottom goes, have a look at the graph of the FTSE over the past 15 years and tell me that doesn't look like the mother of all double tops. Haven't checked the graph for the DOW but it will be similar. I'm sure people will continue laughing at you all the way down...:)
I see that you are one of the hard of thinking. Let's go through this slowly for you.
You asked a question about a rhetorical statement. I've explained to you what the statement was retorting to. Your question has no value because nobody was suggesting that China nuking America would help them get their debt back. Despite this you are still after an answer?
The answer is that you are a thief of oxygen. Perhaps you should sit quietly and try to understand what people are saying before you try and take part in the debate again.
So you are claiming that something for which you have no evidence must be true, just because it cannot be disproved easily? Sounds like a religious nutjob to me.
Iraq got so much support in America, but not in Europe (where I live) for one simple reason. As a mass (or more accurately a mob) the citizens of America do not know the difference between patriotism, and nationalism, which is a long way down the slippery slope to fascism.
When you attribute causes to unseen entities to support your arguments, you are not just invoking religious paranoia - you are ignoring Occam's razor. There doesn't need to be a shadowy force behind American politicians seeking power over their neighbouring countries, and further control over the population. It can be done in plain view, because it is written into the rules of the game that define politics in America.
We will be the generation that watches the rise of USA Plc. It won't be because of a secret cabal of bankers pulling strings behind the scenes. It will be because corporations want politicians that don't stand in their way of making money through unfettered capitalism, and they can pay for it. Politicians don't want to stand in the way of corporate profits - because a strong economy is the surest way to re-election. Couple this with an expanding government - no government on earth has ever given up powers or control once they have seized them, and you have all the ingredients that you need.
A simple rule in life (call it Occam's rule for people) is:
Never attribute to malice what you can attribute to stupidity.
I'm sure you can understand the connection to shadowy conspiracy theories.
You wish it was only $700B. Read the draft that is circulating. The new agency has a *net* asset cap of $700B at any one time. The amount of debt that it can issue to buy toxic funds is unlimited.
Can I just say a big thankyou from the rest of the world, as you seem to have agreed to bail out all toxic mortgage debt in the world. Quite amazing really, just how generous the US taxpayer is when nobody asks him first. Especially given the unlimited nature of the fund, sure the $700B cap is nice and all, but is your media reporting that as a total, or as the limit on the balance on the books at any one time?
If a datacenter is renting, rather than owning, their building then I can see that pushing up operating costs. As pointed out by the other reply. But money that you are not making is not an operating cost. Your other comment about additional cooling/heating is incorrect as this discussion was about trading increased space for better airflow to reduce heating/cooling costs.
Why does the increased space increase operating costs? I'd assume that the datacenter owns the land / building it is situated on / in. So that's a sunk cost, not a recurring one. Just because the building is bigger doesn't automatically imply higher maintenance costs, especially as it is less full. What is the other cost that increases?
Slashdot Mirror
Why? I'm hoping that they get this granted. It won't make the situation worse for anyone in the tech sector who is already experiencing this kind of crap. After all, having a patent on this process don't make it more legitimate... unless they're going to sue other patent trolls for violation.
Then, it is truly a thing of beauty, and should be widely applauded.
You must be trolling. As you well know there are many instances of infinite sets that can be enumerated effectively.
Now go back to your 47 channels of American Gladiators America. Go back to sleep America. You are free because we tell you that you are free.
Almost right. The GP is asking how to connect to the overlay network (in DHT terms) to find peers to query, rather that how to find the peers that match a query, which is where a DHT comes into it.
Without delving into the Tribble source I would guess the other reply below is correct that a hosts.txt lists some long-lived peers in the network to bootstrap the search from. Once one of them has been contacted then more local peers can be found, and then the DHT protocol can be run over the top.
In fact I didn't even check the Tribble faq to see if they are running a DHT, but it seems likely given that it's the fashionable shiney new thing.
Of course Louis argues against traditional peer review and elitism. In his case it has worked well at keeping his own brand of psuedo-science away from respectable outlets. For those of you arguing against him - don't bother. Just google him and read some of the bizarre rants that he has come up with in the past. Then remember that peer review was designed to keep him, and those cranks like him, away from mainstream academia.
No you've completely missed the point. Each user takes their own photograph of themselves. This never leaves their device. When a connection is to be setup the other user photographs them. There are now 4 photos in the set:
Alice's photo of Alice
Alice's photo of Bob
Bob's photo of Alice
Bob's photo of Bob
Each user now has a pair of images that should be similar (but are not the same). Alice has a photo of her and Bob, Bob has different photos of Alice and himself.
The images are hashed in some sense to generate a seed for the key. The assumption is that the image hash is robust enough that the two different images of Alice generate the same seed, and the same for Bob.
So if I now take a third pair of pictures, one of Alice and one of Bob, and the hash is robust then I can recompute their seed, and derive their key. As I said originally, it's some interesting vision work to come with a robust hash like that, but it is not actually secure.
If you still don't believe me, then reread the article and consider that the security of the system relies on Alice's picture never leaving her device. The same applies for Bob. To perform key agreement that means the same seed must be derived from the separate images.
This sounds like interesting work as I'm sure that the hashing of the photos to generate the passwords is quite interesting research. But from the summary (on the uni site) the work is quite flawed as a security measure. If I see Alice and Bob taking pictures of each other in order to establish a secure link then all I need to do is photograph them both covertly and I can regenerate their password.
Thanks for writing that - it's a really insightful post. It's interesting that Greenspan was still trumpeting the successes of computer modeling for subprime debt in 2005. I think the first mention that I heard from an Actuary friend that a wave of shit was about to hit the fan was roughly 3-4 years ago over a poker game. Back then rumours were starting to circulate that the aaa rating on subprime securities was garbage.
Once of the basic difficulties in modeling markets is that the rules change. The rules get encoded into the simulation as the basis of the model. Over the timescales that investment houses operate on, it looks like these assumptions are static. But when they do move, on a longer 15-20 year timescale everything changes, and the old model just produces garbage.
The point that you make about the interconnectedness of uncorrelated risks is a great example of this. In the boom years those risks were uncorrelated, so the correct model was one that treated them separately. But the inevitable bust is a Black Swan - at any given point it has negligible probability, but we know that it will occur sooner or later. When it does the basic rules of the game change and the effects are huge.
OK, so your answer is that you don't think there should be large changes to the source in isolation from the main branch. I can buy that being possible in small projects. I can also buy making those changes switchable in some cases.
The reason that I say some cases is that sometimes a change cannot be isolated from a large swathe of code. When this happens every point that it touches the code becomes a point of failure. Isolating the changes in the source control guarantees that they cannot affect the stable release until the dev team is ready.
By adding switches to the code to choose versions you are moving that responsibility from source control into runtime code. Although that may work, it is also a potential source of failure.
I agree that it is a good aim to start small, and change things incrementally, but experience tells me that this is not always possible. Say for example you're working on a compiler, and after a few versions have been released it becomes apparent that one of the internal representations is wrong. Maybe it was a bad decision in the design, or maybe the functionality required has shifted over time. Whatever the reason this type of major surgery - changing a basic datatype that everything depends on does occur some times in big projects.
In those cases it doesn't seem feasible to change it all in one go, freezing the mainline code. Or to code conditionals in each bit of code affected - as this introduces too many potential points of failure.
This type of case is the rational for branching and merging in source control. It's there because there is a real need for it. Even if the theory says that you shouldn't get into that position, in practice is can be unavoidable. It is a major decider in which version control to use.
I use subversion day to day. The merging support doesn't seem to be fantastic - although it is better than I remember cvs being. Lots of people seem to recommend git because it has a strength in merges. I was curious if you had some silver bullet to prevent merges, but it would seem that you just have the same regular grade bullets as everyone else. :)
Imagine for a second that the timescale for developing a new feature is longer than your timescale for releasing minor builds on your stable branch. That's why the branch gets the name experimental. This is a common requirement in lots of software, and not simply bad management.
How would you deal with this without branching and merging?
So as a cmd-line user who only uses centralized repositories (with only a couple of other developers), what kind of benefits would git provide over subversion?
So how would you handle maintaining a stable branch with a small amount of day-to-day changes at the same time as an experimental branch where live dev work is going on?
There are plenty of reasons why this is necessary, and it drives people to branch merge big chunks of codebases. What alternative method would you use?
Actually wikipedia tries not to select a version of the truth. The reason that most talk pages are filled with arguments and claims that articles lack NPOV is that the aim is to write something that is Neutral Point Of View. Rather than have one person select the truth according to their knowledge and biases, the idea is that all the people who argue about a topic can contribute something.
As long as they can back it up with a citation.
It's a low bar, but it's high enough to cut out 90% of the crap. As long as citations are provided then multiple view points can coexist in the same article. The point is to prevent somebody deciding on truth, to present the argument so that the reader can make up their own mind. To do so in such a way that there are links into the background literature that can be followed, or debate on the talk page to see which areas were the most contentious.
In short, it's the complete opposite of what you describe.
There is a link from the article to a good essay that he's written recently to explain the current meltdown in terms of a black swan.
Why would you blame "greedy" individuals? They were behaving rationally and trying to maximise their own profits. Surely the people who rigged the game to make their bets look one way have a lot more blame coming their way. Expecting people to make sub-optimal financial decisions so that they can collectively improve the situation at a macro-scale is silly. Besides, if anyone wanted to they are following a strongly dominated strategy and would be drowned out by the hoards of people searching for equilibrium. Those people were stupid, but that doesn't make them to blame for the current crisis. Hopefully there will be a massive wave of fraud charges amongst the bankers and estate agents who deliberately rigged the game. Sadly the central bankers and politicians who helped them do will get away with it.
As far as reaching bottom goes, have a look at the graph of the FTSE over the past 15 years and tell me that doesn't look like the mother of all double tops. Haven't checked the graph for the DOW but it will be similar. I'm sure people will continue laughing at you all the way down... :)
I see that you are one of the hard of thinking. Let's go through this slowly for you.
You asked a question about a rhetorical statement. I've explained to you what the statement was retorting to. Your question has no value because nobody was suggesting that China nuking America would help them get their debt back. Despite this you are still after an answer?
The answer is that you are a thief of oxygen. Perhaps you should sit quietly and try to understand what people are saying before you try and take part in the debate again.
Try reading the context that I replied to:
Who is holding all your debt right now (and they got nukes, so it counts)?
Dude, I ain't even from Iceland...
So you are claiming that something for which you have no evidence must be true, just because it cannot be disproved easily? Sounds like a religious nutjob to me.
Iraq got so much support in America, but not in Europe (where I live) for one simple reason. As a mass (or more accurately a mob) the citizens of America do not know the difference between patriotism, and nationalism, which is a long way down the slippery slope to fascism.
When you attribute causes to unseen entities to support your arguments, you are not just invoking religious paranoia - you are ignoring Occam's razor. There doesn't need to be a shadowy force behind American politicians seeking power over their neighbouring countries, and further control over the population. It can be done in plain view, because it is written into the rules of the game that define politics in America.
We will be the generation that watches the rise of USA Plc. It won't be because of a secret cabal of bankers pulling strings behind the scenes. It will be because corporations want politicians that don't stand in their way of making money through unfettered capitalism, and they can pay for it. Politicians don't want to stand in the way of corporate profits - because a strong economy is the surest way to re-election. Couple this with an expanding government - no government on earth has ever given up powers or control once they have seized them, and you have all the ingredients that you need.
A simple rule in life (call it Occam's rule for people) is:
Never attribute to malice what you can attribute to stupidity.
I'm sure you can understand the connection to shadowy conspiracy theories.
Thanks for the link. It's always entertaining to read the work of a grade A nutjob:
Shadowy figures in conspiracy behind the government.... Check.
Upcoming world war III and new world order.... Check.
One world government.... Check.
The only thing missing was that it forgets to blame the Jews. There are certain standards to keep up when writing this kind of trash.
You wish it was only $700B. Read the draft that is circulating. The new agency has a *net* asset cap of $700B at any one time. The amount of debt that it can issue to buy toxic funds is unlimited.
Can I just say a big thankyou from the rest of the world, as you seem to have agreed to bail out all toxic mortgage debt in the world. Quite amazing really, just how generous the US taxpayer is when nobody asks him first. Especially given the unlimited nature of the fund, sure the $700B cap is nice and all, but is your media reporting that as a total, or as the limit on the balance on the books at any one time?
If a datacenter is renting, rather than owning, their building then I can see that pushing up operating costs. As pointed out by the other reply. But money that you are not making is not an operating cost. Your other comment about additional cooling/heating is incorrect as this discussion was about trading increased space for better airflow to reduce heating/cooling costs.
Why does the increased space increase operating costs? I'd assume that the datacenter owns the land / building it is situated on / in. So that's a sunk cost, not a recurring one. Just because the building is bigger doesn't automatically imply higher maintenance costs, especially as it is less full. What is the other cost that increases?