No scientific journal editor would even consider allowing a reference to an anonymous source.
On the other hand, authors' names are often removed from papers before refereeing, because editors and referees are supposed to judge a paper on its content rather than its authors' reputations. If you believe that journalism should be scientific in its methods, perhaps sources should be judged on the verifiability of what they say rather than on their reputations. Of course, sometimes the source's identity is indispensable - if the information cannot be verified directly then you have to fall back on reputation - but if an anonymous source leads to verifiable information, why should the source's identity be relevant?
The most desirable, in my opinion, would be some sort of content-neutral caching mechanism that would move data closer to consumers.
I agree - something like a caching HTTP proxy, but with files identified by location-independent URNs rather than URLs, would save a huge amount of bandwidth. The question is how to deploy it incrementally.
There's already a spec for URNs based on hash trees, so we could start with a browser plugin that handles hash tree URNs, fetching them from the same server as the embedding page if no proxy has been configured. Maybe hash tree requests should go to a different port so that ISPs can see how much transit traffic they'd save by installing a proxy.
The proxy should deal with IP addresses, not DNS names, so that P2P requests can be proxied.
How much is your bandwith worth to you? Would they follow a model like empornium where I would have to keep my share up? What if I didn't want to give out *any* bandwith would I still get to download?
I've been working on a quantitative answer to this question. (There are some notes here but they're a bit out of date.) My idea works like this: peers upload to their neighbours in order to obtain downloads from their neighbours (payment in kind, as in BitTorrent). When a peer has to choose which neighbour to upload to, it works out the number of extra bytes it's likely to be allowed to download as a result of uploading to each neighbour, and it uploads to the neighbour that offers the best value. (You can work out the expected value by keeping track of each neighbour's upload/download ratio, there's no need for neighbours to make explicit "bids".)
Uploading to a neighbour changes its upload/download ratio, so if you have several neighbours that are equally good providers then you'll upload to them in round-robin order; if some neighbours are better providers than others, they'll get a bigger share of your upload bandwidth. Neighbours that don't allow you to download will have the lowest priority. Essentially, neighbours bid for your bandwidth in a blind auction, but rather than offering currency they offer payment in kind, and rather than naming explicit prices they're assessed on the basis of their past behaviour. Unlike the Tit For Tat rule used by BitTorrent, this rule is motivated directly by self-interest, so it *should* be harder for selfish users to gain an unfair advantage (although I can't prove it yet).
It makes sense because it scales better than centralised distribution ever will. An ideal P2P network can serve a million users for the same cost to the publisher as serving a single user. Publishers don't care how many bits are transmitted overall, just how many bits they have to transmit personally, so BitTorrent makes sense for publishers.
Users also don't care how many bits are transmitted overall, just how many bits they have to transmit personally. BitTorrent allows each user to choose between leeching (which results in slow downloads) and sharing (which results in fast downloads), so BitTorrent also makes sense for users.
It's the comedy of the commons: nobody cares about overall efficiency as long as their own goals are being met.
you can use the awesome power of supply-demand curves and free markets
Why do libertarians always sound like infomercials?;-)
The advantage of BitTorrent's tit for tat algorithm over micropayments is that you don't need a currency - peers pay each other in kind, by uploading to those who allow them to download. Digital currencies need a central clearing house (or tamper-resistant hardware) to prevent double spending, but payment in kind is completely decentralised.
There's not much meat in the abstract of Clarke's Defcon presentation, and no clues on the Freenet site. Can anyone explain the new routing algorithm or point me to some documentation?
As far as I can tell from the wiki the new clipboard manager won't affect middle button pasting (which uses the "primary" selection rather than the "clipboard" selection). Perhaps you could change the mapping of your middle mouse button from 2 to 6?
Ingredients: useful features 10%, developers' pet features 10%, marketing-inspired features 17%, bugs 19% (approx.), legacy code that makes no sense but breaks things if we remove it 28%, documentation 24% (of which useful documentation 3%). Warning: may contain traces of Knuth.
The fingerprints will also be stored in a central database. (Personally I have no problem with the biometric aspect of the proposals - if everyone has to have an ID card then it should be linked as strongly as possible to the owner's physical identity; photos and fingerprints are preferable to DNA and iris scans because they can be verified by the bearer as well as the authorities.)
Part of the problem is that automation has made tracking and profiling easier than ever before. If everybody carried an ID card with an RFID chip it would be trivially easy to track people's movements, find out who they associate with, and round them up at short notice. You might not object to the current government having that kind of power, but countries can turn from democracies to dictatorships in a matter of decades. Measures like ID cards, which centralise knowledge and power, arguably accelerate the trend towards dictatorship because few people can resist the temptation to increase their power over others.
In my opinion we haven't been given a single good argument for ID cards. The often-quoted cost of ID fraud is mostly credit card fraud, which ID cards wouldn't fix; benefit fraud costs the country far less than well-known tax loopholes. Terrorism of the kind we've seen recently can't be prevented without draconian restrictions on freedom of movement, and nobody has explained how ID cards would reduce other kinds of crime.
People talk about the convenience of a government-approved photo ID, but passports already provide that function. In practice people don't carry them - they're rarely needed and it's inconvenient (and expensive) if they're lost or stolen. The same would be true of voluntary ID cards, and the inconvenience would be much greater for mandatory ID cards.
Even if you think I'm being paranoid about the potential dangers, why introduce ID cards when the benefits haven't been demonstrated and the costs are likely to be high?
On the other hand, arresting a North Korean or Chinese hacker would require the cooperation of an unfriendly government, whereas arresting a British hacker sends the same warning without ruffling any diplomatic feathers.
Those laws don't sound like a specification for a new system so much as a description of the current system - the "patchwork of identity one-offs" that the article sets out to criticise.
1) User Control and Consent - in the current system, users must create a separate identity for each site.
2) Minimal Disclosure for a Constrained Use - most sites don't require any information that leads back to your real identity, unless money's changing hands. Some require an email address for verification, but those are disposable.
3) Justifiable Parties - again, using a separate ID for each site limits the scope of the information you reveal.
4) Directed Identity - in the current patchwork, your email address or URL is your omnidirectional ID, and website accounts are unidirectional IDs.
5) Pluralism of Operators and Technologies - the patchwork has this by definition.
6) Human Integration - requiring a manual login at each site (rather than, say, single sign-on) keeps the human in the loop.
7) Consistent Experience Across Contexts - I'd say most users are familiar with the login/password/keep-me-signed-in routine by now.
You can create a record of your actions without actually tying that record to yourself. This way you give your anonymous speech more credibility without compromising your privacy.
A strong (cryptographic) pseudonym doesn't give anonymous speech any more credibility than a pen name would. Pseudonymous publishing is vulnerable to a man-in-the-middle attack: the attacker republishes everything you publish, under his own key. Once your ideas have built up some credibility, the attacker starts subtly modifying what you say. Readers see that the new works are signed with the same key, so they assume that they have the same author. (If anything, a strong pseudonym *damages* your credibility in this situation, because it creates a stronger assumption of identity than a pen name would.)
Interesting idea, but what's to stop people from making inaccurate predictions in order to influence policy? Obviously they'd lose money that way, but lobbyists are prepared to lose money if they can influence policy.
Maybe banks should start issuing their customers with USB tokens? Tokens are like smart cards - they can perform public key operations to verify the user's identity without leaking any private information to phishers.
Slashdot Mirror
Incidentally, the 37-foot diameter units described in the article generate 25 kW each - I wonder if they'd be suitable for domestic use?
Tell that to Ho Chi Minh, Che Guevara and Ahmed Shah Massoud.
Freedom is an illusion, designed as tool of manipulation for grumbling slave strata yearning for dominion.
If freedom is an illusion, what defines a slave?
On the other hand, authors' names are often removed from papers before refereeing, because editors and referees are supposed to judge a paper on its content rather than its authors' reputations. If you believe that journalism should be scientific in its methods, perhaps sources should be judged on the verifiability of what they say rather than on their reputations. Of course, sometimes the source's identity is indispensable - if the information cannot be verified directly then you have to fall back on reputation - but if an anonymous source leads to verifiable information, why should the source's identity be relevant?
I agree - something like a caching HTTP proxy, but with files identified by location-independent URNs rather than URLs, would save a huge amount of bandwidth. The question is how to deploy it incrementally.
There's already a spec for URNs based on hash trees, so we could start with a browser plugin that handles hash tree URNs, fetching them from the same server as the embedding page if no proxy has been configured. Maybe hash tree requests should go to a different port so that ISPs can see how much transit traffic they'd save by installing a proxy.
The proxy should deal with IP addresses, not DNS names, so that P2P requests can be proxied.
I've been working on a quantitative answer to this question. (There are some notes here but they're a bit out of date.) My idea works like this: peers upload to their neighbours in order to obtain downloads from their neighbours (payment in kind, as in BitTorrent). When a peer has to choose which neighbour to upload to, it works out the number of extra bytes it's likely to be allowed to download as a result of uploading to each neighbour, and it uploads to the neighbour that offers the best value. (You can work out the expected value by keeping track of each neighbour's upload/download ratio, there's no need for neighbours to make explicit "bids".)
Uploading to a neighbour changes its upload/download ratio, so if you have several neighbours that are equally good providers then you'll upload to them in round-robin order; if some neighbours are better providers than others, they'll get a bigger share of your upload bandwidth. Neighbours that don't allow you to download will have the lowest priority. Essentially, neighbours bid for your bandwidth in a blind auction, but rather than offering currency they offer payment in kind, and rather than naming explicit prices they're assessed on the basis of their past behaviour. Unlike the Tit For Tat rule used by BitTorrent, this rule is motivated directly by self-interest, so it *should* be harder for selfish users to gain an unfair advantage (although I can't prove it yet).
Users also don't care how many bits are transmitted overall, just how many bits they have to transmit personally. BitTorrent allows each user to choose between leeching (which results in slow downloads) and sharing (which results in fast downloads), so BitTorrent also makes sense for users.
It's the comedy of the commons: nobody cares about overall efficiency as long as their own goals are being met.
Why do libertarians always sound like infomercials? ;-)
The advantage of BitTorrent's tit for tat algorithm over micropayments is that you don't need a currency - peers pay each other in kind, by uploading to those who allow them to download. Digital currencies need a central clearing house (or tamper-resistant hardware) to prevent double spending, but payment in kind is completely decentralised.
Cuba is over 500 miles long.
There's not much meat in the abstract of Clarke's Defcon presentation, and no clues on the Freenet site. Can anyone explain the new routing algorithm or point me to some documentation?
I'm sure the CIA will pay for a balloon at the American end.
As far as I can tell from the wiki the new clipboard manager won't affect middle button pasting (which uses the "primary" selection rather than the "clipboard" selection). Perhaps you could change the mapping of your middle mouse button from 2 to 6?
If web pages still look blurry, add the following line to ~/.bash_profile to disable antialiasing in GTK2 and Gecko:
export GDK_USE_XFT=0
Which OS family has been ported to three out of four of those platforms?
Ingredients: useful features 10%, developers' pet features 10%, marketing-inspired features 17%, bugs 19% (approx.), legacy code that makes no sense but breaks things if we remove it 28%, documentation 24% (of which useful documentation 3%). Warning: may contain traces of Knuth.
It's OK, it's SVG, you can rescale it later.
So is Perl. :-)
The fingerprints will also be stored in a central database. (Personally I have no problem with the biometric aspect of the proposals - if everyone has to have an ID card then it should be linked as strongly as possible to the owner's physical identity; photos and fingerprints are preferable to DNA and iris scans because they can be verified by the bearer as well as the authorities.)
In my opinion we haven't been given a single good argument for ID cards. The often-quoted cost of ID fraud is mostly credit card fraud, which ID cards wouldn't fix; benefit fraud costs the country far less than well-known tax loopholes. Terrorism of the kind we've seen recently can't be prevented without draconian restrictions on freedom of movement, and nobody has explained how ID cards would reduce other kinds of crime.
People talk about the convenience of a government-approved photo ID, but passports already provide that function. In practice people don't carry them - they're rarely needed and it's inconvenient (and expensive) if they're lost or stolen. The same would be true of voluntary ID cards, and the inconvenience would be much greater for mandatory ID cards.
Even if you think I'm being paranoid about the potential dangers, why introduce ID cards when the benefits haven't been demonstrated and the costs are likely to be high?
On the other hand, arresting a North Korean or Chinese hacker would require the cooperation of an unfriendly government, whereas arresting a British hacker sends the same warning without ruffling any diplomatic feathers.
1) User Control and Consent - in the current system, users must create a separate identity for each site.
2) Minimal Disclosure for a Constrained Use - most sites don't require any information that leads back to your real identity, unless money's changing hands. Some require an email address for verification, but those are disposable.
3) Justifiable Parties - again, using a separate ID for each site limits the scope of the information you reveal.
4) Directed Identity - in the current patchwork, your email address or URL is your omnidirectional ID, and website accounts are unidirectional IDs.
5) Pluralism of Operators and Technologies - the patchwork has this by definition.
6) Human Integration - requiring a manual login at each site (rather than, say, single sign-on) keeps the human in the loop.
7) Consistent Experience Across Contexts - I'd say most users are familiar with the login/password/keep-me-signed-in routine by now.
So what needs fixing, exactly?
A strong (cryptographic) pseudonym doesn't give anonymous speech any more credibility than a pen name would. Pseudonymous publishing is vulnerable to a man-in-the-middle attack: the attacker republishes everything you publish, under his own key. Once your ideas have built up some credibility, the attacker starts subtly modifying what you say. Readers see that the new works are signed with the same key, so they assume that they have the same author. (If anything, a strong pseudonym *damages* your credibility in this situation, because it creates a stronger assumption of identity than a pen name would.)
Interesting idea, but what's to stop people from making inaccurate predictions in order to influence policy? Obviously they'd lose money that way, but lobbyists are prepared to lose money if they can influence policy.
This should bring down the cost of maintaining that huge undersea network of hydrophones and magnetometers that doesn't exist. ;-)
Maybe banks should start issuing their customers with USB tokens? Tokens are like smart cards - they can perform public key operations to verify the user's identity without leaking any private information to phishers.