5 ms·

It would be valuable for a large entity to censor content that would make it look bad; by disrupting operations & breaking existing links to content.

dane-pgp 8y ago
This looks like a job for: Blockchain! Seriously, though, imagine a public, distributed, append-only log, containing the hashes of the content stored by archive.is, with the data effectively signed by a (e.g. wallet) private key of the archive maintainers. People could volunteer to locally store copies of all the data that was archived, but only serve the data that goes "missing" from archive.is. With such a dead man's switch in place, there would be no harm in archive.is taking down any content that they were threatened (legally or otherwise) over, and then other sites (hopefully too many to attack) could publish the taken down content, alongside the cryptographic proof that the data was originally published by archive.is. You're right though that there would still be disruption, as the original archive.is links would (in the worst case) return unhelpful error pages.
- beaconstudios 8y ago
  or archive.is could just post regular signed backups. No need for blockchain, as there is no need to form a consensus in this case.
  - dane-pgp 8y ago
    Yes, regular signed backups would be almost as good. What a blockchain adds is the inability to fraudulently backdate material (in case a key was compromised) and a financial incentive for making sure that at least the existence of the hash remains a public record (unless the value of the entire blockchain becomes zero, which is more likely for some blockchains than others). I suppose I'm inspired by this idea: https://www.reddit.com/r/Bitcoin/comments/6aiysw/the_whole_internet_archive_timestamped_with/ https://www.reddit.com/r/Bitcoin/comments/6aiysw/the_whole_i...
    - beaconstudios 8y ago
      a blockchain doesn't uniquely add that - there are cryptographic constructs that prevent that sort of attack, such as (I believe, I'm not a cryptographer) merkle trees that enforce a causal ordering. Really, the cost of a blockchain massively outweighs the value it brings to the table. The only instance where a blockchain is needed is where you need a tamper-proof distributed consensus between untrusted parties.
      - dane-pgp 8y ago
        Yes, you can create a merkle tree which enforces a causal ordering, but anyone with the (potentially compromised) key can generate such a tree with whatever ordering they want (or missing items). I concede, though, that the existence of multiple signed orderings would at least prove that the key had been compromised. You're also right that there is a huge cost to creating a secure blockchain from scratch, but the incremental cost of using an already secure public blockchain (such as Ethereum's) can be lower than the amount of value it brings to the table. Proving the existence of data, at specific timestamps in the past, when keys may have later been compromised, does seem like a use case that requires tamper-proof distributed consensus between untrusted parties. Whether that use case is particularly valuable or not is harder to judge, though, I admit.
- skywhopper 8y ago
  This misses the point of censorship. A party determined to sow misinformation can hijack a domain to do so. All the blockchain in the world won't solve that problem.
  - MorrisofOrange 8y ago
    That why we will solve censorship by mining my blockchain node hash with a sharding fork! /s
  - DoctorOetker 8y ago
    it's hilarious, to underscore the fragility of the conventional domain name system under pressure of politics and money, and portray it as an argument against blockchains,... this is literally what Namecoin (probably the second significant blockchain proposal) was designed for! of course lots of time has past now, and better mechanisms have been formalized (especially Algorand)...

dane-pgp 8y ago

This looks like a job for: Blockchain! Seriously, though, imagine a public, distributed, append-only log, containing the hashes of the content stored by archive.is, with the data effectively signed by a (e.g. wallet) private key of the archive maintainers. People could volunteer to locally store copies of all the data that was archived, but only serve the data that goes "missing" from archive.is. With such a dead man's switch in place, there would be no harm in archive.is taking down any content that they were threatened (legally or otherwise) over, and then other sites (hopefully too many to attack) could publish the taken down content, alongside the cryptographic proof that the data was originally published by archive.is. You're right though that there would still be disruption, as the original archive.is links would (in the worst case) return unhelpful error pages.

beaconstudios 8y ago
or archive.is could just post regular signed backups. No need for blockchain, as there is no need to form a consensus in this case.
- dane-pgp 8y ago
  Yes, regular signed backups would be almost as good. What a blockchain adds is the inability to fraudulently backdate material (in case a key was compromised) and a financial incentive for making sure that at least the existence of the hash remains a public record (unless the value of the entire blockchain becomes zero, which is more likely for some blockchains than others). I suppose I'm inspired by this idea: https://www.reddit.com/r/Bitcoin/comments/6aiysw/the_whole_internet_archive_timestamped_with/ https://www.reddit.com/r/Bitcoin/comments/6aiysw/the_whole_i...
  - beaconstudios 8y ago
    a blockchain doesn't uniquely add that - there are cryptographic constructs that prevent that sort of attack, such as (I believe, I'm not a cryptographer) merkle trees that enforce a causal ordering. Really, the cost of a blockchain massively outweighs the value it brings to the table. The only instance where a blockchain is needed is where you need a tamper-proof distributed consensus between untrusted parties.
    - dane-pgp 8y ago
      Yes, you can create a merkle tree which enforces a causal ordering, but anyone with the (potentially compromised) key can generate such a tree with whatever ordering they want (or missing items). I concede, though, that the existence of multiple signed orderings would at least prove that the key had been compromised. You're also right that there is a huge cost to creating a secure blockchain from scratch, but the incremental cost of using an already secure public blockchain (such as Ethereum's) can be lower than the amount of value it brings to the table. Proving the existence of data, at specific timestamps in the past, when keys may have later been compromised, does seem like a use case that requires tamper-proof distributed consensus between untrusted parties. Whether that use case is particularly valuable or not is harder to judge, though, I admit.
skywhopper 8y ago
This misses the point of censorship. A party determined to sow misinformation can hijack a domain to do so. All the blockchain in the world won't solve that problem.
- MorrisofOrange 8y ago
  That why we will solve censorship by mining my blockchain node hash with a sharding fork! /s
- DoctorOetker 8y ago
  it's hilarious, to underscore the fragility of the conventional domain name system under pressure of politics and money, and portray it as an argument against blockchains,... this is literally what Namecoin (probably the second significant blockchain proposal) was designed for! of course lots of time has past now, and better mechanisms have been formalized (especially Algorand)...