3 ms·

> Yes. Peer-to-peer networks have existed long before Web3 ever became a thing. Bittorrent worked fine without a blockchain. IPFS works fine without a blockchai

> Yes. Peer-to-peer networks have existed long before Web3 ever became a thing. Bittorrent worked fine without a blockchain. IPFS works fine without a blockchain. Soulseek and Limewire didn't need blockchains, same as Napster. Fine, but for IPFS you need a normal domain name. Oops, now you're centralized/cancellable again. Every single person who thinks blockchain = database just doesn't understand centralization vs decentralization, and trust vs trustless. A database is run by a central authority.

zinekeller 4y ago
> Fine, but for IPFS you need a normal domain name. Huh? Who told you that? Sure, if you are lazy enough to rely on a translator (which exists for Tor and IPFS), domain names would be in the picture, but you could use clients... that do that directly.
- concinds 4y ago
  I meant to use IPFS for a "mainstream" site.
smoldesu 4y ago
You're saying nothing. I take it you're suggesting that DNS is bad, and should be replaced by a Blockchain. Fair enough, let's humor this funny little adventure you're suggesting. 1. The most immediate problem is going to be scaling. Since we can't rely on centralized peers (or even federated trackers like on Bittorrent), our ledger has to be both distributed and operate with relatively low block times. This introduces a problem: anyone who wants a "complete" ledger needs to trace their domain back through the history of the blockchain. This would require a local copy of the ledger on your client (extremely large), or server-side domain resolution (unbelievably slow, even for a single user). The internet as we know it simply wouldn't fit on a Blockchain. 2. Let's pretend like that isn't an issue though, and you only need to serve a few dozen sites to a few dozen users (the fate of many such crypto projects). Now we get to the exciting stuff; decentralized domain markets! I'm assuming this is where people step in to suggest smart contracts and NFTs get used, but there's a bigger problem on the horizon; lack of regulation. Since DNS is a centralized server, it can determine when abusive transactions are conducted and rectify them. If Microsoft decided to buy every domain NFT overnight, nobody could stop them. Now Microsoft owns your "decentralized" internet! Yippee! 3. Even though the mental gymnastics are getting exhausting, let's assume one more time that this project has perfectly worked out every social engineering kink and logical exploit in their market system. ...then what? Domain owners can now be openly hunted by observing the wallet their domain name lives in. Domains can be stolen like those poor Bored Apes. Companies can no longer take solace in a B2B transaction with their DNS provider, since that doesn't determine the ownership of their domain anymore. The logic of domains breaks down past this point, you may as well just use unique identifiers, IP addresses or hashes by now. The idea of a web as we know it wouldn't exist in this reality. It would be a lawless no-man's land, a reflection of the catatonic crypto landscape as it exists currently.

zinekeller 4y ago

> Fine, but for IPFS you need a normal domain name. Huh? Who told you that? Sure, if you are lazy enough to rely on a translator (which exists for Tor and IPFS), domain names would be in the picture, but you could use clients... that do that directly.

concinds 4y ago
I meant to use IPFS for a "mainstream" site.

smoldesu 4y ago

You're saying nothing. I take it you're suggesting that DNS is bad, and should be replaced by a Blockchain. Fair enough, let's humor this funny little adventure you're suggesting. 1. The most immediate problem is going to be scaling. Since we can't rely on centralized peers (or even federated trackers like on Bittorrent), our ledger has to be both distributed and operate with relatively low block times. This introduces a problem: anyone who wants a "complete" ledger needs to trace their domain back through the history of the blockchain. This would require a local copy of the ledger on your client (extremely large), or server-side domain resolution (unbelievably slow, even for a single user). The internet as we know it simply wouldn't fit on a Blockchain. 2. Let's pretend like that isn't an issue though, and you only need to serve a few dozen sites to a few dozen users (the fate of many such crypto projects). Now we get to the exciting stuff; decentralized domain markets! I'm assuming this is where people step in to suggest smart contracts and NFTs get used, but there's a bigger problem on the horizon; lack of regulation. Since DNS is a centralized server, it can determine when abusive transactions are conducted and rectify them. If Microsoft decided to buy every domain NFT overnight, nobody could stop them. Now Microsoft owns your "decentralized" internet! Yippee! 3. Even though the mental gymnastics are getting exhausting, let's assume one more time that this project has perfectly worked out every social engineering kink and logical exploit in their market system. ...then what? Domain owners can now be openly hunted by observing the wallet their domain name lives in. Domains can be stolen like those poor Bored Apes. Companies can no longer take solace in a B2B transaction with their DNS provider, since that doesn't determine the ownership of their domain anymore. The logic of domains breaks down past this point, you may as well just use unique identifiers, IP addresses or hashes by now. The idea of a web as we know it wouldn't exist in this reality. It would be a lawless no-man's land, a reflection of the catatonic crypto landscape as it exists currently.