You have to first show that you can see from his perspective.
I saw that perspective and I explained why that worldview is wrong. I don't understand why I should support a point of view that's mathematically invalid.
Is this true?
More or less. You can of course shrink it to just one hash (like it's done on ETH) and store the file itself separately, but even in this case all you own is the hash which isn't scarce and proving that it corresponds to the bytes created by a certain person requires external sources of truth. Therefore a hash isn't by itself sufficient to establish the true ownership and it shouldn't be on blockchain as well because it's just spam.
Very good. Yes, standard practice is to "shrink" the images by hashing. Thus you obtain another token for your image, in fact a non-fungible token for your image, because the hash was a cryptographic one. This means that the hash is irreversible, ownership of the images hash-token is the same thing as just ownership of the image. There is no change in scarcity at this stage.
I think we may have found a critical point of confusion here, in speaking of hashing. When I (or the smart contract) hashes an image, we are scrambling the image with an asymmetric hashing. It's not a simple checksum, is what I mean. So hashing the image can only be done with the owners public key, and the resulting hash is hers to unlock— she needs no external oracle, she has her private key.
Are we still in agreement?
reply