[fiatjaf]

The innovation of blockchains is not on the how-to-store-things side, but how to keep a state of things every nodes agree with.

Stellar can store data on Postgres because it is just a small database of how much money each account has at each ledger. Past ledgers can be erased from the database (which makes it not a blockchain in any sense anymore).

Bitcoin has a history of all transactions organized in blocks not because it's a fancy new database technology, but because that way is the way it worked out better to keep a synced state between nodes, kind of an append-only log.

You could, if you wanted that, read the Bitcoin database and translate it into a set of rows of a who-has-how-much Postgres table.

Or, better, you could implement a Bitcoin client that stored its blocks as rows in a Postgres table, but I think it would be much more resource-intensive than the databases the Bitcoin clients are using today.

[runeks]

The Bitcoin blockchain is a compact serialization format used as the input to the proof-of-work function. The proof-of-work is the important part, and for that we need a centrally, agreed-upon serialization format where the latter blocks inherit PoW from earlier blocks.

A Bitcoin node can store blocks/transactions in whatever way it wishes -- Postgres, Dropbox, SQLite DB on a floppy disk -- as long as it's able to deliver to nodes in the canonical serialization format, because it's needed to verify proof-of-work (and signatures for the transactions).

[Frogolocalypse]

> Past ledgers can be erased from the database (which makes it not a blockchain in any sense anymore).

A block snake? That sounds catchy.