Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

By my understanding, this system requires every transaction to be published to everyone, and to be remembered forever. That sounds like it inherently can't scale very far.

Details:

The basic data structure is a hash-chained list of blocks of transactions, with each block including one transaction of XX bitcoins from /dev/null to whoever generated that block. A node trying to calculate a next block include all (valid) transactions it knows. If the list branches, nodes ignore the shorter branch (this makes transactions more secure as they get older, assuming the swarm has more computing power than any troublemakers; so the time for a transaction to "clear" is whenever you trust that there are enough later blocks). In order to know if a transaction is valid, the node needs to have the prior transaction(s) that the involved bitcoins came through last (and any other subsequent transactions of that prior one). That prior transaction could be arbitrarily far back.



1. There had been talk of pruning information that the blockchain don't need, especially information that got confirmed a gazillion time.

2. There had been talk of a lightweight client that download the block header or something.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: