Author: Mark Friedenbach 2014-01-07 00:21:25
Published on: 2014-01-07T00:21:25+00:00
In this message, Peter Todd comments on Thomas Voegtlin's Python-levelDB implementation of the UTXO (Unspent Transaction Output) hashtree. He notes that there are two tree structures being discussed, a 256-way PATRICIA trie and a bitwise non-level-compressed trie structure or proof-updatable trees. The latter enables stateless application of one proof to another, allowing mining and mempool operations without access to the UTXO set. However, performance is closely tied to key length, making H(scriptPubKey) the more desirable option. Todd suggests creating per-block indexes of all scriptPubKeys, spent or unspent, queryable via partial prefix method as it would be quite easy to do so.
Updated on: 2023-06-07T22:41:33.548806+00:00