Author: Peter Todd 2017-02-25 04:12:02
Published on: 2017-02-25T04:12:02+00:00
The conversation revolves around a proposal to cluster entries by entry time in order to optimize performance. The proposed solution involves using position entries instead of hashes for the values being stored, with each entry receiving a value that is the binary representation of its position. The discussion then moves to the implementation and review of a merkle-set library, which is deemed complex and lacking sufficient explanation for its advantages. The benefits of an insertion-ordered data structure are discussed, with the author arguing that new coins are more likely to be spent and therefore warrant such a structure. The idea of optimizing a plain vanilla UTXO set solution before attempting the insertion ordering trick is suggested, as it would provide a meaningful basis for comparison. The author also points out that insertion ordering is not a simple trick, but rather a fundamental change to the data structure.
Updated on: 2023-06-11T21:38:45.406509+00:00