Author: Andrea 2018-06-24 08:19:00
Published on: 2018-06-24T08:19:00+00:00
In a message to the bitcoin-dev mailing list, William Casarin suggested that including a human-readable part in the encoding of Payment Secret Transaction Batches (PSBT) would be a good idea. However, Andrew Chow disagreed, stating that it was unnecessary. Similarly, Peter D. Gray suggested that PSBT should be binary only and hex for developers and JSON interfaces, but Chow believes that the large size of PSBTs will make them cumbersome to move around as hex. Instead, he suggests using an encoding that most wallets already support, such as Base64. Chow also notes that the magic code is still necessary for the binary format of PSBT to prevent normal transaction deserialisers from accidentally deserialising a PSBT. While he agrees with Gray's suggestion to use a filename extension and a mime-type to match, he plans to update Appendix A of PSBT and define symbols for the numeric key values to assist implementers. However, he admits that the tables in the appendix are not working and he will attempt to fix them despite his frustration with Mediawiki. Finally, Andrea comments on the proposed changes, agreeing that having per-input key derivation and per-output data would be more convenient for signers and that the definition of global type could be removed entirely from the revised spec.
Updated on: 2023-06-13T03:19:12.605884+00:00