Author: Alan Reiner 2014-02-12 18:21:23
Published on: 2014-02-12T18:21:23+00:00
The discussion revolves around two things - tracking by inputs and outputs and defining a new canonical transaction hash/ID. The absence of I/O-based tracking affects bookkeeping, as outgoing transactions cannot be properly associated with customer accounts/actions. While a canonical ID can prevent some buggy behavior, it does not address the issue of mutation that invalidates a chain of unconfirmed transactions. It is not safe to double-spend the original transaction without atomic exclusion, as both the replacement and the original could be confirmed. The solution lies in designing software to use static IDs instead of full transaction hashes.
Updated on: 2023-06-08T02:38:20.065770+00:00