Author: Rusty Russell 2019-09-27 02:08:27
Published on: 2019-09-27T02:08:27+00:00
In this email exchange, Rusty provides feedback on a draft. He questions the use of "requirements" in reference to data and suggests that actors have requirements instead. Rusty asks for clarification on how to encode q, which is a coefficient used to estimate set difference. He also requests clarification on whether it's one or two sets of truncated IDs per peer. Finally, he recommends removing the phrase "To the best of our knowledge," from a statement about PinSketch being more bandwidth efficient than IBLT since he has implemented and experimented with IBLT and found it to be worse.
Updated on: 2023-06-13T21:29:26.037931+00:00