Laundry list of inter-peer wire protocol changes

In a discussion on the Bitcoin-Dev mailing list, Anthony Towns expressed his preference for Rusty Russell's construction of shachain over the existing one, citing that it would make it more convenient to work out the hash for any given index, as long as you know the seed. He also mentioned that the elkrem tree requires knowledge of the desired height before mapping to indexes. However, he did admit that it is not a significant difference and probably doesn't matter much. Towns also suggested that an explanation with a diagrammed n-way tree structure would be helpful. In response to a query about R value vs keypair, he stated that it should still be considered an R&D idea rather than trying to release it in v1.0. Regarding multi-sig txs, Joseph pointed out that by allowing more than one hash on commit txs, escrow-style services could be enabled. While there was some discussion about whether it was multiple signatures or hashes, it was ultimately agreed that it was multiple hashes.