Author: Eric Lombrozo 2015-08-24 17:58:55
Published on: 2015-08-24T17:58:55+00:00
The author of the context is torn on the filtering mechanism. While he supports connection filtering, he runs his own full node instances to connect to due to three fatal flaws: no mechanism for short proofs of tx nonexclusion, txout unspentness, block validity, poor privacy, and lack of incentives to run servers. He believes that BIP37 was necessary as a step towards a client/server architecture. However, he finds the filter mechanism useful because no "special" server is required. The author would rather make the distinction between trustless peers and a client/server model more explicit.The email conversation included in the context discusses the NODE_BLOOM and NODE_NETWORK codes. It is stated that NODE_BLOOM is distinct from NODE_NETWORK, and it is legal to advertise NODE_BLOOM but not NODE_NETWORK. This is not against spec, and proposals may make this more useful in the future. The suggestion to reword the statement about NODE_BLOOM is agreed upon by Wladimir J. van der Laan.
Updated on: 2023-06-10T20:45:33.668807+00:00