Published on: 2015-08-30T21:10:18+00:00
In an email conversation, Peter R and Daniele discuss the optimality of a greedy algorithm mentioned in a paper. Peter agrees with Daniele that the algorithm is asymptotically optimal in a certain case. However, Peter points out that "non-greedy" solutions to the knapsack problem are more relevant when choosing from a smaller number of items with similar sizes to the knapsack itself. Nonetheless, Peter argues that for the average transaction size and block size limit in Bitcoin, quantization effects become small, making the greedy algorithm a reasonable approximation.Tom Harding joins the discussion and informs Daniele that the optimal way to choose transactions for inclusion in a block is a knapsack problem, which is NP-complete. Tom refers to Peter R's analysis on fee markets without blocksize limits, which shows that the advantage of large miners is a side-effect of coinbase subsidization. As block rewards decrease, so will the advantages of large miners. Tom explains that larger blocksizes, as proposed by BIP100-101, aim to increase transaction fees by expanding the network's capacity and reducing mine centralization.In a bitcoin-dev mailing list, the risks of a contentious/schism hardfork are discussed. A document is shared to improve understanding of these risks. The discussion includes Peter_R's analysis on fee markets with uncapped max blocksizes, where it is noted that critiques made towards his work were not definitive. Simulations show that larger miners have advantages due to coinbase subsidization, but this advantage diminishes as block rewards become comparable to transaction fees. It is suggested that raising the max blocksize in a controlled manner can ease the transition into the fee market dynamic. The main critique of uncapped max blocksizes is the advantage that large miners have over smaller ones, but this advantage is attributed to block subsidies. Once block rewards become comparable to transaction fees, this advantage diminishes.The writer expresses concern over the acceptance of complex miner voting protocols as a hard fork option, while considering Mike and Gavin redirecting the Bitcoin protocol as risky. They believe that Peter_R's analysis on fee markets without blocksize limits should be given more weight by the community. The writer suggests controlled increases to the max blocksize to gradually transition into the fee market dynamic. They argue that advantages of large miners do not come from propagating large blocks but rather from block subsidies. As block rewards decrease, these advantages diminish. The writer hopes this matter is considered seriously, especially in the upcoming scaling workshops.Overall, the discussion revolves around the optimality of algorithms, the impact of blocksize limits on fee markets, and the advantages of large miners in the context of block rewards and transaction fees. There is a focus on the potential solutions proposed by BIP100-101 to increase transaction fees and reduce mine centralization.
Updated on: 2023-08-01T15:41:15.474273+00:00