@go1111111
I think we basically agree that insofar as the UASF idea is about economically important people/groups/businesses having an effective vote via their economic weight, it is sound, but that whether those entities run listening nodes is of minor importance at best.
As far as I can see, an economically significant entity who wants policy X isn't substantially hindered in their ability to leverage their economic influence to bring about policy X by failing to run a listening node that holds to policy X nor is their influence substantially reinforced by running one (as opposed to a listening node with a policy that always agrees with the mining majority, or even just an SPV wallet). They can exert their influence in other far more powerful ways as we are seeing with exchanges and other businesses taking sides, investors trading the forks, website owners censoring or promoting, whales funding development, and indeed miners making choices.
On network topology, the jargon is detailed mainly because there are a lot of different shapes a network can take on, and there are certain results, like the finding about Sybillability in
On Bitcoin and Red Balloons, that depend on the specific topology. It becomes less necessary if we are working with images, and I would like to move in that direction, but on the main it's probably sufficient to just distinguish between two network types: mesh and (nearly) complete graph.
Mesh topologies look somewhat like a mesh, i.e., like the screen-door-shaped distributed topology in the centralized/decentralized/distributed network diagram famous in the Bitcoin world. I have seen this mentioned by key figures like Greg Maxwell and it's pretty easy to see all sorts of small-block claims starting to make sense when you look at the diagram and imagine that it approximates the Bitcoin network: easy to Sybil, fragile without many "nodes," easy to partition, hard to ensure consistent and speedy propagation of even modestly big blocks, fraud proofs needed for SPV, zero-conf is hopelessly broken, contentious hard forks are dangerous, not antifragile, etc.
These issues are radically different when the topology is close to a complete graph, where every node* is connected directly or almost directly to every other node as a consequence of protocol rules (a few of which might have since been removed because Core saw them as superfluous) as well as economic incentives driving network resources to be concentrated toward fast connectivity to the mining nodes that have proven most prolific.
Bitcoin truly
cannot be a mesh, because if it were magically made into a mesh the network would spontaneously reorganize into a near-complete graph because of the incentives. Making it more mesh-like as Core imagines it by adding a ton of miner-mimics (listening nodes) would itself
be an attack, from which Bitcoin is designed to self-heal.
*Further confounding the matter is an equivocation between the mining network (with its mining nodes) and the larger listening network (with its listening nodes), which I think is rooted in a misunderstanding of the paramount importance of mining incentives. This accounts for the "small blockers already know mining network is tightly connected"...they just want to call that "not the Bitcoin network." The worldview they have introduced since Satoshi left is a tangle of word games and interlinking circularities.
I have a lot of experience with this phenomenon in other fields, but I think people seriously underestimate how
once you get a faulty assumption baked into a system it starts becoming self-reinforcing. Like a government subsidy, people adjust to it and it soon looks like it would be impossible to do without. It becomes very difficult to disentangle. "Who would build the roads?"
The only way to root out a potentially faulty assumption is to explore at length the possibility that the assumption was faulty from the start and see if the whole system makes more sense when reconstructed from the ground up without it. This is the proper way to evaluate a new paradigm without status-quo bias.
Here this means evaluating the full package as a package, that is, examining each unorthodox piece in turn with the provisional assumption that all the other unorthodox pieces are true (instead of false): that Bitcoin is a mining network and therefore a near-complete graph, that SPV is secure without fraud proofs, that a glut of listening nodes is in fact a burden on the network and doesn't affect governance, that there is no defense against irrational or non-profit-seeking miners except a PoW change (and even that is questionable), etc. Then see whether the whole thing hangs together better than the present orthodoxy, or not. And it's a nice bonus if it also comports exactly with Satoshi's original vision.
