1
1
It's easy to see that currently 50% of the hashrate supports Segwit.
It's also simple to find that about 1000 nodes are supporting it. But I can't seem to figure out how significant that amount is, since all the total node counts are very rough estimates, and not all nodes are public.
So what is the best way to find out how solid the support under non-mining nodes is right now?
1
The total node count is not relevant. It's easy to spin up nodes, or even fake nodes that just report a certain user agent string. That doesn't mean that they care about the rules, plan to keep following them if the hashrate majority goes another way, whether they'd even notice if they're forked off, or even whether they actually implement the rules.
What does matter is how much economic activity is following the fork's rules, and whether they're planning to keep doing so regardless of what happens.
Measuring that is something you cannot do inside the protocol. You need to look at company statements, press, ...
Just an idea.. Would it be possible to reliable track user support if each transaction contained a signal instead of each block? That would require fake nodes to loose money on fees. – Muis – 2017-07-13T19:40:18.893
That sounds like requiring a soft fork in order to track a different soft fork. Furthermore, it doesn't solve the problem. Fake nodes wouldn't be sending any transactions, so how would they incur fees? – Jestin – 2017-07-14T02:16:59.090
@Jestin If fake nodes don't send any transactions they also have no voting-rights, because 1 transaction = 1 vote. – Muis – 2017-07-16T15:07:12.100
@Muis, it's not a node creating a transaction that is effectively a vote, it's a node relaying a block or transaction. Relay nodes don't pay fees to relay other people's transactions. – Jestin – 2017-07-16T15:11:32.767
@Jestin Relays? I think you're misunderstanding my proposal. It doesn't require a softfork, everything stays compatible, and offcourse relay nodes do not get voting rights this way, why should they? – Muis – 2017-07-16T15:16:08.107
A USAF is complicated, especially when trying to think of it in terms of voting. Not all votes are equal. For example, a single exchange's vote is probably worth more than nearly everyone else's (non exchange) votes combined. Would counting support on their transactions really be a good measurement? It's more about if miners depend on an entity than how many transactions they do. – Jestin – 2017-07-16T15:32:46.177
@Jestin Exchanges are responsible for a very large share of the transactions (every withdrawal), even though technically its not their BTC, so yes I believe that would be a reasonable measurement. But let's continue here https://bitcoin.stackexchange.com/questions/56734/why-are-transactions-not-used-for-signalling
1
No measurement is going to give you the full picture due to sybil attacks, incomplete datasets, data source abstraction/summarization, etc.
However, there are resources out there that are trying to make it easier to see through the fog:
Nodes: https://coin.dance/nodes
Blocks (miners): https://coin.dance/blocks
Poli (companies): https://coin.dance/poli
1
does that help ? https://coin.dance/blocks
– Sufiyan Ghori – 2017-07-13T18:38:33.983