mleku on Nostr: this all relates to stuff i've said about how i think we can solve the network ...
this all relates to stuff i've said about how i think we can solve the network intelligence problem of lightning
1. channels should be advertising their balance as a single bit, meaning one side is up = true, if balance is other side, then false
2. channels should advertise a single value indicating the range of their total size, like, each number of that value represents let's say 20000 sats, so 50 means 1Msats, or maybe make it powers of 2 so each additional bit is 2x as much as the previous, so 2 means up to 40ksat, 20 means 2^20*20ksat (idk, if i sat down and thought about a good scale i could do it, but in general 100k is minimal channel, mainly because average payment is somewhere around 20ksat, so yeah
you cound wrap it into a single 8 bit value, or 2 hex chars, and make it a signed integer so it gives you both direction and size range
it would be voluntary, you can send nothing, but clients constructing payment routes could decline to use them for this case, and this would allow a lot of avoiding of shit small channels, which likely is what contributes to this 20% failure rate
1. channels should be advertising their balance as a single bit, meaning one side is up = true, if balance is other side, then false
2. channels should advertise a single value indicating the range of their total size, like, each number of that value represents let's say 20000 sats, so 50 means 1Msats, or maybe make it powers of 2 so each additional bit is 2x as much as the previous, so 2 means up to 40ksat, 20 means 2^20*20ksat (idk, if i sat down and thought about a good scale i could do it, but in general 100k is minimal channel, mainly because average payment is somewhere around 20ksat, so yeah
you cound wrap it into a single 8 bit value, or 2 hex chars, and make it a signed integer so it gives you both direction and size range
it would be voluntary, you can send nothing, but clients constructing payment routes could decline to use them for this case, and this would allow a lot of avoiding of shit small channels, which likely is what contributes to this 20% failure rate