I am not for or against this, I will just describe some details about how this could work.
One way is to require that zindex=1 and if it isn't, don't render the data. The other way is to use our RPC's which dynamically calculate the anonset. Those can be slow for calculating the global anonset, but we can make them fast by embedding the data into the SD code. For instance, let's say we know the anonset at block H1 is X.
Now we want to update the current anonset at the current blockheight H2. We only have the calculate the anonset between H1 and H2, and then add it to X. This is much faster than calculating the global anonset in a loop, which is very slow and wasteful of CPU resources. And eventually it will be slow enough that the RPC method will timeout.
I am not for or against this, I will just describe some details about how this could work.
One way is to require that `zindex=1` and if it isn't, don't render the data. The other way is to use our RPC's which dynamically calculate the anonset. Those can be slow for calculating the global anonset, but we can make them fast by embedding the data into the SD code. For instance, let's say we know the anonset at block H1 is X.
Now we want to update the current anonset at the current blockheight H2. We only have the calculate the anonset between H1 and H2, and then add it to X. This is much faster than calculating the global anonset in a loop, which is very slow and wasteful of CPU resources. And eventually it will be slow enough that the RPC method will timeout.
As was mentioned in #105 the anonset info would be cool to have.
I am not for or against this, I will just describe some details about how this could work.
One way is to require that
zindex=1and if it isn't, don't render the data. The other way is to use our RPC's which dynamically calculate the anonset. Those can be slow for calculating the global anonset, but we can make them fast by embedding the data into the SD code. For instance, let's say we know the anonset at block H1 is X.Now we want to update the current anonset at the current blockheight H2. We only have the calculate the anonset between H1 and H2, and then add it to X. This is much faster than calculating the global anonset in a loop, which is very slow and wasteful of CPU resources. And eventually it will be slow enough that the RPC method will timeout.