coinstudent2048[, no idea really how it was decided. It was originally 0.3 xmr per block, but then block time was increased to 2 minutes, so it grew to 0.6. So I always think its better to say 0.3 xmr per block minute

because, who knows, at some point blocks could get longer again

gingeropolous: Yeah I think it's better to say 0.3 xmr per block minute. From what I read, the increase to 2 minutes is trade-off between fast transactions and scalability (blockchain grows slower) + stability when Monero becomes popular (less 'orphans'). I don't understand much of these details, but my questions are answered. Thanks!

* gingeropolous: Yeah now I think it's better to say 0.3 xmr per block minute. From what I read, the increase to 2 minutes is trade-off between fast transactions and scalability (blockchain grows slower) + stability when Monero becomes popular (less 'orphans'). I don't understand much of these details, but my questions are answered. Thanks!

a key reason to increase to 2 minutes was to reduce orphan rate

as for 1 minute blocks making the blockchain smaller, I don't see 2 150 kB blocks being smaller than 1 300 kB block

I could easily be missing something

No I read that the 2 minute blocks make "blockchain grow slower", but again I am not sure; I am not well-versed in this.

* gingeropolous: Yeah now I think it's better to say 0.3 xmr per block minute. From what I read, the increase to 2 minutes is trade-off between fast transactions and (scalability (blockchain grows slower) + stability when Monero becomes popular (less 'orphans')). I don't understand much of these details, but my questions are answered. Thanks!

yes, 2min blocks reduce the rate of orphaned blocks

Will growing block sizes potentially increase the number of orphaned blocks again?

I doubt it. block size doesn't really affect mining rates

block propagation is slower and miner with low network bandwidth will get more orphan blocks

I mean miner nodes

maybe. most blocks are fluffy blocks

My fork will have 30 second blocks (trading orphans for better game playability) so we can find out (though how network size affects that is unclear).

your fork?

I forked monero.

(to make a game on a chain, not a straight coin)

hmm, I thought the OpenCL code for RandomX was still radeon-specific

there's generic OpenCL code too, it works on NVIDIA

hyc: Does the advent of fluffy blocks mean block time could be reduced again?

If you can get deterministic transaction ordering, e.g. by fee, and pre-emptively sync mempools, it seems like you could get block times arbitrarily low

since you would only need the header and the delta from the target node's mempool

(i.e. if I know my neighbor's mempool at time T, I only need to send the delta from now to time T, and any "weird" transactions I noticed, and he'd figure it out)

If a scammer mines (first) a block with tx that double spends an output in a tx that the rest of the miners have, it doesn't matter does it.

"it doesn't matter" referring to "deterministic transaction ordering, e.g. by fee, and pre-emptively sync mempools".