Quote from: see360 on August 31, 2014, 09:30:06 PM

Quote from: bigreddmachine on August 31, 2014, 06:18:03 PM

My questions then I suppose were really these (again, I am genuinely interested in an asnwer, and not trying to FUD):

What are the repercussions of the much higher rate of orphans? Are there repercussions?
Can this result in more transactions being voided?
And finally, since all PoS implementations rely on centralized checkpointing, this means that the operator of the central checkpoint and those nodes closest to that central checkpoint are able to sync much faster than other nodes. What advantages does this give to those nodes in terms of # of Orphans compared to slower-to-sync nodes farther away?

Repercussions... I'm no expert, I'm just a small amateur miner, but I think:

1. higher orphan rate is shared equally by all JPC miners if the network is balanced, so no one miner is at a disadvantage. The orphan rate might be increasingly unfair if one miner has over half the network hashrate, but then there's a bigger concern that all coins have in this scenario.
2. In my experience with over 100 JPC transactions, I have never had a send/receive transaction voided, but I do get a high orphan rate, which is different from voided transactions. Both the orphaned block and the accepted block that goes into the block chain in its place should have the same transactions, just a different block got the privilege of being recognized by the network.
3. I must have a misunderstanding of the PoS implementation if there is centralized checkpointing. Maybe someone else could address this question.

In practice the short target block time intervals have resulted in extremely fast transactions (confirmations).

2) Not necessarily the same set of txes: first, block is created with existing ATM of creation set of transactions, next it is mined by HW, then if the solution found it is submitted. Orphan txes appears on the network usually when aside chain exist and is mined.

3) centralized checkpointing exist for 51% pos/pow attack prevention to make it impossible to override very long sub-chain. It is not good idea to immediately sync-checkpoint the most recently found block as this leads to appearance of many forks. It is much better to checkpoint the blocks that are N deep in the chain and allow the usual way the longest chain to win. Please note, that the "longest" is not in terms of number of blocks: for PoW/PoS coins the cumulative block trust value is accounted for the chain, for PoW only coins the cumulative work is accounted.