Quote from: Jutarul on November 02, 2012, 05:57:27 PM

Quote from: gmaxwell on October 20, 2012, 05:43:56 PM

I've seen some POW ideas that I find intriguing for other reasons like amiller's idea of using fast random access to the txout set as proof of ability to rapidly process Bitcoin transactions but I've not seen anything that prevents people with custom hardware from getting a substantial advantage over the public except by getting specialized hardware into the hands of the public.

I second that. We'd need more ideas of how to implement a fair POW system which cannot be easily corrupted. Satoshi ruled against Proof-of-IP, because it's even easier to attack.

A fundamentally fair system rewards the miner by their contribution to the network (in terms of transaction processing and related services). The difficulty increase does not add any beneficial factors to the bitcoin network, except resilience against a 51% attack. Thus the focus should lie on how to implement resilience against 51% attacks without the consumption of a lot of energy.

Well we have the PPCoin experiment running, so we can see if proof of stake is a way forward, but they still use PoW for minting.

The other thing that PoW does that some folks forget, is it gives us an unhackable time source that does not rely on external time servers or anything else that could potentially be spoofed. You know locally when each block happened, so you have a reliable timestamp, and don't have to trust anyone. Moving to a pure PoS system would remove that timing mechanism, which is why PPC is a hybrid and was initially launched with check-pointing servers until the PoW activity became high enough.

You have to have a constraint that can reliably be applied to all members of the network. There are a lot of choices, but relatively few that are immune to manipulation. Bitcoin is based on pure CPU PoW, LiteCoin uses CPU and cache/memory IO, and PPC uses pure CPU as it's constraint for block generation to give us this timestamp. Other theoretical constraints include network latency or bandwidth (potentially manipulable by building well connected private clouds, but there could be some benefits to this like the Swiss are doing,) persistent storage, or even real-world physical token passing (and 2 guys pass a bag of tokens back and forth as fast as possible to do a 51% attack.)

When I start thinking about a potential future where we want to control power usage, I'm noticing that really what you are talking about is "increasing the cost of mining equipment so that power cost/performance parity is reached earlier" So instead of an ASIC that costs $1300 and uses 1W/GH, you have a computer that costs $2500 and uses 5W/(GH equiv) of power. Assuming that folks like bitcoin and keep mining, it would end up being a completely power neutral move that would have no advantage in security, minting, or timekeeping. (my opinion, but I think the conclusion is pretty obvious)

The good news is that you are stressing out over a short term issue that does not need to be solved because Satoshi designed the solution into it. Block halving will make this look like the 1849 gold rush (http://en.wikipedia.org/wiki/California_Gold_Rush) once the reward drops to a level that is not economic for all but the most efficient operations. These larger and more efficient operations will be motivated by profits, but will likely be more interested in maintaining the network security to ensure bitcoin profits in other endeavors rather than directly from mining itself. I would not be surprised at all if the BitCoin world of 2030 is almost 100% institutional (finanacial houses/banks) mining, with partnering agreements on planned hash levels to help reduce costs, and a big warehouse full of standby gear that can be powered on if a rouge does appear on the network.

So even if PoW is not ideal, it still seems to be the best solution, and I don't see any real advantage to changing it from a simple solution to a more complex one if it does not bring a substantial additional value.