Bitcoin without bitcoins?

[DumbFruit]

So I've been thinking about this on and off for a while, and came back to my attention with the release of the libbitcoin-consensus library in 0.10.0 (Not that it actually has anything to with this.)

It seems to me that it should be possible to run a blockchain that is abstracted from the implementation details of any cryptocurrency that relies on it.

There would be at least two blockchains. One would be the Proof of Work blockchain and the Bitcoin blockchain.

The Proof of Work blockchain would be identical to the Bitcoin blockchain as it functions now except that it does not contain any transactional information, only a list of arbitrary hashes.

If someone wanted to provide Proof of Work for Bitcoin, they would be hashing the previous block header on the Proof of Work blockchain as well as the hash of a current block for Bitcoin (Along with any other hashes they'd want to include) + nonce. The Bitcoin block itself would also contain the hash of the block header of the Proof of Work block that included the previous Bitcoin block. It's possible that several Proof of Work blocks would not include any hashes for Bitcoin blocks.

[Proof of Work Block]
Previous Proof of Work block hash
CryptocurrencyBlockHashes[1,2,...,MAX]
nonce


[Cryptocurrency Block]
Proof of Work Block header hash that contained the previous Cryptocurrency block hash.
Transactional Information


The main advantages I see with this system is that it allows arbitrary experimentation in cryptocurrency designs while retaining PoW infrastructure, it creates a marketplace for Proof of Work hashes so that transaction costs can be computed by the actions of the market participants, and it allows centralization of cryptocurrencies while retaining decentralized consensus of their state. The Proof of Work provider could also include other services like "smart property", not just cryptocurrencies.

Transaction fees could be calculated by a cryptocurrency in any fashion they like and offered to the miner that includes it on the Proof of Work blockchain. Miners would simply reject fees that are too low to include the cryptocurrency block on the Proof of Work blockchain. I would think cryptocurrencies would employ strategies that provide increasing fees the longer it isn't included on the Proof of Work blockchain.

I think a Proof of Work miner would need to run a full node for any cryptocurrency it's providing work for. Though the miner could choose any arbitrary combination of cryptocurrencies.

Thoughts?

Update 5/13: Well I guess this was too stupid to consider, or else I didn't explain well enough how this is different from merge mining? What I described is almost the same as merge mining, the difference is that there is no need to pay miners in any particular cryptocurrency. In merge mining, bitcoins are required in order to pay fees, but in what I described the way the miner is paid can be in any form the miner wishes which could be bitcoins, or any other cryptocurrency, or even fiat if you wanted to do that for any reason. The only requirement to add to the Proof of Work blockchain is proof of work, aka; A hash meeting difficulty requirements.

[virtualx]

So I've been thinking about this on and off for a while, and came back to my attention with the release of the libbitcoin-consensus library in 0.10.0 (Not that it actually has anything to with this.)

It seems to me that it should be possible to run a blockchain that is abstracted from the implementation details of any cryptocurrency that relies on it.

Correct, you can create a version that is abstracted from the implementation. I do think this would not be for all alts because they can have very big differences and thus the level of abstraction would be very high.

The main advantages I see with this system is that it allows arbitrary experimentation in cryptocurrency designs while retaining PoW infrastructure, it creates a marketplace for Proof of Work hashes so that transaction costs can be computed by the actions of the market participants, and it allows centralization of cryptocurrencies while retaining decentralized consensus of their state. The Proof of Work provider could also include other services like "smart property", not just cryptocurrencies.

This allows for experimentation but less than with an alt, because you are limited to the protocol that exists.

[DumbFruit]

I read back up on merge mining, and actually since you can do fee-less blocks it's possible to do this in the existing Bitcoin protocol, but it's different from Merge mining in these ways;

1.) Merge mining does double mining, whereas what I'm describing all mining resources are directed at the Proof of Work blockchain. The  cryptocurrency blocks are built at the same time as a Proof of Work block.

2.) There is no pretense of being a Bitcoin block, so a dedicated Proof of Work block would be more efficient. No transactional information or scripting would be included.

3.) Bitcoin would be only optionally supported by miners.

The main advantages I see with this system is that it allows arbitrary experimentation in cryptocurrency designs while retaining PoW infrastructure, it creates a marketplace for Proof of Work hashes so that transaction costs can be computed by the actions of the market participants, and it allows centralization of cryptocurrencies while retaining decentralized consensus of their state. The Proof of Work provider could also include other services like "smart property", not just cryptocurrencies.

This allows for experimentation but less than with an alt, because you are limited to the protocol that exists.

The only limitations that I'm aware of is that the cryptocurrency must be a Proof of Work cryptocurrency (Which is the only type of cryptocurrency that achieves decentralized consensus that I know of.) and the confirmation times would be longer. Are there other limitations?

Sidechains are more strictly limited by design. Merge mining, and this slight twist on it, are much more lenient.