Tru-therium

[psztorc]

Into (reader can skip this)

Everyone has been picking on Etherium lately. To me, this suggests a growing fear that Eth may *gasp* be extremely useful, and *double-gasp* might, possibly uniquely have the ability to challenge and destroy Bitcoin's hard-won network effects.

Just for kicks, I'm going to propose my own Ethereplacement : Trutherium! No, its not a joke, but its pretty close because I have essentially no cryptography experience and barely-acceptable coding skills (especially compared to this crowd). Honestly, though, I've mostly been waiting patiently to say "Trutherium" for a few days now...they both contain "th"! My joke idea should be easy to criticize if you're knowledgeable, so please don't hold back.

Motivation
Gavin's Bit-Therium, is pretty cool (the same Bitcoin we know and love, with Etherium). However, he assumes a majority of trustable oracles, which I don't like. Oracles should have to buy-in/sell-out to discourage Sybil/retirement attacks, should suffer if they try to attack the network (successfully or otherwise), shouldn't know what other oracles plan to do, should have an incentive to hide what they plan to do..

Oh, I'm describing part of my own project there.

Sketch of the Idea
With Truthcoin, we have reports combined into a consensus via eigenvector. The economics of it were finely-tuned to prediction markets, but can substitute PM-States for contract-states, LMSR for arithmetic-accounting, and the user-generated 'Votes' for contract resolution reports made by oracles. Complex payouts could be built into the Market section, with several Decisions being assembled Wall-St-style into 'derivative' contracts (to build polynomial payoffs for example). The fee structure could be set to auto-.5 if not enough fees had been paid, or to predominantly rely on the listing fees (not trading fees).

Wednesdays are busy for me...might write more about this later. Again, I don't take this very seriously, having watched mission-creep steal the life out of many a good project. I just thought I would join the club, because I don't believe in trusting oracles.

[somnicule]

What do you think the benefits of this would be, compared to implementing VoteCoin-style functionality via ethereum contracts? I've played around with the code a little, and it seems possible (and relatively simple) to implement a reputation currency, a contract to create and register decisions, and a contract to manage the ballots and voting. I've made a little headway with regards to this already, though it's likely I'll come across some obstacles.

[psztorc]

This idea is on top of Truthcoin. Ie, "Ethereum by using Truthcoin".

So you're saying 'what would be the benefits of this, compared to "Truthcoin by using Ethereum" '. The answer is that, with this, you don't need Ethereum at all.

[zack]

Explanation of truthereum:
Truthereum has the decisions, and the voting pools, from truthcoin. Instead of prediction markets, truthereum has a Turing-complete language to create contracts. If we used the language to define prediction markets, then truthereum would have all the same abilities as truthcoin.

The only difference between bit-ethereum and truthereum is that bit-thereum's oracles sometimes have an incentive to collude and commit a 51% attack, whereas truthereum's voters never have an incentive to cheat.

If a contract's outcome can be determined with software, then ethereum is far superior. It would have lower fees. It is less expensive to pay computers than people. For example, if two people decided to bet on whether the hash of the next block will be odd or even. Such a contract can be resolved with software. We shouldn't need to pay humans.

If a contract's outcome can only be determined by humans and not software, then truthereum or bit-ethereum are needed. For example, if two people bet on whether or not hail will destroy crops on a certain farm over a certain period of time. Software cannot detect weather.
A contract like this requires a specialized pool of voters who already pay attention to the weather in a certain area.

Truthcoin's prediction markets is the first contract type supported by truthereum.
The special prediction market where you can only buy shares, and you cannot sell... This is the second contract type supported by truthereum. It is optimized for funding public goods.
In the bribery thread, I outline a third contract type that truthereum could support which is optimized for bribery.
Probably bounty-markets would be created too, for rewarding assassinations.

If truthcoin existed, and we wanted to add the bribery thing after the fact, it would be really difficult.
We have to convince everyone to update their software at the same time.
If we were using truthereum instead, then anyone would be able to create new contracts at any time, and the contract would instantly be available to everyone.

I am struggling really hard to make the simplest version of truthcoin possible, and will not consider adding a scripting language of any sort, let alone a Turing-complete one.

[martinBrown]

This idea is on top of Truthcoin. Ie, "Ethereum by using Truthcoin".

So you're saying 'what would be the benefits of this, compared to "Truthcoin by using Ethereum" '. The answer is that, with this, you don't need Ethereum at all.

Let me see if i understand correctly. The idea is something akin to Bit-thereum, which is having bitcoin transactions escrowed by m-of-n independent oracles, each with one key of the multi-sig set. As I understand the OP, rather than m-of-n independent oracles, a TruthCoin implementation is used. And the outcome of Truthcoin decisions is used to provide the "data outside the blockchain" which ultimately arbitrates the outcome of bitcoin transactions.

I would call that idea "Bit-Truth" (analogous to Bit-thereum, which would be better named Bit-Oracles). I like the name Tru-thereum for an implementation of TruthCoin on Ethereum (i.e. truthcoin as an ethereum contract(s) / ether sub-currency).

Note that Ethereum is not specifically about mediating "data outside the blockchain", its just that the powerful scripting language (and storing of arbitrary state) makes it easy to write contracts which could do that (i.e. a SchellingCoin contract or an m-of-n Oracle-consensus contract). What Gavin Andresen basically said is let's just use independent semi-trusted off-chain blackboxes (as a workaround to the limitations of bitcoin script) for the storing the arbitrary state and executing the contract, and then give them keys to an m-of-n multisig bitcoin address.

[psztorc]

Right, Ethereum would be a fully trustless version of Truthereum, which would only aggregate semi-trustless oracles in a theoretically trustless way.

There would be no need to build Truthereum on top of Truthcoin-Ethereum, because you could just use Ethereum.