In short, codius requires parties to select a group of reputable nodes to run the contract code but in exchange is technically simpler and faster.
Ethereum requires zero trust and gives pretty much a 100% guarantee the contract will be executed as written, but is technically more challenging and contracts run slower.
Actually Ethereum requires you to trust the mining pools that make up 51% of the hashing power of the network.
Codius allows you to select oracles based on whatever criteria you like -- including using mining. The nodes don't necessarily need to be reputable, you just need them not to collude
The "trustlessness" in Bitcoin & Ethereum are conditional on the lack of collusion between miners.
It's particularly jarring to hear people clamor that it's "trustless", confusing the computational security of crypto with the much weaker game-theoretic security of Bitcoin.
Marketing I suppose.
As far as we know, blockchain systems are the best way so far to minimize the amount of trust. It's all well and good to talk about how imperfect blockchains are, but until something better comes along it seems like not much can be gained by parsing out different gradiations of "trust" in the way you suggest.
Codius (and Orisi.org) can deliver external inputs to the contracts - it's something no blockchain system can do by design.
In other words - if you want your Ethereum contract to rely on weather, a certain website, or a currency price, you need Orisi, Codius, or some other similar system.
> Codius (and Orisi.org) can deliver external inputs to the contracts - it's something no blockchain system can do by design.
False, when I make a bitcoin purchase, and external input (the amount I spent) makes it into a bitcoin script. In other altcoin systems you can put in arbitrary datafeeds, like the weather (admittedly with a gatekeeper that spoon feeds it to the blockchain) Also, if someone wanted to create an altcoin that allowed totally arbitrary inputs (like Codius) I see no reason why it would be technically impossible.
I see that Codius requires a way to find non-colluding nodes to process your contracts. Mining is terribly inefficient and slow (can be fast with the GHOST or DECOR protocol though).
Looks like Codius might fit perfectly with the Bitshares Delegate system. The top 101 delegates could become the trusted list of Codius code processors. All Delegates could check the output hash of the contract of every other delegate and kick out the malfunctioning or bad delegate out of Bitshares altogether, just as they already do for transaction confirmation issues.
On top of that the Codius processors could be paid instantly with Opentransactions tokens which represent Bitshares based bitUSD/EUR/RMB, etc.
I highly doubt Ethereum could do better than this. This setup would be fast, secure, super profitable for delegates, and it would not suffer transactions per second limitations as much.
Ethereum requires zero trust and gives pretty much a 100% guarantee the contract will be executed as written, but is technically more challenging and contracts run slower.