On Sun, Dec 31, 2017 at 6:18 AM, Jonathan Frederickson silverskullpsu@gmail.com wrote:
whereas with blockchain and hashgraph that trust is a mathematical inviolate (cryptographical level of) certainty.
Sure, assuming every aspect of that contract can be expressed on the blockchain. Financial transactions with cryptocurrencies can be (I'll sell you this ERC20 token for this much ETH, if and only if enough people pledge), but I'm still skeptical that such contracts will be very useful outside of a few narrowly defined niches.
it's early days yet. cryptokitties is an actual first real-world effort to create something that isn't "trading or mining or speculation". this video is both extremely funny, informative and insightful at the same time:
https://www.youtube.com/watch?v=hu8X0aZhDpA
A smart contract for preordering a new device, for example, has no idea whether the devices have been shipped, whether they were lost in transit, etc. The moment you involve humans (which is almost always necessary), you're back to needing dispute arbitration. A mathematically inviolable contract does you little good when a key part of the contract as seen by one of the parties (i.e. actually receiving the thing) can't be expressed as part of it.
yehyeh. but... let's think it through. there's a couple of scenarios
(1) one of the uses is: insurance. insurance companies are now investigating blockchain for declaring and underwriting insurance. what do you think would happen if an *insurance* company decided, affer making a publicly declared inviolate contract with someone, "oh we don't want to actually pay out, it's just a mathematical note, we didn't really actually truly mean it"?
the consequences of disregarding a [physical] violation / dishonouring of a publicly-notarised atomic contract is actually much more serious than it first seems.
(2) numerically-expressible contracts such as "subtract this number and add it to this other number"
this is basically the whole basis of crypto-currencies (except ethereum). the transaction is done and inviolate and there *is* no backing out. the "number" is transferred (added) to your "balance sheet", and that "balance sheet" propagates forward within the network.
the number [the "value" of the crypto-currency] *has forward continuity*. further atomic transactions empower further atomic transactions empower further atomic transactions.... you never *actually* escape from the inviolate nature of the atomic-contract-public-ledger-system.
thus, for *this* scenario, the objection that "human dispute arbitration" is needed simply does not apply.
(3) computationally-expressible contracts such as ethereum
this is where it gets *really* interesting as e.g. cryptokitties demonstrates (which, btw, has caused a major meltdown of ethereum as it's sucking up THIRTY PERCENT of the ENTIRE ethereum network resources).
whilst you _can_ actually replicate the contract system (cryptokitties unfortunately have not published the FULL genetic algorithm / contract) it is clear that the cryptokitties site is - as things stand - entirely set up to be the "enabler". it is NOT however the "arbitrator". the ETHEREUM NETWORK is the "arbitrator and prover that you bought, sold or birthed a kittie".
it is just unfortunate that the relevant algorithms have not been made public so it is not possible to create your own kitties... you *can* i believe however independently *buy and sell* them using a replica of the buy and sell algorithms... i'd like to see someone do that.
regardless: this is just another specialisation and extension of the extreme degenerate case, "subtract this number and add it to this other number", such that *as long as* you remain *within* the cryptokitties "system", All Is Well.
(4) static contract / document storage including describing arbitration procedures (potentially in computational form)
this is a proposed "hybrid" system which would be REALLY interesting, and is a general form of (1) and (3) that i've been thinking about for some time.
so where there is an actual cross-over into the real world, it's the ethereum-style (computation style) contracts that get particularly interesting, and even more so when you have machine-readable documents (think "Microsoft Access Forms"). in effect, this moves our dependence on individual computers over to DISTRIBUTED networks, exactly as the microsoft millenium research group envisaged about 20 years ago and used DCOM as the basis to do it.
so overall i believe you're missing the fact that it's only because *right now* we still use "individual" computers (or MIS-place our trust in other PEOPLE's computers), but imagine what would happen if ethereum's computation system BECAME THE DEFAULT FOR ALL COMPUTING ALGORITHM EXECUTION WORLD-WIDE OVERNIGHT.
that we *haven't* moved to that model - world-wide - is the reason why i believe that what you say, "we need to fall back on human-powered dispute arbitration" is true... *at the moment*.
l.