Quote from: NewLiberty on November 15, 2014, 02:35:29 PM

Quote from: Odalv on November 15, 2014, 02:09:52 PM

Quote from: NewLiberty on November 15, 2014, 01:23:25 PM

Quote from: Odalv on November 15, 2014, 12:54:57 PM

Quote from: NewLiberty on November 15, 2014, 02:15:39 AM

Sure they CAN be combined with multi-sig and oracles. But if this is not ALWAYS done (and it appears that it is likely to in fact be rare) then each SPV proof that does not so combine them, increases the incentive for a 51% attack (with miners not running OP_SIDECHAINPROOFVERIFY support, due to the ability to unlock SPV locked coins, and perhaps make some use of them).

If you have big enough user base then even 100% hash power attack cannot spend this BTC b/c it is not valid transaction.

1. you can ignore OP_SIDECHAINPROOFVERIFY
2. if some miner will add transaction what will spent BTC from address using OP_SIDECHAINPROOFVERIFY into block without valid "SideChain proof" then this block will be ignored b/c it will contain invalid tx.

Thanks, I've been intrigued on the details of that since the San Jose 2013BF conference.
This was something that Eli Ben-Sasson worked out
https://www.youtube.com/watch?v=YRcPReUpkcU

The TX are invalid for nodes running OP_SIDECHAINPROOFVERIFY, but if not...
This is an implementation of this SCIP (Succinct Computational Integrity and Privacy) so it has to be checked, yes?

It struck me as the opening of the window for deterministic computing.
Essentially verifiability of what all computers are doing all the time.
It sets up a powerful framework. Terrifying, important, and exciting.
I get why the core devs want to work on it. New math changes things.

Yes, very interesting things bitcoin can do :-).
Did you read comments ?
- Can proof of computation be used to create forced work that is useful? e.g. Protein folding instead of hashing?
- I think there does exist a possibility where the work can be replaced by running another algorithm such as Protein folding and proving through SCIP that a certain number of computational steps were executed. The other nodes then verify this proof, and accept it as work.

Useful proof of work also changes incentives
The thing of it is... SCIP is way bigger than Bitcoin.
I don't see why folks don't get the trepidation around plugging Bitcoin into this, as the first and biggest thing it is applied to.

Eli describes a process that happens all the time with me.
You hear a new notion, the first response is "Oh, that's easy" & "This is great"
Then after some thought, "Its impossible"
Then after a bunch of really hard work, you might be able to make it possible.

With respect to the complexity implications of this for Bitcoin (and the huge world beyond), I'm in that third phase for quite a while now.

The way something works is one thing to understand, it is another to understand how something breaks.

Computational ordering and process linking and all sorts of interesting hash collision potentials that could make for some phenomenal weirdness were they to come about. Sometimes REALLY unlikely things do happen.

Bitcoin is just an experiment right? We don't really think it is going to change everything, do we? So its OK to just try it and see what happens....right?

Wink

scipSC can be started today using
a) oracle
b) Federated Peg: oracle & scip-proof

=> "oracle & scip-proof" can be changed into
c) scip-proof when 70% of nodes adopt OP_SIDECHAINPROOFVERIFY