Post
Topic
Board Development & Technical Discussion
Re: Really Really ultimate blockchain compression: CoinWitness
by
Peter Todd
on 19/08/2013, 21:01:15 UTC
Maybe the distinction between soft-fork and hard-fork is less important than I thought, but I'm still a bit confused on why this is a soft-fork. With regard to a valid transaction that passes the SCIP verification (from the off-chain system back into the Bitcoin network), can you please explain why the older nodes would consider it to be a valid transaction? If that's the case, it means that the older nodes would allow anyone to spend such outputs, without checking anything at all?

That is exactly the case; see BIP 16 for an example where a soft-fork was used in that way. CoinWitness could be implemented similarly.

Basically it isn't safe to use CoinWitness until >50% of the hashing power rejects invalid CoinWitness transactions, but unless you are a miner you don't actually need to upgrade your node.