Quote from: Mike Hearn on April 07, 2014, 03:40:22 PM

For bitcoinj, the way I think we'll go for now is just standardising some local API bits. It's too early to come up with a client->server protocol standard: we need to wait for things to settle down and best practices to be established. Current bitcoinj plan:

https://groups.google.com/forum/#!topic/bitcoinj/Uxl-z40OLuQ

Hi Mike,

Thanks for the link to those design notes. I see your point and agree that it makes more sense to provide a plugin interface for oracles and leave the wire protocols to be bespoke until the whole space matures.

One thing that I didn't understand from your document was this line:

Quote

In an ideal world, BIP 70 would be more widely adopted by now and P2SH would not be required or useful.

How does BIP 70 prevent the need for multisig addresses? My understanding was that BIP 70 operates at a higher level on top of the blockchain, by simplifying the workflows involving payments and address management. But we would still need P2SH multisig scripts in order to provide blockchain-level security involving risk analysis services / oracles.

I am interested in building a RA service as you call it, so would be keen to know when bitcoinj has this implemented. Even if you had just a basic sketch of the plugin API so that I know what type of calls would need to be implemented, that would be great for me to get started on trying to build something.

Thanks.