Re: [Bitcoin-development] Payment protocol thoughts

2012-10-02 Thread Gregory Maxwell
On Tue, Oct 2, 2012 at 6:44 PM, Mike Hearn wrote: > A simple way to solve this problem is just use the SSL identity of the > server that is taking part in the protocol, but it's not much harder SSL itself (as opposed to using the certs as you suggest) is not non-reputablable, so it's not enough f

Re: [Bitcoin-development] Payment protocol thoughts

2012-10-02 Thread Mike Hearn
I think it's worth pondering the different things we may want in future, even if that future is quite far out, just to ensure we have a robust design that won't box us in later. Brainstorming feature ideas now doesn't commit anyone to implementing them, but it may help improve the final v1 design.

Re: [Bitcoin-development] Payment protocol thoughts

2012-10-02 Thread Peter Vessenes
I meant sent twice, a. No double-spends that I'm aware of. Sorry for the loose verbiage! Peter On Tue, Oct 2, 2012 at 11:07 AM, Jeff Garzik wrote: > On Tue, Oct 2, 2012 at 1:52 PM, Peter Vessenes wrote: > > This is small, but an interesting tidbit from BTC Foundation payments; > > roughly 3-5

Re: [Bitcoin-development] Payment protocol thoughts

2012-10-02 Thread Jeff Garzik
On Tue, Oct 2, 2012 at 1:52 PM, Peter Vessenes wrote: > This is small, but an interesting tidbit from BTC Foundation payments; > roughly 3-5% of our initial members double-spent. WOW, that's terrible. To be specific, do you mean a) paid twice or b) sent BF coins, then sent the same coins el

Re: [Bitcoin-development] Payment protocol thoughts

2012-10-02 Thread Peter Vessenes
There are tons of scenarios, some discussed here on this list previously, which would benefit from out of band messages and notes. This is small, but an interesting tidbit from BTC Foundation payments; roughly 3-5% of our initial members double-spent. WOW, that's terrible. I presume that's becaus

Re: [Bitcoin-development] Payment protocol thoughts

2012-10-02 Thread Gavin Andresen
I agree we need a payment protocol, but instead of thinking of all of the things we might possibly want I would like to solve a few boring problems that we have right now. Absolutely critical: + Bitcoin addresses by themselves are insecure against man-in-the-middle attacks. We need a payment prot

[Bitcoin-development] Payment protocol thoughts

2012-10-02 Thread Mike Hearn
I've been thinking about the requirements for a payment protocol lately. It seems we have consensus that we need one of these. Pieter has a gist on the topic here: https://gist.github.com/1237788 IMHO we'll want to move away from "send X BTC to address Y" and more towards "upload to me transaction