Hi Alan, I am using an approach similar to your proposal in a service I am developing. I have, however, chosen to sign using the following scheme: 1. take sha512 of document (=hash512) 2. take ripemd160 of hash512 3. create 512 bit data structure, where the first 352bits are '0', and the rest is the ripemd160 of our hash512 4. sign it with the key
This procedure prevents an evil site from fooling you to sign a transaction spending your own coins. So bottom like never sign a full sha512 with a key for any other purpose than a transaction. (The above could easily well have been implemented as just truncating the hash512 to 256 bits, feel free to propose the optimal scheme). /M On 04/04/2012, at 08:23, Wladimir wrote: > Alan, > > On Wed, Apr 4, 2012 at 2:01 AM, Alan Reiner <etothe...@gmail.com> wrote: > There is all this fanfare around P2SH and how multi-sig is the solution to > all these security problems, but how the hell do you use it? I believe that > BIP 10 (or successor) is critical to the success of multi-sig, because the > greatest barrier to using multi-sig will be the ability to actually execute > them in less than 14 steps. And if every client implements it differently, > there's even less chance it will be used (assuming the userbase reaches any > level of client diversity). > > That is a laudable goal. > > So your proposal is about signing "Preformatted messages from sites" to make > financial transactions more secure, not arbitrary user-to-user messages such > as email. That really restricts the scope, which is good. > > In this case there is no use for S/MIME, which deals with encoding/signing > multipart mail messages. And no need to deal with MIME headers, html, or > embedded images, and such. And we can simply require one character encoding, > no need to support hundreds. > > The "request signing" bitcoin URL makes sense in my eyes. Less copy/pasting > is good. Do mind that there is usually a URL size limit (depending on the > browser) so this cannot be used for long messages/contracts. A possible > solution would be to make an option to pass the address where the message can > be retrieved (and maybe also where the signature must be sent, to save a > copy-paste back?). > > Looking at existing solutions, the only other "sign request" that I know of > is the CSR (https://en.wikipedia.org/wiki/Certificate_signing_request) but > the functionality and goal is very different. > > It'd be useful (and IMO most important) to write down some use-cases in which > this makes P2SH easier and less involved. How many steps can be eliminated of > the 14? > > Wladimir > BTW: we also still need a BIP to define URL signing / authentication itself. > > ------------------------------------------------------------------------------ > Better than sec? Nothing is better than sec when it comes to > monitoring Big Data applications. Try Boundary one-second > resolution app monitoring today. Free. > http://p.sf.net/sfu/Boundary-dev2dev_______________________________________________ > Bitcoin-development mailing list > Bitcoin-development@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/bitcoin-development ------------------------------------------------------------------------------ Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev _______________________________________________ Bitcoin-development mailing list Bitcoin-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bitcoin-development
