On Thursday 21 July 2005 14:15, Werner Koch wrote:
On Thu, 21 Jul 2005 13:11:56 +0100, Ian Grigg said:
But it recalls to mind what we do in contract issuance. In
our model, we add strings to every keyId in the chain. These
"roles" then inform the software of how to prepare and check
This works well when using a new key for each role.
Assuming you would add the rules as different UID to one key you can't
see from a signature which role/UID was used to sign the document.
The Signer's User ID is a solution to this; however it is far easier
to create separate keys.
I agree entirely, that's in fact what we do, and
the software rejects any additional roles that
might be present.
There's also the issue that generally different
roles have different security models, so for
example ones [certification] key might be better
off locked in the safe.
iang
--
Advances in Financial Cryptography, Issue 2:
https://www.financialcryptography.com/mt/archives/000498.html
Mark Stiegler, An Introduction to Petname Systems
Nick Szabo, Scarce Objects
Ian Grigg, Triple Entry Accounting