> I haven't written a signer, but deal a lot with signing at deSEC. > > It's been said a few months back that its difficult to avoid > collisions in a multi-signer setup, because you don't know ahead > of time what key the other signer uses. > > Avoiding them is certainly possible, but requires extra coordination. > It seems to me that the complexity of that is (much) larger than > continuing to allow ~one collision.
In my opinion this is a quality of implementation issue. We should not design a multi-signer protocol that has collision even if there is no document that requires it. You are right that it requires extra effort. But it has also benefits. For example testing software if all code paths properly handle keys with key tag collisions is also unpleasant. THere is also the issue that software that stores public keys in files with the traditional filename scheme may have a hard time handling key tag collisions. _______________________________________________ DNSOP mailing list -- [email protected] To unsubscribe send an email to [email protected]
