On 5/3/22 01:59, Christian Huitema wrote:
There are many technical problems, but there also some pretty fundamental User Interaction issues. The way I think of this problem is "I want to find the electronic address of the person whom I call Alice Example". That kind of name is not unique in general, but it is unique enough for me -- cryptographers often refer to this as a "pet name". Doing that in a centralized service is hard. You have to assume that a variety of phishers are going to try insert their own set of metadata in the service database. I might have better chances asking my friends, who may well understand who I refer to as "Alice Example". And maybe we could develop some kind of friend-to-friend service overlaid over a social network. But if we are not careful, it will be easy to leave enough holes to let a whole raft of phishermen through...
Sure, but even without the problem of key discovery, if you're going to send mail to Alice Example, you have to find her email address anyway, which is globally-unique. Somehow we have managed to do that for decades without a global directory. So I don't think it's necessary to solve the directory problem to solve the problem of key discovery.
And if you have an email address, you can use a DNS lookup (maybe DoH) on the domain name portion of that address to find an oracle that's associated with that domain name. And you can use some protocol that uses TLS and X.509 certs to verify that you're talking to the authoritative server for that domain name.
So the key lookup service doesn't have to be centralized, which simplifies things a bit.
But I certainly agree that there are some pretty fundamental user interaction issues to be sorted out, even without tackling the directory problem.
Keith
