On Tue, May 3, 2022 at 10:46 PM John Levine <johnl@xxxxxxxxx> wrote:
It appears that Lyndon Nerenberg (VE7TFX/VE6BBM) <lyndon@xxxxxxxxxx> said:
>For an IMAP store things get a bit tricker. It would need a protocol
>exstention that allowed for on-the-fly decryption.
IMAP does header and content searches, and invariably indexes the
message store and update the indices as messages arrive so the
searches are reasonably fast. A lot of them use SIEVE or something
like it to do mail sorting as the mail arrives. Other than storing all
your keys in your IMAP server, how do you plan to do that? It is my
impression that the vast majority of mail users leave their mail on
the server and see the same folders from multiple devices, so this is
not an edge case, it's the most common case.
These are not trivial problems. The people who write mail servers are not
completely inept, and if there were easy solutions to key management problems,
we would have solved them.
There are many things that only look simple AFTER they have been solved. The Web for instance.
The key advance in the Web was jettisoning a lot of Ted Nelson's ideology. The focus on payments and copyright and most importantly his insistence that links must never break. What made the Web possible is 404 not found.
We have hobbled ourselves with a lot of bad decisions in the PKI space. First and foremost, the P stands for Public. We only ever really focused on managing public keys and left private key management to the users to sort out.
Another major issue is that PKIX and PGP are both focused on managing credentials for people that conflate authentication and authorization. This in turn results in the need for certificate expiry and the need for frequent certificate refresh. WebPKI is a little better but it is welded to the DNS which in turn is designed for discovery of Internet hosts, not users and the entire naming system is designed for organizations, not people. This in turn cripples PGP and S/MIME because a binding of a key to 'alice@xxxxxxxxxxx' is only telling me about the example.com account 'alice', it is not a binding to Alice herself because Alice doesn't own that name.
The problem with deployment is that there are a lot of vested interests and everyone is determined that it won't be their ox that is gored. This is not my first rodeo though.
The Web did have one major technical advance that Ted did not anticipate: URIs. DECNET file names provided something similar but Dan Connolly's addition of the scheme prefix changed the game.
The key to making key management work is to make use of threshold cryptography. At the moment, getting folk to even take notice of my work is a bit difficult. But as I said, this is not my first rodeo. My demo reels show how easy key management can be if we use the right tools:
I have a very simple value proposition for IETF-ers taking notice of Threshold Cryptography:
Most employers send employees to forums like IETF so they are up to date with current technology. They are paid to be the eyes and ears of the company. As Principal Scientist, I made it my business to be sure that I understood every technical development that my CEO might call me up to ask a question about at any time without notice.
So I am going to be announcing this work at HOPE in NYC just before IETF Philadelphia. Former member of CERN Web team proposes new open infrastructure to secure the Internet is a story designed for the media. I used to get a lot of earned media when I was at VeriSign and I will soon be hiring a PR team to the Mesh Foundation. People can ignore me if they like but having your CEO find out about the next thing through the press could be a career limiting move.
The current collapse of the criminal-currency schemes might well make this a very opportune time to launch the Mesh. If people recall, when the Web launched, Time Warner and co were investing hundreds of millions of dollars in what they called 'interactive television' which was not interactive at all, not in any degree. It was the collapse of Interactive TV that led to all those dollars being poured into the Web. As pretty much everyone in this forum knows, 'Web/3' is a fraud, it is not a 'decentralized' anything, it is just another scam designed to empty the pockets of the gullible.
What I have is the real deal. Web3 sets expectations that the scammers have no intention of meeting. The Mesh delivers. I think a lot of 'cryptocurrency' companies are going to find that a pivot to the Mesh is the best way to save their investments.
