Two score years ago, great minds looked at the confusion of networks and networking protocols and realized the need for a network of networks. The inter-network protocol was a protocol that went between the gaps in the networks and connected them together. Once this Internetwork protocol was established between the networks, people realized they could run the same protocol on the LAN and it would provide the same capabilities as the network protocol but with Internet capabilities.
Today we are facing a real problem in the messaging space. I have accounts on Skype, Signal, Keybase, Facebook, LinkedIn, G+ and I only cover half the set of things people ask me for.
A major part of the reason we got into this situation is that 20 years ago, AIM was dominant in the messaging space and would sue the pants off anyone who tried to interoperate.
We have fragmented protocols that all do the same thing. Jabber and SMTP do exactly the same thing: they post messages. Only some minor differences in the client UI make them seem different.
There is an old joke about 20 standards becoming 21. But that is because creating a new protocol is the wrong approach.
We need to have an Internet for applications. That is some sort of infrastructure that can bind all the existing systems together into one so that if I want to talk to Fred and send a message, the infrastructure will work out which common protocols we support and sets up the call. Or rather tries to because Fred might not want to talk to me right now or by that method or ever.
Such a system will of course inevitably converge on one communication protocol. But that puts the cart before the horse. Before the pioneers could build IP Networks, they focused on connecting the existing networks.
The requirements for such a system are not difficult, it isn't much more than a contacts directory that the user controls and can be ported from one device to another without regard for vendors. Oh and it should have PKI client credential support built in.
My employer is no longer in the business of selling certificates. We are not a CA. Instead of making money selling credentials that let other people know who our customers are, how about we try to curate Webs of trust and tell people what a good certificate to talk to Fred is?
I am about 70% of the way to complete building one. Anyone interested in trying it out?
