I tried to explain that the laws are not what Bob Blakley calls "desiderata" - things that we would like to see. They are the objective characteristics of an enduring identity system at Internet scale.
Oh, but how often we turn out to be wrong about these seemingly objective characteristics. Kim's Law 5 on Pluralism of Operators and Technologies understands that there will be more than just one identity authority, but wishfully states that there must be one "encapsulating protocol (a way of agreeing on and transporting things)". This reminds me of a heated BOF three years ago at the IETF for a WG I ended up being heavily involved in, chartered to take the Jabber protocol and standardize it as XMPP. At the time there was already another IETF WG doing instant messaging, and a friend of mine got up at the microphone, objecting to the formation of the WG, clearly quite upset, saying:
But you can't have two instant messaging protocols! That's like -- that's -- that's like having two IPs!
Of course the irony is that the IPv6 WG was probably right down the hall working out how to co-exist with IPv4 during an extended, possibly interminable transition period. And today we have both XMPP and SIMPLE and gateways between instant messaging protocols -- not just two, but probably more like eight (including MSN, AOL and other pre-existing systems). Much as we'd like computer systems to be simple, considerations of backward compatibility and competition between aesthetic models, among other things, keep things exciting.
1 comment:
If the comments you're referring to are the ones I'm thinking of, the comparison was not to having two versions of IP, as much as having two competing and fundamentally incompatible proposals for IPv6.
Post a Comment