This first meeting will attempt to figure out how E should cope with the
ways actual IP addresses, and the relevant platform networking stacks, fall
horribly short of the idealized notion of IP addressing. Many of these
problems have already been addressed in E (primarily by Eric Messick & Bill
Frantz (others?)) but many problems remain. The key issue is, of course,
how to bring about three-party handoff
http://www.erights.org/doc/elang/concurrency/images/fundamental.gif despite
these problems. And to determine which problems we can postpone fixing
till after 1.0.x.
In a simple idealization of IP, each machine has one permanent IP address which its running applications can find out. Anyone with IP connectivity to the internet can contact that machine merely by using that same IP address. A wonderful fantasy. As we'll see in the following series of notes, many of our problems trying to create true distributed capability security are caused by misguided approaches to try to make the IP world more "secure".
