On the specifics, I'm a bit more skeptical. Although your formalism does
adhere to your position on <base>, that's only because you've imposed the
assumption (unnatural in my opinion) that the "addr" map be surjective.
Surjectivity assumptions in a distributed system are often a sign of trouble.
More generally, the careful distinction between the address/resource/entity
aspects of a "Web document" that you developed for Expires has not yet made
it over to the conceptual foundation of "Web links". Do links exist at the
address, resource, or entity level, so to speak? Hopefully all of them.
For interactivity and statefulness, in particular, it will be important
to have linking concepts at the entity level.
I'd like to take up the challenge of working with you toward a better
model, because I see it as an essential precondition to any meaningful
Internet Draft on Web interactivity.
---------------------------------------------------------------------
Paul Burchard <burchard@math.utah.edu>
``I'm still learning how to count backwards from infinity...''
---------------------------------------------------------------------