A good long while ago, I'd meta-proposed that most of the features &
functions you'd want with interactivity could be accomplished with
a very simple add/delete edit protocol. That is, to generate the
new scene, browsers would be sent a message saying "delete this, add
that". This is kind of an anti-thesis of "messaging between objects",
where objects, where-ever they are, whatever they are, "tell" other
objects to do things (that the other objects know how to do).
Maybe its time to open a (philosphical) debate on the various
ways of acheiving interactivity? Sort of like the earlier
proceedural-vs.-declarative langugage debate?
--linas