Right, but that's because there's very little notion of "object" in HTML.
People just don't *think* about HTML pages that way. (At least, they
didn't used to -- there's just a glimmering of that appearing now.)
VRML, on the other hand, is intrinsically *intensely* object-oriented,
in the most literal sense.
Therefore, we can't count upon the HTML side of the world to come up
with an adequate solution. They don't have the strong motivation for
good object-base solutions. We do; therefore, we need to at least help
to drive this effort along.
BTW, this really is more than *just* a caching problem -- it's really
closely tied into the URI effort, and I'd advise anyone seriously
interested to take a look at that. Unfortunately, I don't think that
the URI team has yet come up with quite the concept we need, although
they've got a bunch of stuff that should be useful. (URNs are, at
least, a step in the right direction.) We probably will ultimately
want to use all of the major URI concepts, and maybe add one or two
of our own...
-- Justin
Who is still musing on the notion of putting
his keywords proposal into some sort of
URC format, but first we need a concept
of querying for URC qualities...
Random Quote du Jour:
"My interests are aerobics, surfing, and holistic frog-licking."
-- The 976iverse