> Description of loading less-detailed scene first...
...
> If the server is intelligent, it could do some of this automatically--
> it could at least deliver up a bounding-box representation of the scene
> first. However, once of the nice properties of HTTP is that the server
> doesn't care what kind of files it is serving up to the HTTP clients.
> If all you care about is rendering speed (and not the length of time to
> transfer scenes across the net), then browsers can do this
> automatically (Inventor's viewers already support the notion of "render
> as bounding boxes", for example).
I would like to see VRML as allowing authors a way of specifying objects
at a number of levels of detail, e.g. close up a tree has a trunk and
branches formed from cylinders together with a large number of polygonal
leaves. Further away it is represented as a simple 2D image, and in the
far distance as a color + cross sectional area. Low end systems could
choose to use one of the less detailed models.
Introducing explicit levels of detail also gives you the chance to
attach behviours which are only triggered when you come within a certain
range, e.g. the clicking of a clock, or retrieving extra info as a kind
of networked annotation (personal possessions in a work cubicle, ...).
-- Best wishes,Dave Raggett
----------------------------------------------------------------------------- Hewlett Packard Laboratories email: dsr@hplb.hpl.hp.com Filton Road tel: +44 272 228046 Stoke Gifford fax: +44 272 228003 Bristol BS12 6QZ United Kingdom