Quadtrees?

[coder at ibm.net]

On 28/02/97 at 06:58 AM, cg at ami-cg.GraySage.Edmonton.AB.CA (Chris Gray)
said:

>My thinking has been to use a 3D sparse array, which can contain objects
>of differing sizes, hence visibilities. I'm pretty vague on how the
>details would work, however. :-(

Sounds a little like an R*-Tree.  I recently reposted an old post from
Wout's list with definitions and descriptions of quadtrees etc.  For me
the beauty of an R*-Tree is that it allows very easy localised processing. 
To find all locally relevant data nodes, just ascend the tree from the
current location until you find a rectangle which bounds the requested
area.  Another nice point is that multiple "thematic" schema can be easily
bound within the same R*-Tree.

--
J C Lawrence                               Internet: claw at null.net
----------(*)                              Internet: coder at ibm.net
...Honourary Member of Clan McFud -- Teamer's Avenging Monolith...