Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!uwm.edu!bionet!ig!ames!amelia!orville.nas.nasa.gov!uselton
From: uselton@orville.nas.nasa.gov (Samuel P. Uselton)
Newsgroups: comp.graphics
Subject: Re: Tesselating the sphere
Summary: A related paper was presented at SPIE/SPSE Electronic Imaging
Keywords: mapping, quadtrees
Message-ID: <4966@amelia.nas.nasa.gov>
Date: 20 Feb 90 19:36:04 GMT
References: <155@tacitus.tfic.bc.ca>
Sender: news@amelia.nas.nasa.gov
Reply-To: uselton@orville.nas.nasa.gov (Samuel P. Uselton)
Followup-To: comp.graphics
Organization: NASA Ames Research Center, Moffett Field, CA
Lines: 14

On Friday, Feb 16, Gyorgy Fekete of NASA Goddard, presented a paper
at the SPIE/SPSE Electronic Imaging Workshop in Santa Clara.  Unfortunately,
the proceedings won't appear for a couple of months, so all I have is the
abstract.  He described a system in use at Goddard which started with
an icosahedral initial decomposition, followed by a subdivision similar
to the one in the referenced article: if more detail is needed, a triangle 
is divided into four, in the obvious way (split the edges at midpoints and
connect).  The title of the paper is "Sphere Quadtrees: A new Data Structure
to Support the Visualization of Spherically Distributed Data".
It is from the session "Extracting Meaning From Complex Data" (Vol 1259
for ordering purposes).  He used a quadtree (obviously :-) ) and I don't
remember a discussion of bit counts, but it could clearly be done that way.
He probably has references in the paper and might send a preprint if you
are in a hurry.