| |
Published by the IEEE Computer Society
| Article Contents | ||
| REFERENCES | ||
| Download Citation | ||
| Download Content | ||
| PDFs Require Adobe Acrobat | ||
Fig. 1. Persistent structure. (a) The node-copying technique, where each persistent node has three extra fields. (b) Persistent binary search tree (where we use the alphabetical order to compare the keys (letters)) after simulating a sequence of updates. Each node has one extra field. The number associated with a node/pointer denotes its version stamp. The numbers 1 to 7 on the top horizontal line denote the entry-point array
. Version 5 is shown in red.
Fig. 2. (a) The time-octree. (b) An example of the time tree for time interval
. Each internal node labeled
covers the time span
, and each leaf labeled
corresponds to time step
.
Fig. 3. Line-sweep process to insert/delete intervals
to the time-octree. There are four time cells
whose
intervals are respectively
. The vertical red line is the sweep line. The
interval endpoints subdivide the scalar-value range
into ranges
,
,
,
,
,
,
,
,
, where version
of the time-octree corresponds to range
. Isovalue
lies in range
, and version 5 of the time-octree contains exactly time cells
, which are active for
;
and
belong to the octree of
and
belongs to the octree of
.
Fig. 4. (a) Standard representation. Inserting/deleting a leaf
can create/remove many nodes (shown in red excluding
). (b) The compact representation. (c) With the compact representation, inserting/deleting a leaf
(as the event in (a)) can create/remove at most one internal node, the degree-2 fork node
.
• The authors are with the Computer Science and Engineering Department, Polytechnic Institute of New York University, Brooklyn, NY 11201.
E-mail: cwang05@students.poly.edu, yjc@poly.edu.
For information on obtaining reprints of this article, please send e-mail to: tvcg@computer.org.
REFERENCES
| | x | ||
