Merge pull request #867 from Idclip/tsc_notes

TSC Notes

Author: Idclip

tsc/meetings/2020-11-03.md

@@ -0,0 +1,50 @@
+Minutes from 69th OpenVDB TSC meeting, Nov 3rd, 2020, (EDT)
+
+Attendees: *Nick* A., *Jeff* L., *Ken* M., *Dan* B.
+
+Additional Attendees: Johannes Meng (Intel), JT Nelson (Blender),
+Andre Pradhana (DW), Bruce Cherniak (Intel)
+
+Regrets: *Peter* C.
+
+Agenda:
+
+1) Confirm quorum
+2) Secretary
+3) Timezones
+4) OpenVDB Grid Layout (Cell centered vs Node centered)
+5) Next Meeting
+
+1) Quorum was confirmed.
+
+2) Secretary was Nick Avramoussis
+
+3) Timezones
+
+Agreed to move with daylight savings - i.e. whatever the meeting invite says!
+
+4) Node Value Locations (Cell centered vs Node centered)
+
+A user has run into issues where the distinction between how values may
+potentially be stored in OpenVDB numerical grids vs OpenVDB Points grids was not
+clear. The problem is essentially that methods to retrieve node bounding
+information from a given VDB do not necessary respect the positioning of cell
+values. This is especially true for CoordBBoxs whose integer coordinates can
+either represent the minimum to maximum-1 bounds (Node centered) OR the
+minimum-0.5 to maximum+0.5 bounds (Cell centered). The distinction between the
+two is left to the underlying algorithm. Comments were made in regards to the
+position information of points within a PointDataGrid. Points are stored
+relative to the cell center; that is their voxel offsets are between -0.5,+0.5
+and not 0,1. This difference is irrelevant when considering the fast discarding
+of candidate nodes; instead one must account for the differences of a volume
+modeling a continuum and a points grid with discrete data stored potentially
+throughout a given cell. Jeff observed that the ray intersection code could
+potentially be incorrect for cell centered values. All agreed that better
+documentation is a must. Nick, derived or specialized implementation of bounding
+boxes which represent either state with clearer API methods would also help show
+this distinction. Ken to investigate further and draw up images representing the
+problem for further discussion and to hopefully include in the docs.
+
+5) Next Meeting
+
+Next meeting is November 10th, 2020. 1pm-2pm EDT (GMT-4).