Vis. techniques for EB data - basic and research
Disclaimer
What follows is just "chicken scratchings" for the
LDRD Final Report.
Taxonomy
Conceptually straightforward techniques/extensions
Note: Although these are conceptually straightforward extensions,
a considerable amount of work is necessary to implement any of them in
existing vis. packages because none provide even basic support (e.g. data
structures) for this.
-
Flatten data and treat embedded boundaries (EB) separately
-
Independent collections of data on grids and treat EB separately (render
together)
-
Cut and paste and render (replace vis. of coarse data with vis. of finer
data where possible and show the result, seams and all) and treat EB separately
Other techniques/extensions
-
EB/vof's treated and visualized as 1st class citizens in AMR data rather
than extracting them as geometry and visualizing them separately (ANAG
needs storage capabilities for EB inside AMRv)
-
Visualization of result in the continuous problem domain rather than in
the discrete computation domain - collaborative effort required
-
"Grid" visualization
-
Data selection and interrogation referenced all the way back through visualization
tools to Boxlib (lacking in basic infrastructure)
Tacks
Leverage off NCSA work.