Difference between revisions of "NURBS TODO"

From BRL-CAD
m (Reverted edits by Caddui (Talk); changed back to last version by Sean)
(several items completed including V&V and prep caching)
 
Line 1: Line 1:
= Optimize BREP/NURBS ray tracing=
+
= Profile BREP/NURBS ray tracing=
* Benefit: faster ray tracing runtime, user visible
+
* Benefit: better ray tracing runtime understanding, supports test case development
 
* Time Estimate: 2 days
 
* Time Estimate: 2 days
 
* References:
 
* References:
 +
** src/libbrep
 
** src/librt/primitives/brep
 
** src/librt/primitives/brep
 
** src/librt/opennurbs_ext.*
 
** src/librt/opennurbs_ext.*
  
 
= Optimize BREP/NURBS prep=
 
= Optimize BREP/NURBS prep=
* Benefit: faster ray tracing runtime, user requested, user visible
+
Perform limited tree splitting during prep, split dynamically during ray trace, store lock-free.
* Time Estimate: 2 days
+
* Benefit: faster ray tracing throughput, user requested, user visible
 +
* Time Estimate: 1 week
 
* References:
 
* References:
 +
** src/libbrep
 
** src/librt/primitives/brep/brep.cpp:rt_brep_prep()
 
** src/librt/primitives/brep/brep.cpp:rt_brep_prep()
 
** src/librt/opennurbs_ext.*
 
** src/librt/opennurbs_ext.*
  
=Implicit-to-NURBS=
+
= Convert existing entities to NURBS=
Implement conversion of implicit primitives to BREP/NURBS primitives. Partially complete, primitives are fixed at origin, some primitives unimplemented.
+
Implement hooks to describe entities in BREP/NURBS form. Partially complete for most, but some complex entities remain unimplemented.
 
* Benefit: partial requirement for interactive visualization of implicit CSG geometry in OpenGL
 
* Benefit: partial requirement for interactive visualization of implicit CSG geometry in OpenGL
* Time Estimate: 3 days
+
* Time Estimate: 2 days per entity
 
* References:
 
* References:
 
** src/librt/primitives/${PRIM}/${PRIM}_brep.cpp
 
** src/librt/primitives/${PRIM}/${PRIM}_brep.cpp
  
=BREP/NURBS-to-NMG=
+
= BREP/NURBS-to-NMG=
 
Implement conversion of BREP/NURBS to NMG.
 
Implement conversion of BREP/NURBS to NMG.
 
* Benefit: requirement for polygonal export, partial requirement for OpenGL visualization, user visible
 
* Benefit: requirement for polygonal export, partial requirement for OpenGL visualization, user visible
* Time Estimate: 1 day
+
* Time Estimate: 2 days
 
* References:
 
* References:
 
** src/librt/primitives/brep/brep.cpp:rt_brep_tess()
 
** src/librt/primitives/brep/brep.cpp:rt_brep_tess()
  
=BSPLINE/NURBS review=
+
= Old BSPLINE/NURBS review=
 
Identify useful concepts implemented in the old BSPLINE/NURBS code that are not present in openNURBS.
 
Identify useful concepts implemented in the old BSPLINE/NURBS code that are not present in openNURBS.
 
* Benefit: code reuse, potentially very useful functionality.
 
* Benefit: code reuse, potentially very useful functionality.
* Time Estimate: <1 day
+
* Time Estimate: 1 day
 
* References:
 
* References:
 
** src/librt/primitives/bspline/*
 
** src/librt/primitives/bspline/*
  
=BSPLINE/NURBS reuse=
+
= BSPLINE/NURBS reuse=
 
Port any useful BSPLINE/NURBS code to openNURBS data structures and routines.
 
Port any useful BSPLINE/NURBS code to openNURBS data structures and routines.
 
* Benefit: code reduction, allows the old BSPLINE/NURBS code to be removed after functionality is migrated.
 
* Benefit: code reduction, allows the old BSPLINE/NURBS code to be removed after functionality is migrated.
Line 45: Line 48:
 
Implement boolean evaluation of BREP/NURBS surfaces.
 
Implement boolean evaluation of BREP/NURBS surfaces.
 
* Benefit: partial requirement for interactive visualization of BREP/NURBS in OpenGL.
 
* Benefit: partial requirement for interactive visualization of BREP/NURBS in OpenGL.
* Time Estimate: 7 days
+
* Time Estimate: 2 years (for 100% robustness)
 
* References:
 
* References:
 
** ESOLID: http://research.cs.tamu.edu/keyser/geom/esolid/
 
** ESOLID: http://research.cs.tamu.edu/keyser/geom/esolid/
Line 56: Line 59:
 
** ftp://ftp.cs.unc.edu/pub/users/manocha/PAPERS/INTERSECT/IJCGA.pdf
 
** ftp://ftp.cs.unc.edu/pub/users/manocha/PAPERS/INTERSECT/IJCGA.pdf
 
** http://portal.acm.org/citation.cfm?id=237751
 
** http://portal.acm.org/citation.cfm?id=237751
 
=Plate-mode BREP/NURBS=
 
Implement plate mode BREP/NURBS objects. Similar, if not identical, to plate mode BoTs, 'brep' objects need the ability to have a thickness described so you could automatically handle non-solid NURBS and sheet-metal style geometry without requiring surface solidity.
 
* Benefit: allows for import and visualization of BREP/NURBS surfaces that are not solid, user visible
 
* Time Estimate: 2 days
 
* References:
 
** include/rtgeom.h
 
** src/librt/primitives/bot/bot.c
 
  
 
=Implicit CSG to BREP/NURBS CSG test scripts=
 
=Implicit CSG to BREP/NURBS CSG test scripts=
Line 79: Line 74:
 
* References:
 
* References:
 
** doc/docbook/articles
 
** doc/docbook/articles
 
=Cache objects=
 
Implement data cache objects.  Create new LIBRT API routines for storing cache data in a v5 container.  This would amount to storing binary data into a 'binunif' object marked as hidden with attributes for date/time stamping and MD5 or SHA256 checksumming.
 
* Benefit: faster ray tracing runtime, applicable to other primitives (NMG, BoT)
 
* Time Estimate: 1 day
 
* References:
 
** include/raytrace.h
 
** include/db5.h
 
** src/librt/binunif
 
** src/librt/primitives/*
 
 
=Raytrace prep cache optimization=
 
Cache BREP/NURBS prep results to a specified cache file.
 
* Benefit: faster ray tracing runtime for BREP/NURBS, user requested, user visible
 
* Time Estimate: 2 days
 
* References:
 
** src/librt/primitives/table.c
 
  
 
=BREP/NURBS wireframe=
 
=BREP/NURBS wireframe=
Line 123: Line 101:
 
* References:
 
* References:
 
** src/librt/primitives/brep
 
** src/librt/primitives/brep
 
= Report on using BREP/NURBS for analysis=
 
Description needed here.
 
* Benefit: documentation, user visible
 
* Time Estimate: 3 days
 
* References:
 
** doc/docbook/articles
 
 
= Analytic comparison of BREP/NURBS and BoT=
 
Description needed here.
 
* Benefit:
 
* Time Estimate: 1 day
 
* References:
 
** sh/*.sh
 
** regress/*.sh
 
** bench/run.sh
 
 
= Analytic comparison of BREP/NURBS CSG and implicit CSG=
 
Description needed here.
 
* Benefit:
 
* Time Estimate: 2 days
 
* References:
 
** sh/*.sh
 
** regress/*.sh
 
** bench/run.sh
 

Latest revision as of 17:55, 26 August 2016

Profile BREP/NURBS ray tracing[edit]

  • Benefit: better ray tracing runtime understanding, supports test case development
  • Time Estimate: 2 days
  • References:
    • src/libbrep
    • src/librt/primitives/brep
    • src/librt/opennurbs_ext.*

Optimize BREP/NURBS prep[edit]

Perform limited tree splitting during prep, split dynamically during ray trace, store lock-free.

  • Benefit: faster ray tracing throughput, user requested, user visible
  • Time Estimate: 1 week
  • References:
    • src/libbrep
    • src/librt/primitives/brep/brep.cpp:rt_brep_prep()
    • src/librt/opennurbs_ext.*

Convert existing entities to NURBS[edit]

Implement hooks to describe entities in BREP/NURBS form. Partially complete for most, but some complex entities remain unimplemented.

  • Benefit: partial requirement for interactive visualization of implicit CSG geometry in OpenGL
  • Time Estimate: 2 days per entity
  • References:
    • src/librt/primitives/${PRIM}/${PRIM}_brep.cpp

BREP/NURBS-to-NMG[edit]

Implement conversion of BREP/NURBS to NMG.

  • Benefit: requirement for polygonal export, partial requirement for OpenGL visualization, user visible
  • Time Estimate: 2 days
  • References:
    • src/librt/primitives/brep/brep.cpp:rt_brep_tess()

Old BSPLINE/NURBS review[edit]

Identify useful concepts implemented in the old BSPLINE/NURBS code that are not present in openNURBS.

  • Benefit: code reuse, potentially very useful functionality.
  • Time Estimate: 1 day
  • References:
    • src/librt/primitives/bspline/*

BSPLINE/NURBS reuse[edit]

Port any useful BSPLINE/NURBS code to openNURBS data structures and routines.

  • Benefit: code reduction, allows the old BSPLINE/NURBS code to be removed after functionality is migrated.
  • Time Estimate: 2 days
  • References:
    • src/librt/primitives/bspline/*
    • src/librt/primitives/brep/*

BREP/NURBS CSG evaluation[edit]

Implement boolean evaluation of BREP/NURBS surfaces.

Implicit CSG to BREP/NURBS CSG test scripts[edit]

Develop test scripts for comparing conversion of implicit CSG to BREP/NURBS CSG

  • Benefit: provides validation and verification via ray tracing.
  • Time Estimate: < 1 day
  • References:
    • regress/*.sh
    • sh/*.sh

BREP/NURBS documentation[edit]

Document the new BREP/NURBS primitive (user overviewdocumentation: how to import, use, export, limitations).

  • Benefit: provides user documentation explaining capabilities, limitations, and benefits; user visible
  • Time Estimate: 3 days
  • References:
    • doc/docbook/articles

BREP/NURBS wireframe[edit]

Improve BREP/NURBS wireframe. Investigate using knots on subcurves.

BREP/NURBS source code cleanup[edit]

Consolidate Owens and Reeves approaches to ray intersection.

BREP/NURBS healing[edit]

Implement an rt_heal(rt_db_internal *dbi); routine. The routine tightens up trimming curves and edge/vertex pairings so that they align perfectly while still preserving the topological structure.

  • Benefit: more robust evaluation for improved (more stable & consistent) raytracing, import, export; routine interface later applicable to NMG/BoT too
  • Time Estimate: 3 days
  • References:
    • src/librt/primitives/brep