Minutes of the team meeting 23 September 2005
Last meeting with Marek. Thanks to him for his nice work
and best wishes for his future activities.
Review of the posters
Ilka: 2 posters see http://root.cern.ch/root/posters/Allposters.ppt
Bertrand 1 poster ""
Olivier 1 poster ""
Timur 1 poster ""
Richard 1 poster see http://root.cern.ch/root/posters/richard.ppt
Gerri 1 poster
Miro Morach 1 see http://root.cern.ch/root/posters/miro_poster.ppt
Sketch of the talks
ROOT I/O: status and developments (30') Markus Frank
We will review the latest improvement to the ROOT I/O, including
new convenience interfaces. We will review in details the new
extended and optimized support for STL collections.
see Markus draft talk at: http://root.cern.ch/root/drafts/markus.ppt
Trees, TTreeSQL: status and developments
We will review the latest improvement to the ROOT TTree, including
support for indexing TChains, new convenience interfaces.
We will also review in details the new SQL interface.
CINT: status and developments (20')
We will review the latest development. In particular we
will describe Masa's new bytecode executor engine and the
plan for improving the CINT by leveraging Reflex.
General - TVirtualViewer3D & TBuffer3D
OpenGL - Logical/Physical Objects, Scene, Cameras
Rest about GL viewer only:
Clipping - GL Planes + CSG
Scene Rebuilds - on demand model
Rendering - multipass, timeouts, draw sorting
Level of detail
-PROOF demo status
-demos with graphics
October and December releases
- Mostly fixes and consolidation of the additions made in the current release
- TGeoShape::Capacity() implemented for most primitives. Needed for accurate
computation of geometry weight.
- Full support for scale transformations. Requires delegating distance computation
functionalities to TGeoNode
- Other changes required by the interface G4-TGeo
Future plans and wish list:
- Support for parameterized shapes. This will reduce the geometry size in memory for
certain geometries defined in G3 style.
- CAD geometry import
- Geometry builder GUI
His program of work is here:
* TLatex new symbols (at least #ell)
* Clean up the old OGL viewer (TPadView3D ... etc ... done by Fine)
* Get the color of a bin in a 2D histogram drawn with "COLZ"
* With options "c" or "l", drawing is done through the bins centers.
One could implement an algorithm extrapolating to the edges.
* PostScript: Markers size must be revisited in PostScript(and PDF)
* PostScript output: merge TPostScript and GL2PS. (needs investigation
* Implement progressively 3d plot (LEGO, SURF) with OGL. (needs
* Implement TPie, to draw pie chart (original code in HIGZ).
1. Clipping - resolve bugs with plane set clipping (mostly working) +
2. Add Translation/Scale/Rotate manipulators for all normal + clipping
shapes (first two done as part of clipping).
3. Fix coding convention violations + comment all code properly.
4. Draft external control (scripting) components of TVV3D - camera
(view) and clipping objects initially.
5. Intial support for large number of track/particles efficiently, or
externally 'drawn' objects registered with viewer - to be discussed with
6. CSG clipping - dep. on Timur.
7. GL-in-Pad updates.
7. Stabilise external control components of TVV3D.
8. Event display framework - first draft.
9. New GUI suitible for event display components + embedded pad version.
10. Add heads up camera for easier navigation inside detectors.
11. Look at scene rebuild efficiency - pre-create/recycle
12. Add more native GL tesselated shapes (as per spheres/tubes) if time
- look at most common ones in detectors geometries.
13. Consider decoupling from pad completely - emit signals for
selection, modification, request refilling of scene.
14. More clipping shapes - dependent on CSG working.
15. Create clipping plane(s) directly from camera - parallel to film
plane - with control to "push/pull away".
16. Restore TArcBall free perspective camera.
17. Investigate making TGLCylinder respect LOD flags at draw.
18. Interaction cursors.
19. Box zoom.
20. Lighting improvements - spotlights + carried light camera (miner's
my plans are as following:
for the October release: sPlot, TF23 minimization, sample quantiles in TMath and
qqplots from those sample quantiles
for the December release: Loess (nonparametric locally weighted polynomial fitting)
- XPROOFD and all related stuff
- Improvement and validation of the style manager functionality;
- Development of TAttPad and TFxx editors;
- To complete the missing functionality of the TStyleManager;
- Undo/Redo tools
- New User's Guide edition in December
- Completed GUI chapter + v.5 updates
Plans for next October release:
o consolidation of Vector package.
o add missing methods in rotation
o add transformation in 4 dimension
o define interfaces of IGenFunction and move from MathMore
o have first version of standard random package ?
* define interface to be used by TF1
* add additional algorithms from GSL: fft, multidim. integration, multidim. root finders
(problem: need probably to include GSL LA for some of these)
* add GMinuit using C++ Minuit as a tar file. C++ Minuit interfaces are not exposed, only used
Plams for December release:
* complete interfaces to functions and algorithms to have integration with TF1 like classes
* add some missing mathematical functions commonly used (incomplete beta and gamma)
* complete integration of new C++ Minuit in ROOT. Implement a backward-compatible interface.
* complete new Fitter interfaces
Tour de table
* The maximum number of histograms in a THStack plotted as lego plot was
18. Now there is no limit anymore.
* THStack colors are now stored in a dynamic arrays. Previously they were
stored in a fixed length arrays of 10 entries.
* New option "LINE" to paint TGraph2D. A 3D polyline is painted.
* Since skewness and kurtosis were available in TPaveStat, the default
value set by TStyle::SetOptStat(Option_t *stat) was not correct.
* Release notes ready
* Coding conventions ok in TGX11
* A problem with transparency: with TPad has been raised several time.
Now each time a TPad is painted a fill area is painted even if it
white. That is a problem in PS file because there is no way to remove
this white background which is not wanted when a PS file is included in
a document. I am working on it.