Re: [ROOT] Expansion of the Geomtry classes for simualtion codes

From: Valeri Fine (Faine) (fine@bnl.gov)
Date: Wed Oct 04 2000 - 02:17:26 MEST


Hello Christian,

Thank you very much for your information concerning TVolume classes.
I hope you realized these classes were designed with GEANT-like simulation
in mind. Alas the functionality these classes provided is till far away of what GEANT3
did many years ago.

 We had provide a utility to create TVolume's from the GEANT ZEBRA structure on fly.
 This gives the possibility to visualize GEANT geometry via OpenGL directly.
(see our ROOT '2000 
http://root.cern.ch/root/R2000Program.html
ftp://root.cern.ch/root2k/pavelR2000.ppt
http://root.cern.ch/root/R2000Html/pavelR2000/index.htm

ftp://root.cern.ch/root2k/StarFrameWorkAbd3D.ppt

and CHEP'2000 talks 



----- Original Message ----- 
From: Christian Holm Christensen <cholm@hehi03.nbi.dk>
To: <fine@bnl.gov>
Cc: <roottalk@pcroot.cern.ch>
Subject: [ROOT] Expansion of the Geomtry classes for simualtion codes


> Hi Valery, 
> 
> I have a request/suggestion to the expand the TVolume class, or some
> sibling, or whatever, with a TMedium and TField members. 
> 
> The idea behind that, would be to have some class that are similar to
> the objects in GEANT3, that is (sort of ordered by containment or
> reference) 
> 
>          TVolume
>     |
>     +-TMedium
>     |    |
>     |    +-TMaterial/TMixture
>     |    |
>     |    +-TField
>             |
>             +-TNode
>         |
> +-TShape 
> 
> Of course it would be beneficial to have a list of TMedium and TField
> objects in TGeometry or similar.
> 
> I believe such a set of classes would be very useful for detector
> simulation, and generic enough to suit most detector simulation codes
> (like GEANT3 and GEANT4) needs. I did play around with something like
> that some time ago and I have some (ugly) code if you want it. 
> 

 Our experience teaches one needs 3 design loops at least to get more-less 
"generic" object-model and class library. That somebody but the class author 
 may use. This means any  real implementation  of the object model is a valuable 
 asset.

 We at STAR have had several discussions which direction  TVolume should go.
 We are not ready to present our results yet but appreciate any idea and codes.

There are many  ways to do things  (and to screw things).

> Anyway, that's my to pennies worth on that subject. 

  I appreciate alot you give us a chance to see your sources and the real application 
  you built with that (if any).

  With my best regards,
                          Valery



This archive was generated by hypermail 2b29 : Tue Jan 02 2001 - 11:50:34 MET