Discussion on the the IO and folder structure:

Yves said that the multiplication of files probably would make the access more
unstable, especially on HPSS.

Bjorn expressed his concern on the consistency of the proposed file structure.
It can be difficult an misleading if one uses many events and many 

Predrag explained that if the file structure would be fixed in a directory, 
than the management caused no problems. 

Rene thinks that the proposed structure is very similar to the existing
schema and has similar problems as AliRun. It would be better to have an 
lightweight AliDetector class where you add the needed functionality. 
 

Federico: The biggest problem is the geometry. One has to know how it is 
retrieved for a given purpose. The parameters are needed very often, but one
also might use the information about the materials (for the Kalman filter).

It was proposed to review the new design and implementation by a small group of
experts.

Working groups:

Virtual MC: Rene, Ivana, Peter
New IO: Rene, Fons, Piotr, Yves, Jiri, Ivana
Fluka: Endre, Alberto, Andreas, Marian
Track references: Marian, Fabrizio, Marek, Nicola, Christoph, Bjorn

Alberto explained the structure of Fluka/AliRoot interface and some features
of Fluka transport.