Disclaimer : Please forgive my ignorance. Please take the following discussions as suggestions.

Data Structure -> Root?

The data structure is a building block for event reconstruction. It defines how the data to event interface will have to be written. Root is well-suited for physics applications. It supports various data writing schemes. It has an example Event object which is coupled to tracking. Great for writing calibration and reconstruction software.

Geant -> Geant 4? Geometries?

It would be ideal for the standard Qweak simulator to use Geant4. G4 is C++ based, whereas G3 is in FORTRAN. Incoming students are likely to have zero FORTRAN experience, but may have C++ experience. Learning C++ and its applications would be much more useful than forcing students to learn FORTRAN.

Geometries : XML or G4?

Other collaborations dealing with G3 and G4 are using XML based geometries. This format allows for both G3 and G4 simulations to be run on the same detector geometry, preserving G3 capability. This is very useful for the continuation of detector studies carried out by G3 users. There exist utilities that convert G4 geo to XML geo and vice versa. There may also already exist utilities for G3, but I am not aware of them.

Geant4 also supports g3tog4 which converts G3 geometries directly to G4, but does not have backward compatibility. However this may be useful for G3 by converting to G4 and then into XML which may be used by G3. Is this feasible?

Maps

What is the current map system? How robust should detector geometries be? There must be some allowance for changes due to design changes, calibrations, and run-time changes such as wire failures.

Calibrations

At some stage calibration software will have to be written. What calibrations will each individual detector undergo before event calibrations? There should be a list of these by each detector group, agreed upon by the tracking group. These need to be considered so that calibrations do not take ~6 months due to poor preparation. This is often the case when calibrations are performed by students that are at an introductory experimental physics/programming level.

Tracking Group as it relates to software

There at least needs to be an email list for people working on tracking. Software designers need to work in parallel to encourage standards. Usage of a forum may also be useful. One exists but is no longer maintained apparently. A forum would become a repository of discussions and a place to go to for help/info.