(More or less detailed) next steps for tracking:
- Finalize track merger (without fit)
- Fit merged tracks (RefitFinal processor might be able to do this out of the box)
- Some initial track performance studies
- (Most likely) iterative tuning of this approach until performance meets criteria
- Alternatively: Use TPC tracks only to get something going quickly (while the combined Silicon + TPC tracking is being refined)
- Need to investigate why RecoMCTruthLinker fails, but might be feasible to solve quicker than the rest
- Could also check if "picking up silicon hits" Clupatra option works link to ILDConfig
Larger picture:
- Pandora will almost certainly need a re-tuning for ILD@FCCee (this will be quite a bit of work potentially, as there is no real expert left that can do this from scratch)
- Might be useful to strip almost all of the sophistication of ILD / CLD away and build it up from "basic principles"
- ML based PF would be very interesting for ILD (IMO)
- Could be done with ILD@ILC (depending on how low level you want to go)
- Could also just use the TPC tracks as starting point (ignoring inner tracking parts)
Technicalities
- Ideally these developments are done via Gaudi algorithms rather than Marlin processors
- Tracking efforts from Victor already going into that direction
Other questions:
- cms energy?
- do we need samples officially produced?
- specific (technical) resources / documentation?