Tracking issues

Regarding 16-series (run 5 production)

1. t-to-d resolution function for data
   • Last discussed: Jan 19 2005
   • Summary: The DCH hit residual pulls on data are ~15% wide for low momentum tracks though they seem consistent with unity for high momentum tracks (which were used for calibration). Since the momentum dependence is not understood and low momentum tracks are more relevant for BaBar physics, we should base the calibration on a more appropriate sample.
   • Who: Wouter Hulsbergen
   • Status: Almost done. The updated resolution function is available, documented, and being implemented (most recent HN discussion of remaining issues is here).
2. t-to-d resolution function for MC
   • Last discussed: Jan 19 2005
   • Summary: The resolution functions used for 'smearing' and 'reconstruction' are not entirely consistent leading to DCH hit residual pulls that are ~15% wide.
   • Who: Gerhard Raven
   • Status: Ready to go. Will increase the resolution by a factor of 1.1 and decrease the smearing to be 1 x resolution instead of 1.1 x resolution Actual implementation is tied to (1.).
3. Tuning of DCH MC parameters for SP7
   • Last discussed: Jan 19 2005
   • Summary: The parameters of the (improved) DCH simulation need to be tuned and validated through tracking efficiency studies. This was complicated and delayed by an observed drop in the pseudoefficiency for low momentum tracks due to a bug in DchTrkGeom which by now has been found and fixed.
   • Who: Jeff Wilson + Tracking Efficiency Group
   • Status: In progress.

Planned for 18-series (run 1-4/5 reprocessing)

1. DCH time-to-distance calibration
   • Last discussed: Jan 26 2005
   • Summary: The t-2-d calibration needs to be updated, the minimal goal being re-running the exsting code (in an up-to-date release). It'd be desirable to go beyond this, though, since there are several specific ideas how to improve the procedure.
   • Who: Liaoyuan Dong
   • Status: In progress.
2. Revisiting SVT local alignment
   • Last discussed: Jan 26 2005
   • Summary: In view of a recent bug fix we should consider re-running the SVT LA (possibly including further improvements) before the reprocessing.
   • Who: Chung Khim Lae, Jürgen Kroseberg
   • Status: Evaluation in progress.
3. Improving material constants
   • Last discussed: Jan 12 2005
   • Summary: studies of reconstructed mass biases indicate that there is room for improving the detector material (and B-field?) description. Before the reprocessing we should collect the information and come up with a (tested!) proposal.
   • Postings: hn 05/01/10, hn 04/12/10, hn 04/12/01, [...], hn 03/07/16, hn 03/08/13, hn 03/08/20, hn 03/08/27, hn 03/10/29, hn 03/11/03
   • Who: Past and on-going studies by Brian Petersen, Jon Coleman, Bill Dunwoodie, Wouter Hulsbergen,...
   • Status: In progress.
4. Implementation of rebuild mode
   • Last discussed: Jan 12 2005
   • Summary: In 'rebuild mode' tracks are re-reconstructed from scratch. It requires the original hit lists. Currently, two types of hits are persisted, namely hits-on-tracks and unassigned hits. We need procedures that create 'raw' hits from hits-on-tracks. The problem must be solved separately for DCH and SVT.
   • Who: Gerhard Raven for DCH, John Walsh for SVT
   • Status: Pretty much done for DCH. More work needed (equivalent to 2 weeks full-time) on the SVT side. Will miss Feb 1 deadline.
5. Track cleanup module
   • Last discussed: Jan 19 2005
   • Summary: Dave has proposed a TrkFixup package collecting various track cleanup tools. The intermediate goal is to provide a tool box at the beta level. Most probably irrelevant for 18-series (unless some actual bugs with obvious fixes are found through these studies).
   • Who: LBL group (Dave Brown, Kerstin Tackmann, Marco Battaglia, Gerry Lynch) +additional manpower to be found.
   • Status: Work has started one some of the issues. Some first results have been obtained (for an example see here). None of this is targeted for the Feb 1 deadline, though.

OUTDATED list (to be cleaned up soon) of other open issues with no(?) definite time scale

1. Hit efficiency of DCH in SP6
   • Last discussed: 04/06/09
   • Postings: hn 03/08/12 hn 04/06/16
   • Who: Jeffrey Wilson
   • Status: done for SP6. validity range must still be changed once SP5 is frozen.
   • Priority: stay tuned.
2. Improvement of DCH amibuity flipping
   • Last discussed: 03/08/06
   • Postings: hn 03/07/03 , hn 03/08/06
   • Symmary: Gerry Lynch has found that the ambiguity flippers in are not doing a very good job, often making the tracks worse rather than better.
   • Who: ?
   • Priority: Low?
3. Move T0 persistence out of tag DB
   • Summary: Because the bunch T0 calculation requires DCH hits, it is not performed when running in cache mode. The T0 values are currently stored in the tag DB. That is not the correct place.
   • Status: needs volunteer
4. SVT tracks with too few hits
   • Last discussed: 04/04/14
   • There is a conflict between the pattern tests performed in the SvtHitAdders and the arbitration process that happens later on: The arbitration can remove hits from tracks. After the hits are removed, the track no longer satisfies any of the initially imposed patterns. This for exmaple leads to SVT tracks with only 1 or 2 remaining hits, which normally is not allowed. This poblem cannot easily be solved. One solution is to drop the arbitration altogether.
   • Postings: Wouter 04/04/08, Eric 04/04/13, Gerry 04/04/14, Gerry 04/04/18
   • Who: Eric, Gerry, Mike
   • Status: aimed for 15-series
   • Time scale:
5. Static functions in need of initialization in SvtWaferID
   • Last discussed: 03/07/30
   • Postings: hn 03/07/29
   • Who: some SVT software expert
   • Status: low priority
6. include splitting the MC match map in TrkSplitTracks
   • Symmary: MC matching is lost for split tracks.
7. ELF crashes in the tracking code
   • Last discussed: 04/04/14
   • Summary: Shahram has collected all tracking related crashes of ELF. Most have been resolved. There is one assert left in DchHitOnTrack. The problem seems to depend on optimization, so most likeli it is not easy to solve.
   • Postings: Shahram 04/04/15, Shahram 04/04/21, Shahram 04/05/18
   • Who: Shahram, Gerhard
   • Status: in progress
   • Time scale: unknown
8. Simplifying DchHitOnTrack, Mini vs Reco
   • Last discussed:
   • Postings: hn 03/09/05
   • Who: some DCH software expert
   • Status: low priority
9. Active hots in DCH OprMon
   • Last dicussed: 04/02/05
   • Postings: hn 04/02/01
   • Summary:
   • Status:
10. repair of the MatEnv proxie
    • Summary: the proxies in MatReco are faulted in beginJob, and by design can be faulted only once. Repair needs modifications in MatReco and several subdetector packages that set pointers to materials in beginJob.
11. Proxies faulting in beginJob
    • Summary: There are some tracking related proxies that fault (ie access the DB) in beginJob.
    • Last discussed: 04/02/25
    • Postings: hn 04/02/24, hn 04/02/24
    • Who: ??
    • Status: urgency unclear (to me)

Low priority

1. mini skim based on PC-filter
   • Last discussed: 03/10/01
   • Who: ?
   • Status: low priority until next (Februari) skimming
2. Procedure for global alignment on magnet-off cosmics
   • Last discussed: 03/09/10
   • Postings:
   • Who:
   • Status: