Known Inconsistencies with KF-based Alignment

[tomeichlersmith]

There are two (small) inconsistencies left on master that I think can be removed now that we are more confident in using KF-based Alignment.

• Removal of odd-number-hit tracks. This was originally done to make the comparison to ST-based Alignment more simple, but we should not be throwing away e.g. 13-hit tracks because that is the whole reason we wanted to switch to KF-based Alignment.

hps-java/tracking/src/main/java/org/hps/recon/tracking/gbl/SimpleGBLTrajAliDriver.java

Lines 577 to 584 in 2216d8c

	if (TrackType == 1 && track.getTrackerHits().size() % 2 == 1) {
	// this is a KF track with an odd number of hits which /cannot/
	// be equivalent to a GBL track so we are going to skip it
	// in a future where KF-based alignment is the standard,
	// we probably want to remove this
	continue;
	}

• The number-of-hits cut in the alignment driver is using the old ST-based style of counting hits (i.e. a module where both sensors combine to form a single hit). We should abandon this style in favor of the simpler KF-based style of counting (each sensor hit is a single hit).
  • I'm not fully committed to doing this since it would then require all steering files that use this driver to be updated.

hps-java/tracking/src/main/java/org/hps/recon/tracking/gbl/SimpleGBLTrajAliDriver.java

Lines 556 to 560 in 2216d8c

	//Kalman
	int actualHitCut = nHitsCut;
	if (TrackType == 1) {
	actualHitCut = 2*nHitsCut;
	}