LCD Sim Weekly
8/2/05

---

NG: Now generating events and placed in LCD FTP area.

sidaug05_scinthcal was not sensitive, so regenerating.

Other files should be okay.

Everyone please look at single particles.

Tim continuing with 100 ab^-1 samples using WHIZARD.

SK: Found a problem with some other files ... sdfeb05 pion files are empty.

location --> ftp://ftp-lcd.slac.stanford.edu/lcd/ILC/singleParticle/sdfeb05/slcio/slic/   ?????

WL: Updated LELAPS so it generates the 64-bit IDs.  Fixed some GeomConverter bugs for GODL conversion.  
compact -> GODL -> run in LELAPS works okay, now.  Problems using event browser.

RC: [update on clusters]

NG: Updates to clusters from Tony?

TJ: Can now get raw energy and correct energy from CalorimeterHits.  Had to make some changes to cluster.
Also affected DigiSim.  Test runs but may have adverse effects.  Assuming single sampling fraction per subdetector,
but needs to be more general, e.g. for different layer thicknesses.

 
[discussion on sampling fractions]

WL: Would like access to sampling fractions for LELAPS, because now they are all = 1.0.

(Need to do sampling fractions by layer.)

RC: Sampling fractions only done once for calorimeters using hep.lcd.

SM: 3 ways to do sampling fractions.  

1) pions (take into account MIPs in Ecal).  

2) neutrals

3) fractions from ecal

NG: EM work from Graham?

GN: EB working on hmatrix going.  Needs to "call it quits" and do writeup.  

MP: Pattern recognition with pi0 decay.  How small an opening angle?  

JG (Jason Grey): Simple cluster builder to build clusters from hits.  From a MIP or not?  Remove clusters after
determined that from MIP.

MC: Uses NN with additional criteria, e.g. isolated hits.

SK: Remove hits from photons (2 possibilities now in org.lcsim).

Cluster unused hits f/ neutrons and K_L (simple cone algorithm).

Quality cut for remaining clusters (try SLAC Neural Net).

NG: RP on tracking?

RP: CCD reco from Nick Sinev and Mike Ronan.  Wanted to bring some functionality to org.lcsim.  

Nick's code -> picks three points, makes circle and makes tracks from various strategies.

Will leave out optimizations for bkgrnds, etc.

JS: 

[slides - ZvTop]

ZvTop makes parabolas to approximate the track.  Creates a Gaussian tube around this parabola.  For all pairs of tubes,
finds the overlaps, finds maximum, and clusters maximum.  Tracks fitted to loc of max in space.

Trying to port from hep.lcd to org.lcsim.

Split into parts:

1) finder - works to first approximation

2) fitter

3) swimmer

MR: 

[slides - TrackingCheater]

For each SimTrackerHit, finds MCP parent and creates a corresponding cheat track.  Uses the MCP's momentum vector
to satisfy FoundTrack.  FoundTrack differs from FullTrack -- nothing about fitting.