About BrHCVertex: (mainly to Bjorn) =========================== I understand that finding a vertex using a histogramming method (BrHCVertex) is supposed to be fast and reliable. But still I am not 100% convinced that that's the best way to calculate production vertices for BRAHMS. Somewhat I still believe that using track parameters rather than hits would do a better job to calculate a vertex position except for events which have too few reconstructed tracks. Before we declare "BrHCVertex" as "the" vertex finder, I would like to see more systematic comparisons between two methods with code optimizations. I also would like to mentioned that it's quite important to be able to calculate "3-dimensional" vertex positions (x,y,z) without assuming x=y=0 especially for the calibration purpose. About BrVertex Module: (mainly to Kris) ============================== I think it's okay to have a general vertex module to include vertex information from the BB's and ZDC's. But I also think that it might not be a bad idea to distinguish (separate) the "vertices" calculated using timing information in the global detectors from the vertices calculated using spatial information in the tracking devices. I personally would rather have a Vertex class (BrVertex) including a few methods (using MTPC1 tracks, using spectrometer tracks, using MTPC1 hits without tracking, and so on) using only spatial information from tracking detectors. I think that the vertex information from BB (ZDC) can be a separate module can belong to brat/bb (brat/zdc) I think. But anyway I don't have any objection to have a more general vertexing module including vertex information from BB and ZDC if we think it's better for the integrity of the brat structure. And one more thing, the current BrVertex includes more information than just x,y, and z: closet approach, number of tracks used, chi-sq, track id used for vertexing, and so on. JH
This archive was generated by hypermail 2b29 : Mon Jun 05 2000 - 13:53:34 EDT