The Kalman filtering track finder in CMSIM is based on the 3-D weighted-means formalism of Kalman filtering method and the fast non-GEANE extrapolation for both track parameters and weight matrix (i.e. inverse of the covariance matrix). The current release is tentative, many tuning and refining works need to be done. The package intend to perform the simultaneous track finding and fitting for both charged primary tracks and secondary tracks decayed from secondary vertices even dozens of cm away from the primary vertices.
The main body of the package (in TKFFFM.F) is self-sustained without relying on GEANT structure. All tracker geometrical information are obtained from the tracker GEANT simulation program via the interface in the steering routine RETR_KFF.F, which also includes all interfaces for reading input from RHIT bank and writing output to TPAR bank. Many adjustable running parameters have not been put into DATA CARD yet, they should be eventually there for being adjusted by users.
The package is highly modulized without much communication by common blocks. Most routines in it can be directly used in a C++ Object Oriented version of the package.