Because the size of the Skim1 output data sets (200-500 tapes) is still too large to be easily used by an individual, the six Skim1 were split again into many sub-streams. Some events that did not pass more stringent cuts were discarded at this point, but not all skims applied additional cuts.
Skim2 processing was accomplished at five institutions, as shown in recon:superstreams. From each , 5-12 sub-streams were written, covering a variety of physics topics and calibration data samples. The sub-streams relevant to the analyses presented in this thesis are presented in recon:substreams. The Skim2 computing models varied by institution, but were generally similar to that used for Skim1. The first Skim2 process began production in January, 1999 and all Skim2 processes were completed by June, 1999.
| Super | Sub | Physics | Number |
| Stream | Stream | Topics | of tapes |
| 4 | FSDD | 15 | |
| 4 | FSDL | and  |
17 |
| 3 | FSCG | Golden Modes | 44 |
In this chapter we detail the methods used for selecting candidates and explain many of the reconstruction methods used in FOCUS. While portions of this discussion are only applicable to reconstruction, other portions are applicable to the reconstruction of the , , and candidates in later chapters.
In this thesis, only the decay mode of the is considered. This decay
mode has an absolute branching ratio of 
%. This decay mode is
chosen for its large branching fraction and ease of reconstruction (all of the
decay products are charged and long lived).
select:dvert explains the vertexing algorithm used in FOCUS and the cuts it makes possible. select:citadl describes some of the cuts available from the algorithm. In select:other two kinematic cuts are investigated. select:vectors explains two ways of improving the momentum vectors for charm daughters. Finally, in select:trees the effects and interactions of some of these cuts with each other are shown and the process of selecting a useful set of cuts for an analysis is explained.