Comments on the status of SvtHitReco and its OO Design

Paola Grosso, Diego Zanin    INFN - Torino

Pete Elmer, John Walsh    Univ. of Wisconsin

The Booch diagrams provided with these comments review the current status of the package SvtHitReco, that is what is currently implemented. For the hit building part, that was previously discussed but not implemented yet, see the former design documents keeping in mind that at implementation time we will likely change some features.

At the moment the package is almost stand-alone. In the near future it will likely use plenty of classes from SvtGeom, while it is not planned to get dependent on SvtReco. It will provide an alternative input to a track finder.

We use, for the time being, only lists of pointers. In the class diagrams most of the links expressing a has-a by value relation with cardinality n, corrispond in the code to a class containing a list by value. Actually, the objects appended to the list exist in the heap but the container class takes care of creating and deleting them through their pointers.

Here we call permanent objects the ones containing the SVT hardware information that will be read from a database, and need not be read each event.

As for the rest, we have not discussed in further details what is saved in the event store and what is deleted at the end of an event. In general, objects whose pointers are listed in a BookKeeper will be saved.


Comments to Diego Zanin
Last modified: Tue Apr 22 16:16:21 MET