Software Meeting Minutes
August 10,2004
D. Lawrence


Attendees: (It was almost 2 weeks ago so I forget)

Agenda:

1. HDDM and in-memory formatting. What's the definitive source
(XML, XSL, both?) Are there any modifications needed to HDDM?

2. Getting married to ROOT: Should we take the plunge with the 
C++ analysis code? Should we avoid this in general and if so, at
what cost?


1.
The topic of HDDM was discussed. The question of how the XML
template, XML schema, and C header file are related in the heirachy of
defining data structures or classes used for in-memory format
of the data. Specifically, tools exist to derive templates from
schemas and to produce schemas from templates. Which should be the
definitive source?

Richard described the relationships and the intended purpose of
the tools. Schemas are the more powerful and complete description
language so it was tenatively decided that we should adopt them as
the definitive source for defining HDDM C/C++ data structures.
We agreed everyone should look up some "Schema for Dummies" type
reference to learn a little more about XML schemas. Richard will
lead another discussion on schemas next week to lay out what
benefits they give us over a simpler XML implementation.

2.
Hall-D data will have need for 3-vector and 4-vector objects
in numerous places throughout the reconstruction/analysis code.
We should adopt a standard for this. Dave proposed we commit to
using the ROOT implementation. This would mean that ROOT would be
required for building Hall-D analysis code. Some discussion of
other packages was had, but there was no clear argument for using
something other than ROOT.