Minutes of Database, event, detector description session 10 May 2000

===============================
Database workplan - David Malon
===============================

  David Malon covered the near term priorities for the database work: 

    - access to existing data
    - data access through the framework
    - detector description infrastructure
    - database support for testbeam
    - other activities
    - viability assessment of the baseline
      - accumulate experience
      - reassessment of ATLAS requirements
      - evaluate scalability/viability testing done by others
    - grid related milestones

  Concerning the viability assessment, the question was asked whether
  other persistency would be explored. David Malon replied that there
  was general agreement that this would be done, e.g. with Espresso
  and/or RootIO.  David Quarrie added that the Gaudi framework has the
  possibility of a "democratic" store allowing different technologies
  to be used. This as well will be explored.
    

=================================================================
Status  of notes on identifiers and hits and digits - RD Schaffer
=================================================================

  RD covered a note which is defining a logical identification scheme
  for readout channels to be used for the software. A first version
  has been circulated to coordinators and db task leaders and has been
  discussed in the subsystems. A more complete version will be
  available and circulated before the end of June.

  A second note defining hits, digits, hits to digits transformation
  and pileup procedure intended to be used in Geant4 has been
  circulated. This note was largely incomplete - it contained the
  current definitions for Geant3 hits and digits.  There has been some
  discussion, and a more complete note will be circulated by end of
  June.

  In the discussion, it was pointed out the approval of offline
  coordinators is probably sufficient for certification of the final
  version of the notes.

  Traudl indicated that it may not be correct to define the digits as
  what is coming out of the ROB's. This was to be discussed further in
  the simulation session.

  There was some discussion on whether the pileup procedure should be
  described in the note. It became clear that it is sufficiently
  non-trivial that it should be clearly described in the note. 

=============================================================
Detector description: status and next steps - Stan Bentvelsen
=============================================================

  Stan gave a status report and summaries of two recent XML workshops
  (at LBNL 8/05/00 and at CERN 14/04/00).  

  He also summarized the status of AGDD. 

  It was clear that a key missing element is a scheme to specify the
  logical identifier scheme in XML. This is needed to identify and use
  the objects created from the XML in the generic model.  David Malon
  wanted to explore the possibility of specifying separately the
  logical model from the physical model in XML.

  There was also some dicussion on how to proceed.  It was concluded
  that there should be some small projects which goes all the way from
  XML to a Geant4 simulation for a small slice of some detector(s).
  This should be done before others are requested to proceed with
  descriptions. (See next point.)

===========================================
Reports from database subsytem coordinators
===========================================

  Steve Goldfarb summarized the work done for the Muons. In
  particular, he proposed to use the Muons for the above mention "test
  slice". This was further discussed in the simulation session the
  next day, concluding that there should be two projects, one for the
  muons and one for the inner detector, which goes all the way from
  XML to a Geant4 simulation for a small slice of some
  detector(s). This should begin immediately with something working
  before automn 2000.

  Tom LeConte summarized TileCal activities.

==================================================
Status of Objectivity Infrastructure - RD Schaffer
==================================================

  RD discussed in a bit more detail how today one can work with Objy -
  adding persistent class definitions to federations.  This will be
  evolved as described by David Malon's talk above.