Common Reconstruction Specifics

Recent Changes

• 29 Mar, 1997 -- Added information on RW files (Stephen)
• 20 Dec, 1996 -- Added section on current items (Stephen)
• 09 Feb, 1996 -- Meeting Update (Bob)

Current Items

• There was a meeting about changes to AbsEnv and AbsEvent. The (sort of) minutes for this meeting are avaiable.
• I've put a table together to tell you which files written on which platform can be read on which platform...just have a look, you'll see what I mean.

Outstanding Items

1. An event data model
2. A calibration architecture, which defines the various steps in processing that will determine the needed constants
3. The geometry architecture, including alignment

Needed Common Implementations

• Improving the coverage of the common code prototype. It now contains
  • Module runtime support
  • Access to event data via the AbsEvent interface
  • Event data input via dbio
  We need to have essentially every MC datum available, we need to improve both design and user documentation, and we definately need more examples.
• Develop access to other data besides beside the event:
  • Parameters for controlling processing, which the program should not change
  • Geometry and calibration data, which might change from event to event
  • Histograms and other common output data under module control
  There has been some minor progress on a "AbsEnvironment" proposal, but first efforts to use it for geometrical data are needed.
• Defining how we plan to associate reconstructed info with the underlying MC "truth". Fabrizio Bianchi, et al, have proposed a method for moving SVT truth data from bbsim via nested dbio templates. The next step is to map that onto a useful proposal for the format of reconstruction objects. This must address issues of object ownership and access via pointers.
• We have limited geometry information in several forms:
  • The bbsim simulation is based on GEANT3 geometry
  • The IFR and DIRC reconstruction code (each) includes some basic geometry definitions
  • The BbrGeom package includes some very simple classes to represent points, directions, various forms of lines, planes and simple boxes. This is intended as a first prototype to be used by the tracking code.
  We need to rationalize these efforts. To do this, we have to
  • understand the implications and requirements of alignment
  • determine whether a common geometry implementation can meet all the requirements
  • explore the possibility of reusing some of the GEANT4 code
  And we have to do all this relatively quickly.

Common Reconstruction Meetings

See also the workshop summaries:

• The third software workshop January 19 and 20 at Ecole Polytechnique, Paris.
• Second software workshop, December 6 through 10 at SLAC.
• First software workshop, October 23 to 28 in Genoa.

Eighth Meeting

The agenda is

• Data base plans and schedule (David Quarrie)
• Status of access to parameters and constants

  An old proposal exists for access to calibration constants and geometry. (The companion one on access to parameters has been replaced). The important idea is to encapsulate the actual access to the constants in a particular place, allowing the access method (database, dbin, dbio, etc) to change with a minimum of disruption to the actual reconstruction code.

  As we make progress on the contents of the calibration information, we should revisit this access proposal to make sure its still sufficient. <<LI>Continuation of the calibration discussion

Seventh Meeting

The agenda is

• Common calibration architecture (See comment in discussion system)

  the original discussion group item has been made more specific and received some comments.

  The next step is to firm these up to serve as a basis for the system-specific considerations. We will discuss the details of this in the meeting.

• Status of access to parameters and constants (this was moved to the next meeting)

Sixth Meeting

Late minutes were posted to the Hypernews list.

The agenda is

• Reconstruction requirements for calibration

  Are people assuming a "Pass 0"? Reconstruction in < N hours? Etc. A discussion group item on this is available to start the discussion

• Discussion of what to do for the April review

  We should work toward a coherent set of documents. A proposed outline being developed by the tracking group is available as an example of what the individual detector parts could cover. This must be combined with coverage of the calibration and geometry architectures, plus data model, schedule, etc, as common "chapters". A schedule and scope for this work must be established.

• Recent updates to common code

  The BbrGeom, AbsTrack and AbsEvent classes have new code in them. CVSNEWS is a new way of tracking CVS changes. We have a release coming up, and should take the opportunity to clean stuff up.

Fifth Meeting

• Storage of and access to parameter data and calibration and geometry data. Some older Common hit -> digi code (note the responses)
• Should we distribute CERNLIB, including the original problem and discussion (note responses) and some larger issues

Fourth Meeting

Third Meeting

Second Meeting

Bob's slides

First Meeting

Go to common code for reconstruction discussion pages or the full discussion list.