COBRAS Coherent Object-Oriented Base for Reconstruction, Analysis

1
COBRAS Coherent Object-OrientedBase for
Reconstruction, Analysis Simulation
Objectives and Plan for 2001

• Vincenzo Innocente
• CERN/EP/CMC

2
Strategic Objectives

• Coherent and Uniform Approach to Physics Analysis
  from Generators to Publication-quality Plots
• Coherency and Uniformity in
• User Interface
• API
• Data Storage
• Software Development
• No compromise on performance
• Overall speed limited by raw I/O and physics
  algorithms

3
ORCA/CARF prototype

• CMS have identified in the use of a common
  Framework and of a base Toolkit the way to
  achieve these strategic goals.
• At the same time the direct involvement of
  detector and physics groups has been recognized
  to play a key role in the development of specific
  software.
• The success of the ORCA/CARF functional prototype
  has convinced us (and our internal and external
  referees) of the feasibility of this approach and
  encouraged us to extend it to the other physics
  software projects.

4
Reconstruction Analysis Software System
Physics modules
Specific Framework
Reconstruction Algorithms
Data Monitoring
Event Filter
Physics Analysis
Generic Application Framework
Calibration Objects
Event Objects
Configuration Objects
adapters and extensions
Utility Toolkit
ODBMS
C standard library Extension toolkit
Geant3/4
CLHEP
Paw Replacement
5
Objectives for 2001

• Make OO Simulation (Generators and Detector
  Simulation) coherent with reconstruction
• Make short-cuts (fast simulation/reconstruction)
  possible in a coherent and transparent way
• Develop a strategy for a coherent
  final-analysis beyond current ORCA
  functionalities
• Understand Requirements for Detector Description
• Continue to improve the current software system
  taking into account feedback from users

6
COBRA

• Identify and isolate base software common to all
  physics software projects
• Build atop of the ORCA success
• Split-out CARF, Utilities and few other base
  packages
• Incorporate base software from other projects
• Modify ORCA, OSCAR, Famos, Iguana to be COBRA
  aware

7
PRS Responsibility

• Identify and isolate detector-specific software
  make the corresponding PRS group responsible for
  it
• Identify and isolate common physics software (for
  either simulation or reconstruction) create
  specific toolkits organize them as a global PRS
  responsibility.

8
Milestone Plan

• Give to Caesar.

• Create COBRA keep current interfaces and
  functionalities
• Make ORCA and Iguana working with it
• Drive OSCAR as a CARF SimApplication
• Make OSCAR Hit-formatting a detector-specific
  code
• Implement coherent and uniform hit persistency
• Split Visualization in detector-specific code and
  common toolkit
• Split Geometry in detector-specific code and
  common toolkit

9
Milestone Plan

• Improve COBRA coherency and uniformity in UI and
  API
• Develop a COBRA generator interface and
  persistency scheme
• Extend COBRA MetaData to cover all aspects of
  simulation and reconstruction for both Production
  and Single-User mode
• Repackage COBRA to reflect new software
  organization
• Reengineer OSCAR to better serve CMS requirements
  for software development, maintenance and physics
  analysis

10
Beyond Current ORCA

• Today ORCA coherent approach stops with the
  production of high-level Reconstructed-Objects
  that eventually can be made persistent.
• (ORCA fully supports batch-oriented,
  high-granularity incremental reconstruction,
  event filtering, and high-granularity physical
  reclustering)
• Is it possible to keep the same coherent and
  uniform approach also in final analysis (up to
  the production of publication-quality results)
  without compromising global performances?
• Better, how to enhance global performance in an
  area, such as final analysis, where human factors
  are often more important than raw computing power?

11
Speculative Milestone in Final Analysis

• Define performance indexes for Final Physics
  Analysis (FPA)
• Evaluate performances for different FPA
  strategies
• Evaluate impact on current COBRA approach
• Design and implement prototype(s)
• What really we are not lacking of
• Evaluate prototype(s) for realistic case-studies
• Evaluate alternatives for the same case-studies
• Decide and proceed toward a functional prototype

12
Conclusions

• Reasonable AFT Objectives for Autumn 2001
• More Coherent and Uniform CARF API and UI
• Higher quality for production (control,
  validation, recovery)
• More detector people able to contribute to
  simulation and visualization software
• ORCA, OSCAR and Famos integrated with COBRA
• A prototype of an Analysis Environment
• A document on Analysis requirements (not in AFT
  scope?)
• A prototype for Detector Description