Supported Operating Systems:
Windows NT, Sun Solaris, Linux
decision taken at the Jan 2000 Software Workshop
Code Management:
cvs is used to do revision control (and has been for a long time)
code units will be organized into 'packages'. The directory structure is such that the package directory will contain src/, <package-name>/ (for public includes), test/ and doc/ directories. There can be other directories. The existing GLASTSIM code has been reorganized into packages.
Responsibility for package updates will be delegated to individual owners. People will not have write access to packages they do not own or share ownership of.
decisions taken culminating at the March 2000 Collaboration meeting.
build & distribution control will be performed via CMT
decision taken following at the May 2000 workshop.
Migration to GEANT4
we will provide a migration path to allow use of both Gismo and GEANT4 in our simulations. Target date for implementation is September 2000.
decision taken at the Jan 2000 Software Workshop
Persistent Object I/O
we will use Root I/O as our persistent object store.
decision taken at the Jan 2000 Software Workshop
Supported Analysis Packages
we will support IDL and Root as analysis packages
decision taken at the Jan 2000 Software Workshop
Automated Code Documentation
We will mandate inclusion of in-line documentation to allow automated production of a 'reference' code manual. The automated tool is DOxygen.
decision taken at the May 2000 Software Workshop
C++ Code Architecture
it is proposed that we adopt the code architecture GAUDI, developed by the LHCb experiment.
decision taken at the Aug 2000 Core Working Meeting
Document Control System
for better or for worse, we have adopted Hummingbird's CyberDocs management system.
decision taken in summer 2000
Project Management Tool
are using MS Project 2000. We are still evaluating its web interface, Project Central.
decision taken in summer 2000
Handling Time-Dependent Constants (proponent: Joanne Bogart; target decision date May 2000)
it is proposed that we adopt the MySQL relational database to control time-dependent constants handling. This may include only meta-data or the constants themselves. TBD.
Coding Rules (proponent: Traudl Hansl; target decision date May 2000)
it is proposed that we mostly adopt the coding conventions put forward by the LHC collaborations. We could make minimal changes to these conventions.
R.Dubois Last Modified: 01/08/2001 09:08