1. MINER A Software The Goals Software being developed have to be portable maintainable over the expected lifetime of the experiment extensible accessible to non-experts allow MINER A to meet its physics goals

2. It have to re-use as many existing software components (developed and maintained outside MINER A) as possible, focusing the collaboration’s limited resources on areas specific to the experiment and improving the reliability of the system as a whole. It should comply with established laboratory computing security policies, in order to re-use as much existing Fermilab infrastructure as possible The Goals

3. Categories of the software internally developed software to operate, simulate, reconstruct, analyze and visualize MINER A data. externally-supported packages which provide services we do not wish to re-invent. specialized, but non-MINER A-specific, software, will be treated as external software, even if supported by MINER A collaborators.

4. Platform Support Fermilab-supported implementations of Linux Windows platform Macintosh OSX 64-bit hardware is not explicitly supported now. it planned for the future

5. Code Management CVS (Concurrent Versioning System) a widely-used utility for source-code management. One can expect that any next-generation utility will be able to import a CVS repository. repository for the MINER A software project is established and maintained at Fermilab

6. Build System and Configuration Management Configuration Management Tool (CMT) used by a number of large experiments, including ATLAS, LHCb and GLAST used by the LHC Computing Grid (LCG) will be supported throughout the LHC era designed to allow the same software base to be built and used on a variety of platforms

7. Software Model The object-oriented model (C++) -entails a longer learning curve -more careful design ₊scales good to large systems ₊encourages an architecture based on interchangeable components with well-defined interfaces ₊Most currently maintained software packages for high- energy physics (ROOT, GEANT4, PYTHIA, etc) either use an object-oriented model or are being ported to this technology.

8. The procedural model (FORTRAN) ₊ has the advantage of familiarity (at least to older colleagues…) -scales poorly to large and complex systems -most physics software packages designed for the procedural model (PAW, GEANT3, CERNLIB, etc) are no longer supported. A hybrid model -problem maintaining data structures in two different programming languages -calls between different languages introduce platform- dependencies that are best avoided

9. Framework A common point of reference for developing and using data-processing applications. well-designed framework: Provides common idioms for writing code and handling data shielding the end-user from the internal “plumbing” (data flow) and “wiring” (control logic) overhead necessary to perform computing tasks, allowing the user to concentrate on the actual physics-related programming

10. well-designed framework: allows individual software components to be freely reused and interchanged supports a master description of detector geometry and running conditions, to be used consistently in data simulation, reconstruction, analysis and visualization, provides a common set of centrally-maintained tools and services used by diverse applications.

11. GAUDI framework C++-based, object-oriented frameworks used by LHCb, ATLAS, HARP and GLAST. will be maintained and developed during the LHC (and MINER A) eras  http://proj-gaudi.web.cern.ch/proj-gaudi/ http://proj-gaudi.web.cern.ch/proj-gaudi/  http://lhcb-comp.web.cern.ch/lhcb- comp/Frameworks/Gaudi/ http://lhcb-comp.web.cern.ch/lhcb- comp/Frameworks/Gaudi/  http://www-glast.slac.stanford.edu/software/Gaudi/ http://www-glast.slac.stanford.edu/software/Gaudi/  https://twiki.cern.ch/twiki/bin/view/Atlas/CoreSoftware https://twiki.cern.ch/twiki/bin/view/Atlas/CoreSoftware

12. Data Persistency and Management POOL framework provides a variety useful services for storing and accessing event data supports a distributed data model and allows the user to reference logical datasets without worrying about where and how the data is physically stored. GAUDI is fully integrated with POOL and is able to read and write data using POOL logical and physical filenames. supports ROOT I/O: POOL data files written in ROOT format are browsable with that program and support links (pointers) between objects in different files persistent storage of LHC data will rely on POOL => continued support over the lifetime of MINER A