Guidelines for paper code preparations

• General guideline: people should be able to reproduce paper figures using provided instructions, codes and histograms

• What should be included: 
  
  • Analysis source code, such as analysis makers, job submission script, plotting macros, etc 
  • Modified files of standard makers from STAR library
  • Small rootfiles containing necessary histograms, including those from running the full statistics, for reproducing paper figures
  • Instructions on how to run the code such that Code QA can reproduce the paper figures
  • Final results based on a small amount (1-2) of input PicoDst/MuDst and/or embedding files. These results will naturally have large error bars, but code QA should be able to reproduce them exactly.
    • Input PicoDst/MuDst and/or embedding files used to generate the results above should be restored and saved on RCF, but not needed to be committed to CVS

• What should not be included:
  • Secondary files, such as library files (.o, .so), figures, etc, that can be produced from source code
  • Log files
  • Empty directories
  • Remove all the directories named "CVS" (check the subdirectories as well). Otherwise, one might get into trouble during committing. 
  • Standard makers that can be obtained from the STAR official library. Instructions should be given on how to retrieve the correct version of these makers. If some files are modified within a standard marker, only the modified files should be committed
  • For common codes, such as StRefMultCorr, that are used multiple times in different parts of the analysis, they should be committed only once.
  • Rootfiles of raw data, embedding data, analysis trees, etc.
  • Private embedding data, analysis trees, analysis histograms etc are recommended to be backed up to HPSS (How-to)