[Mitgcm-support] patches and branches

[mitgcm-support at dev.mitgcm.org]

The latest CVS web has a nice feature 
for filtering files based on tags.

Can we adopt a convention for tagging
branches for patching and or development
so that names are short but meaningful.

Heres my thought for the day

 1. checkpointNN - a checkpoint
    e.g. checkpoint36

 2. b_cNN_PROJ            - a branch from checkpoint
                            NN for project "PROJ".
    e.g b_c36_ATMOS-MERGE

 3. b_cNN_proj_sNN        - phase/stage NN marker for branch.
    e.g. b_c36_ATMOS-MERGE_s03

 4. b_cNN_pNN             - branch for patch NN for checkpoint NN.
    e.g. b_c36_p01

 5. b_cNN_pNN_sNN         - phase/stage NN marker for branch.
    e.g. b_c36_p01_s03

 6. releaseNN             - a release!
    e.g. release01

 7. b_rNN_pNN etc...      - branch for a patch to a release.

 This means that we reserve the following characters
  _    used to separate "words"
  b,r,p,c,s - at the start of a word means
              branch, release, patch, checkpoit, stage
              respectively.
  Other identifiers e.g. PROJ are UPPER CASE ( + numbers ).
  
Chris