[Mitgcm-support] Re: CVS policy

[mitgcm-support at dev.mitgcm.org]

I think thats fine except.....

 1. I imagined that one day we would be able to say
    checkpoint56798 provides a framework for using
     kpp_checkpoint34567
     gmredi_checkpoint9012473248
     calendar_checkpoint14393259
     wrapper_checkpoint0 :-)
     core_dynamics_checkpoint9094
     cg2d_solver_checkpoint57667
     direct_solver_checkpoint312432
     mdsio_checkpoint34225245
     netcdf_interface_34782345
     etc....
    In this approach, by the time we reach checkpoint56798, there 
    would be a specific thing that people could identify as 
    kpp_checkpoint34567
    which would include source, readme's etc...

 2. In checkpoint56798 it ought to be possible to have an
    example run that shows how to use kpp_checkpoint34567 within
    checkpoint56798.
    
 You say that there will be no kpp_checkpoint version and that the 
test tree will be at JPL. Does that mean 1. and 2. won't happen?

Got to go get donuts..... P is away until Wed. and A refuses to
read e-mail on holidays or weekends these days (his ISDN line is
too slow and neither the phone or the cable companies are
willing to run high-speed links to his house).
Since checkpoint56798 is several years out perhaps we should wait and
all talk on the phone on Thur?

Chris