[MITgcm-devel] ecco seaicev4 cost-name change

[Jean-Michel Campin]

Hello An,

I am not really following pkg/ecco developpment, but my preference would go to:
> a) completely replace the old names with new ones in code and put in ERROR + STOP statement if the old names are used,
What does Patrick think of this ?

Cheers,
Jean-Michel

On Tue, Apr 01, 2014 at 06:33:07PM -0400, An T Nguyen wrote:
> hello,
> 
> Back in January 2014, Gael, Ian, and I agreed to implement Ian's seaicev4 proxy cost terms as a function of sst into the pkg/ecco.  I've just updated the code to reflect that.  
> 
> Related to cost formulation, we also agreed to change the cost names to make them more reflective of what each cost is.
> 
> Currently, the 3 cost names are:
> (1) siv4-conc <-- misfit in ice concentration
> (2) siv4-sst    <-- data have ice but model has no ice (deficiency of ice)
> (3) siv4-vol   <--  data have no ice but model has ice (excess of ice)
> 
> I propose changing the names of (2) and (3) to:
> (2) siv4-deconc (deficiency of ice)
> (3) siv4-exconc (excess of ice)
> 
> Any thought / opinion / suggestion on these names?
> 
> Lastly, the implementation part: I can either
> 
> a) completely replace the old names with new ones in code and put in ERROR + STOP statement if the old names are used, OR
> b) completely replace old with new in code with WARNING statement that cost calc for terms (2) and (3) are skipped due to unmatched names, OR
> c) leave both old and new in and if
>     i) only old is provided in data.ecco, set new = old and proceed to calc cost.
>    ii) both old and new names are provided in data.ecco, WARNING of conflict and only use new, proceed to calc cost.
>   iii) only new is provided, proceed to calc cost.
> 
> Option (c) gives backward compatibility (I think) but also is messy and can cause confusion down the line.
> 
> What do you think?
> 
> An
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel