[MITgcm-devel] NAMELIST changes (how to and how not to)

[Patrick Heimbach]

Hi there,

after being puzzled for a day or two
why the ECCO code updated to a new model version
doesn't work anymore I discovered that the
namelist in SEAICE has been reduced in a
very non-smart way - by just dropping a variable (SEAICEuseEVP).

That is certainly NOT the way to go,
unless you want to make sure that the model
is hopelessly incompatible w.r.t. old data files,
or hopelessly difficult to catch such problems.

The way to do it better is as J.M. has demonstrated
for pkg/diagnostics/ :
Create a routine like diagnostics_list_check.F
which checks for retired NAMELIST variables
and stops the model with appropriate warning.
That's better than a crash with cryptic NAMELIST error msg (Altix)
or no crash at all with skipping of remainder (IBM SP).

Cheers
-p.

---
Patrick Heimbach | heimbach at mit.edu | http://www.mit.edu/~heimbach
MIT | EAPS 54-1518 | 77 Massachusetts Ave | Cambridge MA 02139 USA
FON +1-617-253-5259 | FAX +1-617-253-4464 | SKYPE patrick.heimbach