[Mitgcm-support] ECCO setup

[mitgcm-support at dev.mitgcm.org]

Dear all,

a CVS tree has been established under

http://mitgcm.lcs.mit.edu/cgi-bin/cvsweb/development/heimbach/ecco_env/

As the name says, it's still development, but should help to make this
process more transparent and to maintain a converging version.
I suggest, this should be the core code on which to build 
the ECCO release.
Therefore, it is important to maintain the different pieces.
Here's a list of suggested (or already agreed) package masters.
All bugs detected or improvements made should be reported to 
these people:
* model/ eesupp: Alistair & Chris (obvious)
* gmredi:     Alistair & Chris
* kpp:        Dimitris (possibly with help from Armin & Patrick?)
* autodiff:   Ralf
* exf:        Arne & Patrick
* cal:        Arne & Patrick
* cost:       Patrick & Ralf
* ctrl:       Patrick & Ralf
* ecco:       Patrick & Ralf
* optim:      Patrick
List may change over time. Will be included to CVS.
* lsopt:      Patrick

Major changes to previous release (ecco_c28_ralf):

- A few bug fixes in active_file.F, active_file_ad.F

- Renaming of packages subdirectory to comply with mitgcmuv:
  ./packages/   ->   ./pkg/

- All .[hF] routines belonging to packages were removed from 
  ./model/src/ and ./model/inc/
  to their respective package subdirectory,
  to avoid ambiguity in code appearing twice in different dir's.
  
  ./tools/genmake and ./wrk/makefile were adapted accordingly

- Packages that either
  * are not yet used (diva, sph, diagnos)
  * belong to offline optimization (optim, lsopt)
  were moved to ./pkg/offline

Major things t.b.d.:

- Further testing of link ecco <-> offline optim.
  (i.e. link between ctrl_pack.F/ctrl_unpack.F and optim package).
  N.B.: Offline optim itself, i.e. lsopt works fine.

- ./model/ ./eesupp/ is so far based upon mitgcmuv c28
  CVS tree should be changed to automatically link
  latest mitgcmuv into ecco environment.

- kpp:
  * Make sure it works (forward code)
  * Make sure, it is adjointable
  * Provide test case for kpp-specific test case
  * resolve gradient Richardson number-related problem

- gmredi
  * Provide gmredi-specific test case

- exf
  * Equation precipitation has been added to bulk formulae
    -> needs to be tested in conjunction with atmospheric state
       forcing (temperature, wind speed, spec. humidity)
       with precipitation as additional input.
  * For bulk forcing, atmospheric state fields  should be included
    as controls instead of flux fields.

- opbd
  * include open boundary as part of control problem.
  
Best regards
Patrick
___________________________________________________________________
Patrick Heimbach                        Phon: +1 / 617 / 253 - 5259
Massachusetts Institute of Technology   Facs: +1 / 617 / 253 - 4464 
EAPS, Room 54-1518
77 Massachusetts Avenue                     mailto:heimbach at mit.edu
Cambridge MA 02139, U.S.A.            http://www.mit.edu/~heimbach/