Summary of NEC-Cerfacs meeting (15/02/2001) and CERFACS present position
    ************************************************************************

Participants: 
-------------
Kolja Kuse (NEC ESS), Rudi Fischer (NEC ESS), Rolf
Hempel (NEC CCRLE), Guy Lonsdale (NEC CCRLE), Hubert Ritzdorf (NEC
CCRLE), Andrea Piacentini (CERFACS), Sophie Valcke (CERFACS).

Agenda:
-------
-Introduction to the PRISM project, by Kolja Kuse
-Presentation of Oasis, by Sophie Valcke
-Presentation of Palm, by Andrea Piacentini
-Suggestions for the development of the 
 PRISM wp3a -Development of the PRISM System Coupler 

Summary of the discussion:
-------------------------
-It was agreed that the wp3a-Development of the PRISM System Coupler
will effectively be composed of the following tasks:

T1- Descriptions and specifications (mo 1-6)
T2- Development of a Model Generic Interface 
    to present sequential OASIS coupler (mo 1-12)
T3- Development of the PRISM coupler driver (mo 13-24)
T4- Development of the PRISM coupler 
    internal operation parallel library (mo 7-24 + 25-36)
T5- Development of the PRISM coupler parallel 
    communication library (mo 13-24 + 25-36)
T6- Writing of the PRISM coupler documentation (mo 30-36)

-It clearly came out that the precise functionalities of the PRISM
 coupler will have to be defined in the first 6 months of the project
 in accordance with the whole PRISM System Specifications (wp2a) and
 interacting strongly with the model people (wp3b-c-d-e-f-g-h).

-However, it also came out clearly that the PRISM coupler will most
 probably share many functionalities with the Palm coupler in
 development at CERFACS.

-There was a disagrement however on how to distribute the
 tasks T3, T4 and T5 between NEC and Cerfacs; both groups would prefer to
 do T3 and T5.

CERFACS present position (22/02/2001)
-------------------------------------

-As NEC-CCRLE are certainly the experts in communication, it is
 logical that they get tasks T3 and T5.
-As we at CERFACS have more experience in interpolation and other 
 operations to be performed on the coupling fields, it is logical that
 we get task T4.
-However, CERFACS developed a particular expertise in coupler design
 with the development of OASIS and PALM, and it is desirable that the
 design of the PRISM coupler be based on this experience. 
-We would propose the following steps:
1) Present in detail the Palm concepts and developments to the NEC CCRLE
 developers.
2) Identify precisely what concepts of Palm are relevant for PRISM.
3) Define how to incorporate in PRISM some developments already done
in PALM or vice-versa how to make PALM benefit from developments that
will be done in PRISM.