intro to the main modifications to be done ...
Modifications to be done are of three kind:
After the elimination of the small namespaces linkmsg and ftool in release 3.0, it remains now four namespaces in the GetFEM++ project.
It is clear that the separation into these remaining four namespaces is mainly historical. The separate gmm namespace for Gmm++ is clearly justified. The contour of nemaspaces dal and bgeot is more vague. Historically, those two namespaces had their own justifications.
In the very begining of GetFEM++ (the first files was written in 1995) the S.T.L. was not available and the containers defined in the dal namespace was used everywhere. Now, in GetFEM++, the S.T.L. containers are mainly used. The remaining uses of dal containers are eather historical or due to the specificities of these containers. It is however clear that this is not the aim of the GetFEM++ project to developp new container concept. So, the use of the dal containers has to be as much as possible reduced.
Now, concerning bgeot, it was containing some other geometrical object at the begining and was originally designed to be a self-consistent library of geometric concepts. It slowly derived to be like it is now, a collection of algorithms and object definition more or less related to geometry (rtree, kdtree, ftool, polynomials ...).
The conclusion of this is that dal and bgeot namespaces can be advantageously merged to the getfem namespace, reducing to the minimum the use of the dal containers. This should be done preserving the backward compatibility. An intermediary study would be to see if the dal cannot be directly derived from S.T.L. containers preserving the used specificities.
Basic type of integer, real ... used. to be done.