On Mon, 17 Mar 2008, Matthew Knepley wrote: > On Mon, Mar 17, 2008 at 1:06 PM, Barry Smith <bsmith at mcs.anl.gov> wrote: > > > > Fortran90 has namespaces?????? > > Not in the way I was thinking. Damn F90. Anyway, it looks like you > can selectively use interface modules, so we might be able to get > away with redundant names by just not using them together.
> I jsut really hate the idea of putting "PETSc" in front of every > word in the package. It is really the ugliest thing I can imagine > and will make programming that much more of a slog. We already do this for lot of things. [this wasn't the case when we started] - all petsc libraries now have a 'petsc' prefix. - all configure flags have petsc prefix [this wasn't the case - when we started]. - all petsc datatypes now have petsc prefix. And wrt namesapcing, the correct thing to do is expand the namespace to languages without it. [not ingore it]. For eg: All MPI rutines have 'MPI' name space. The way its implemented is: MPI::Comm_rank() - c++ MPI_Comm_rank() - C,F77 [but just because c doesn't have namespace -they didn't decide to use just Comm_rank() for c] The biggest issue is: PETSc tries to use full names for functions, and we usually exceed 32char limit on some machines. And additional 5 chars to each function name might tip us over this limit. Satish
