worked example

All,

I've uploaded about 90% of Michael's worked example to the basin_kernel
CVS.

At this point, there's an important series of questions which we (though
primarily Steve, Enrico, Louis, and I) should address. These include:

1) How do we want to structure/include "operation" functions in the
distribution?

At this point, I have a bunch of starter functions (1 cosmology
function, a bunch of density functions, a proto-convolution function,
etc). In principle, these should be kept separate from the core basin
classes (+member functions). However, some (like compute_density), may
also double as member functions. I think we want to organize this a bit
better. For now, everything I've written is in the monolithic
"basin_goldberg.C" or under "compute_density.C" in the ordinary src
directory.

2) How many different ways do we want to be able to call the same
functionality?

As an example:

Grid g=l.compute_density("x","y","z","mass","newdens");
Attribute dens=compute_density(Attr x, Attr y, Attr z);
etc.
They can easily (or relatively easily, once new objects like pointers to
attributes, and attribute_vectors are introuced) be used as wrappers to
one another, but in general, we won't be able to anticipate everything.

3) Can we think of ways to compactify my code (though helper functions,
new containers and the like, or through global variables), in order to
make things easier on the user end?

Take a look and see if you have ideas.

-----

I'd like all interested parties to look at the code BEFORE tomorrows
meeting so that we can bring suggestions to the table. I don't plan on
either reading my code, or reading this email.

Dave
Received on Thu Jun 8 14:40:35 2006