I know how to write an old-style console I/O program, but I'm new to the gui stuff. I've been plowing through examples in the book by Ivor Horton, which introduces "documents" and "views" as part of a sketching/drawing app example.
Missing from the book are examples of how computationally complex scientific programs should be structured. At first, I thought of putting the guts in the main messag loop, but that didn't seem to fit the paradigm. Then jit occurred to me that my hydrodyamic grids might fit as "documents."
Or perhaps the simulation should be a console program that communicates with the viewer by pipes?
Rather than write a weirdly structured program, I ask for advice on what is the norm these days for scientific programming with a possible future aspect of realtime view, and which can easily be unrolled into a stream format.