Hello all,

We are writing a 2D structured gridding package for a specific application. One of the design issues has to do with whether the block boundaries are defined by just the boundary grid points or by a smooth curve on which the grid points lie.

There are many implementation pros and cons to this issue but from a user's perspective how do you like to think about the block boundary:

1. Just points that can be manipulated as points

2. Points on a smooth curve

Thanks in advance.

(1). If the solver requires only the mesh point location, then, (x,y,z) of a mesh point is all you need. (2). If the solver require the edge curve of the mesh (cell), then, you need to provide the curve information. (3). As a user, I try not to worry about it, unless, I have to write the interface (between the mesh and the solver), which is unlikely for average users who are interested in applications.