I've got numerous "tools" that essentially place a cell. In thinking about upgrading these tools to CONNECT, I'm wondering if my method could be improved. The cell libraries reside on a server. In V8i, my typical approach is to read the cell into a MSElementDescr, with a scale of 1.0, rotation of 0.0, and an origin of 0,0,0. In the tool's dynamics functions, I duplicate the ElementDescr, transform it with TransformInfo, display it, then release it.
I'm wondering if I update my code to where the cell is initially read with my desired scale, rotation, and origin, then just displayed in the dynamics function, if that's any "better" (no duplication of ElementDescr). My thought is that the constant "read" of the cell from the library on the server would be a big performance "hit", or maybe CONNECT, when started, caches the cell library locally in memory such that there may be no real "penalty" for always "reading" the cell.
Bruce