>First I need to clarify some syntax. I will use scxform to refer to the form file (the .scx file) and objform to refer to the form which is an object in the scxform. I read in the early advisories on VFP that you should use a formset in the scxform and put all the objforms you will need for a specific task in it. If you reuse an objform make it a class. When you do this and include the class based on the objform, suprisingly, you don't save anything on the size of the scxform. It simply copies the class into the scxform. While you gain all the encapsulation/inheritance stuff which is great you don't save on loading time.
>
>The question to challenge convention is should you divide your complex scxforms into smaller scxforms that call each other/pass info to each other etc so you can improve loading time as you only load the formset with the objforms you need. Then if you need another scxform it will load resulting in more waiting while the task is ongoing but much less upfront.
>
>I don't think this is the way most people do things but was interested in discussion on the pros and cons of doing this.

That's an interesting point. Having programmed with both single-form apps (first) and formsets (for awhile I thought they were the only way to go), I've finally settled on a mixture, where I use a nucleus formset for screens/objects that are constantly needed, and I farm out less-used stuff to forms which are opened and closed as needed. One problem with this method is that some objects are difficult (as far as usage frequency) to place, in the above two groupings.

But the objective seems right, which is to get the overall most time-efficient product you can, balancing initial load time with interactive wait times...(now that memory is not too much of a concern, I won't go into that aspect).