>>>>Seems like there is no easy answer. We are going to try to see what the overhead is. One would think that the only overhead would be at instanciation, and destroy, but maybe formset events processing has a cost associated. Some of our automated processes are very real time intensive, so a small overhead in event processing may affect the long term results.
>>>
>>>I'll be interested in the results of your test.
>>
>>OK, I'll post them. Trying to decide on some good benchmark tests now.
>>Simple, but reasonably accurate, in a multi-task environment ? Suggestions ?
>
>Maybe testing instantiation performance, overhead associated with thisformset.method() calls, and thisformset.property usage vs. thisform.method() and thisform.property. Just create and destroy a formset object that contains a form and compar this with creating and destroying a form object only. Testing method performance might best be done on a method that does nothing but return. This would focus in an just the calling overhead. Doing many iterations on calculations that reference thisformset properties vs. thisform properties may point out some additional performance differences.

Well, the consensus of the team here at work was that we do not want to invest the time to attempt testing. The one place we are most interested in knowing the results is where we are doing automated processing. These are forms that run in a 24x7 environment, and have DOEVENTS carfully placed in the code to allow a Cancel button to function reliabily. The overhead of the DOEVENTS in difficult to test, since it includes all Window events. We have decided to keep these type of programs on single forms, and suffer any consequences of this decision.

Thanks to everyone who responded, it was helpful !