Hi Jim,
I can't quite count the number of times I've gone over that chapter of yours on data classes to help me with this construction. Thanks.

I was timing the runtime part of the prg file (no form involved) that I originally submitted in this thread. So the time should have summed the time for instantiation of session/DE/cursor, and loading of all files.

When I timed this I got ~.3seconds. Then I got rid of the DE and used USE to open the tables and got the same times roughly. Then I realised that I would want to put the session object into the custom object or some such container that supported the addobject method, for the visual organization of my code. I was real concerned that now I had 4 objects connected and that might get me worse times.

Now with the same timing code but adding custom object addobjecting session the time should have included custom/session/DE/cursors and opentables method. That is when I got the much smaller time. Then I put the custom object into a form and bang we are back to the .3second time again.

The reason I am concerned with time, is that if a website is calling my modules through WebConnect, the .3seconds is about 3 times the normal processing time of an average program run with WC. I wish I could understand where that .004 time is coming from, though it is probably bogus.

>>Thank all for replies to this. Something interesting I've just noticed after getting this working:
>>
>>When I create a session object and reference from it the DE and then addobject 8 cursors with 11 megabytes involved, the time to instantiate everything and open the cursors is about .3 seconds. If I do the same thing but addobject the session object to a custom object, the time for everything to occur is only .004 seconds. The data is really there also as I can access any of the tables!! That was quite a surprise. Any ideas on this?
>
>Terry,
>
>Are you measuring the same time interval for both of them? It is possible that the DE time includes the instantiation of the form, the DE, and the Session object.
>
>Anyway with a difference of .296 seconds, I would go with the best logical design and forget about the time. It is not like you create these things 1,000,000 times in a loop.