>The application handles workorders for a phone company.. For each >workorder, they are building a header table (One record) plus a detail >table(one record per part/labor item), plus a materials table is generated >from the detail table... Add to this, that the main data entry form is >layed out using private datasessions that have to be tracked (so that >reports can be run off the active form's data).. And you have a royal pain >in the A$$.. Now I see what you mean with >For one thing the database is a mess.. Sorry to say, but I've seen worse, though not many of them. One thing that crosses my mind is to cram these little tables into parallel tables, and delete them as soon as workorders are completed, otherwise you'll end up using the slowest piece of software you have: sequential directory search. Another thing that crossed my mind is that before deleting these tables you could copy them into three memo fields of some master table, just for future reference. OTOH, just look at the W95 - it's not only using separate files for all kinds of things, it's using directories as one'd use tables or keys. Big example of great programming, eh?