>And the problem we had at the same time which has NEVER been solved: Why does "Append blank" not work on the same problem network about one time in 100,000. INSERT caused the same problems, and of course there isn't any return value which says "Record Inserted" So we now have loops to actually check to see if the record pointer was on a new record!
I believe the only way to explain this to a layman is "in a 500 rooms hotel, one tap leaks five minutes a month. find and fix.".