No, this has nothing to do with the order of the instances. It has to do with only adding a subset of instances for repeating data. For example, I have a section of my XML data that has x repeating nodes, with a bunch of data underneath. I want a subform to be added for each of those repeating data nodes, but I only want an instance in certain cases. Whether they are displayed or not depends on some complex rules. So I would like to have script in the container subform that loops through the repeating data nodes and decides if an instance needs to be added. If so, the script would do an addInstance and set the data binding to the specific data node. Then the fields inside the subform instance can just reference the data relative to the subform's binding and map correctly to the underlying data.
The alternative is to just bind the subform to the repeating data and perform the data check inside the subform, hiding the instance if it shouldn't be displayed. However the fields underneath are all bound directly to underlying data which will fire a bunch of events, and it seems wasteful to go through all of that processing just to have the whole subform hidden.
