A couple of things I think fall under this topic in terms of this idea: 

1)  Pulling parent/child relationships or peer/peer relationships in a predetermined order without duplication doesn't seem achievable. 

If I display two column sets, for instance, and I want parent 1 in 1 column and parent 2 in another:

• I can set relationship ordinals,  and this ensures I get a different relationship (person) in each column set.  But I can't couple the ordinal with a relationship type.  So I'll get whatever the 2nd relationship is to the student whether it's a parent or not.
• I can set a relationship type on each column filter, but again because I can't couple with an ordinal,  the duplication of those relationships (my guess) times out the lists so it doesn't run at all.  
• Trying to set filters for both ordinal and relationship type for both column sets seems too much for the list to handle, it times out.

So the idea here is to make the advanced lists able to couple ordinal with a specific type of relationship.  Ability to sort by gender on relationships would be amazing also.

2) Displaying data from multiple child objects is a challenge.  I can only get 2 child objects deep before the list times out.  For example,  I built a list with student user base, then parent/child relationship as a child object.  under parent child object I had user base, phone, and address objects.  I could extract parent user base data and parent phone data,  but addresses as a separate object was too much, the list timed out.

3) We should be able to output multiple course enrollments for a student without duplicating rows.