before reading the preferred design model on csns , what I had in my mind regarding RubricScore , was to have two field for the actual score , one , a Transient annotation field that does not get saved in database and only used in the object side , two , a String representation of the score that combined the list together with a separator and then get saved into database , and whenever we retrieve the data back we could just split the individuals based on that separator . but I did not see any mention to an extra field in the preferred design , so I assume we should just go by this and save the individual data piece in the database . which is sort of troublesome in this case since we do not have a certain number for the data columns (depends on the size / scale of the Rubrics) .

the other thing is that we have a list type for a field it gets translated to a "bytea" type . might not be a good question , does this mean that it somehow combines the List elements together and saves it into database ?  and we we get it back by Dao and Entity Manager (behind the scene) it automatically return a list of that one "bytea" type ?